mirror of
https://github.com/immich-app/immich.git
synced 2025-10-29 01:32:34 -04:00
feat: ocr (#18836)
* feat: add OCR functionality and related configurations * chore: update labeler configuration for machine learning files * feat(i18n): enhance OCR model descriptions and add orientation classification and unwarping features * chore: update Dockerfile to include ccache for improved build performance * feat(ocr): enhance OCR model configuration with orientation classification and unwarping options, update PaddleOCR integration, and improve response structure * refactor(ocr): remove OCR_CLEANUP job from enum and type definitions * refactor(ocr): remove obsolete OCR entity and migration files, and update asset job status and schema to accommodate new OCR table structure * refactor(ocr): update OCR schema and response structure to use individual coordinates instead of bounding box, and adjust related service and repository files * feat: enhance OCR configuration and functionality - Updated OCR settings to include minimum detection box score, minimum detection score, and minimum recognition score. - Refactored PaddleOCRecognizer to utilize new scoring parameters. - Introduced new database tables for asset OCR data and search functionality. - Modified related services and repositories to support the new OCR features. - Updated translations for improved clarity in settings UI. * sql changes * use rapidocr * change dto * update web * update lock * update api * store positions as normalized floats * match column order in db * update admin ui settings descriptions fix max resolution key set min threshold to 0.1 fix bind * apply config correctly, adjust defaults * unnecessary model type * unnecessary sources * fix(ocr): switch RapidOCR lang type from LangDet to LangRec * fix(ocr): expose lang_type (LangRec.CH) and font_path on OcrOptions for RapidOCR * fix(ocr): make OCR text search case- and accent-insensitive using ILIKE + unaccent * fix(ocr): add OCR search fields * fix: Add OCR database migration and update ML prediction logic. * trigrams are already case insensitive * add tests * format * update migrations * wrong uuid function * linting * maybe fix medium tests * formatting * fix weblate check * openapi * sql * minor fixes * maybe fix medium tests part 2 * passing medium tests * format web * readd sql * format dart * disabled in e2e * chore: translation ordering --------- Co-authored-by: mertalev <101130780+mertalev@users.noreply.github.com> Co-authored-by: Alex Tran <alex.tran1502@gmail.com>
This commit is contained in:
Kang
GitHub
parent
c666dc6c67
commit
02b29046b3
2
.github/labeler.yml
vendored
2
.github/labeler.yml
vendored
@ -31,7 +31,7 @@ documentation:
|
||||
🧠machine-learning:
|
||||
- changed-files:
|
||||
- any-glob-to-any-file:
|
||||
- machine-learning/app/**
|
||||
- machine-learning/**
|
||||
|
||||
changelog:translation:
|
||||
- head-branch: ['^chore/translations$']
|
||||
|
||||
@ -122,7 +122,7 @@ services:
|
||||
ports:
|
||||
- 3003:3003
|
||||
volumes:
|
||||
- ../machine-learning:/usr/src/app
|
||||
- ../machine-learning/immich_ml:/usr/src/immich_ml
|
||||
- model-cache:/cache
|
||||
env_file:
|
||||
- .env
|
||||
|
||||
@ -113,6 +113,7 @@ describe('/server', () => {
|
||||
importFaces: false,
|
||||
oauth: false,
|
||||
oauthAutoLaunch: false,
|
||||
ocr: false,
|
||||
passwordLogin: true,
|
||||
search: true,
|
||||
sidecar: true,
|
||||
|
||||
17
i18n/en.json
17
i18n/en.json
@ -154,6 +154,18 @@
|
||||
"machine_learning_min_detection_score_description": "Minimum confidence score for a face to be detected from 0-1. Lower values will detect more faces but may result in false positives.",
|
||||
"machine_learning_min_recognized_faces": "Minimum recognized faces",
|
||||
"machine_learning_min_recognized_faces_description": "The minimum number of recognized faces for a person to be created. Increasing this makes Facial Recognition more precise at the cost of increasing the chance that a face is not assigned to a person.",
|
||||
"machine_learning_ocr": "OCR",
|
||||
"machine_learning_ocr_description": "Use machine learning to recognize text in images",
|
||||
"machine_learning_ocr_enabled": "Enable OCR",
|
||||
"machine_learning_ocr_enabled_description": "If disabled, images will not undergo text recognition.",
|
||||
"machine_learning_ocr_max_resolution": "Maximum resolution",
|
||||
"machine_learning_ocr_max_resolution_description": "Previews above this resolution will be resized while preserving aspect ratio. Higher values are more accurate, but take longer to process and use more memory.",
|
||||
"machine_learning_ocr_min_detection_score": "Minimum detection score",
|
||||
"machine_learning_ocr_min_detection_score_description": "Minimum confidence score for text to be detected from 0-1. Lower values will detect more text but may result in false positives.",
|
||||
"machine_learning_ocr_min_recognition_score": "Minimum recognition score",
|
||||
"machine_learning_ocr_min_score_recognition_description": "Minimum confidence score for detected text to be recognized from 0-1. Lower values will recognize more text but may result in false positives.",
|
||||
"machine_learning_ocr_model": "OCR model",
|
||||
"machine_learning_ocr_model_description": "Server models are more accurate than mobile models, but take longer to process and use more memory.",
|
||||
"machine_learning_settings": "Machine Learning Settings",
|
||||
"machine_learning_settings_description": "Manage machine learning features and settings",
|
||||
"machine_learning_smart_search": "Smart Search",
|
||||
@ -245,6 +257,7 @@
|
||||
"oauth_storage_quota_default_description": "Quota in GiB to be used when no claim is provided.",
|
||||
"oauth_timeout": "Request Timeout",
|
||||
"oauth_timeout_description": "Timeout for requests in milliseconds",
|
||||
"ocr_job_description": "Use machine learning to recognize text in images",
|
||||
"password_enable_description": "Login with email and password",
|
||||
"password_settings": "Password Login",
|
||||
"password_settings_description": "Manage password login settings",
|
||||
@ -1438,6 +1451,7 @@
|
||||
"oauth": "OAuth",
|
||||
"obtainium_configurator": "Obtainium Configurator",
|
||||
"obtainium_configurator_instructions": "Use Obtainium to install and update the Android app directly from Immich GitHub's release. Create an API key and select a variant to create your Obtainium configuration link",
|
||||
"ocr": "OCR",
|
||||
"official_immich_resources": "Official Immich Resources",
|
||||
"offline": "Offline",
|
||||
"offset": "Offset",
|
||||
@ -1715,6 +1729,8 @@
|
||||
"search_by_description_example": "Hiking day in Sapa",
|
||||
"search_by_filename": "Search by file name or extension",
|
||||
"search_by_filename_example": "i.e. IMG_1234.JPG or PNG",
|
||||
"search_by_ocr": "Search by OCR",
|
||||
"search_by_ocr_example": "Latte",
|
||||
"search_camera_lens_model": "Search lens model...",
|
||||
"search_camera_make": "Search camera make...",
|
||||
"search_camera_model": "Search camera model...",
|
||||
@ -1732,6 +1748,7 @@
|
||||
"search_filter_location_title": "Select location",
|
||||
"search_filter_media_type": "Media Type",
|
||||
"search_filter_media_type_title": "Select media type",
|
||||
"search_filter_ocr": "Search by OCR",
|
||||
"search_filter_people_title": "Select people",
|
||||
"search_for": "Search for",
|
||||
"search_for_existing_person": "Search for existing person",
|
||||
|
||||
@ -141,7 +141,7 @@ FROM prod-${DEVICE} AS prod
|
||||
ARG DEVICE
|
||||
|
||||
RUN apt-get update && \
|
||||
apt-get install -y --no-install-recommends tini $(if ! [ "$DEVICE" = "openvino" ] && ! [ "$DEVICE" = "rocm" ]; then echo "libmimalloc2.0"; fi) && \
|
||||
apt-get install -y --no-install-recommends tini ccache libgl1 libglib2.0-0 libgomp1 $(if ! [ "$DEVICE" = "openvino" ] && ! [ "$DEVICE" = "rocm" ]; then echo "libmimalloc2.0"; fi) && \
|
||||
apt-get autoremove -yqq && \
|
||||
apt-get clean && \
|
||||
rm -rf /var/lib/apt/lists/*
|
||||
|
||||
@ -41,6 +41,7 @@ class PreloadModelData(BaseModel):
|
||||
|
||||
class MaxBatchSize(BaseModel):
|
||||
facial_recognition: int | None = None
|
||||
text_recognition: int | None = None
|
||||
|
||||
|
||||
class Settings(BaseSettings):
|
||||
|
||||
@ -183,7 +183,9 @@ async def run_inference(payload: Image | str, entries: InferenceEntries) -> Infe
|
||||
response: InferenceResponse = {}
|
||||
|
||||
async def _run_inference(entry: InferenceEntry) -> None:
|
||||
model = await model_cache.get(entry["name"], entry["type"], entry["task"], ttl=settings.model_ttl)
|
||||
model = await model_cache.get(
|
||||
entry["name"], entry["type"], entry["task"], ttl=settings.model_ttl, **entry["options"]
|
||||
)
|
||||
inputs = [payload]
|
||||
for dep in model.depends:
|
||||
try:
|
||||
|
||||
@ -3,6 +3,8 @@ from typing import Any
|
||||
from immich_ml.models.base import InferenceModel
|
||||
from immich_ml.models.clip.textual import MClipTextualEncoder, OpenClipTextualEncoder
|
||||
from immich_ml.models.clip.visual import OpenClipVisualEncoder
|
||||
from immich_ml.models.ocr.detection import TextDetector
|
||||
from immich_ml.models.ocr.recognition import TextRecognizer
|
||||
from immich_ml.schemas import ModelSource, ModelTask, ModelType
|
||||
|
||||
from .constants import get_model_source
|
||||
@ -28,6 +30,12 @@ def get_model_class(model_name: str, model_type: ModelType, model_task: ModelTas
|
||||
case ModelSource.INSIGHTFACE, ModelType.RECOGNITION, ModelTask.FACIAL_RECOGNITION:
|
||||
return FaceRecognizer
|
||||
|
||||
case ModelSource.PADDLE, ModelType.DETECTION, ModelTask.OCR:
|
||||
return TextDetector
|
||||
|
||||
case ModelSource.PADDLE, ModelType.RECOGNITION, ModelTask.OCR:
|
||||
return TextRecognizer
|
||||
|
||||
case _:
|
||||
raise ValueError(f"Unknown model combination: {source}, {model_type}, {model_task}")
|
||||
|
||||
|
||||
@ -38,9 +38,8 @@ class InferenceModel(ABC):
|
||||
|
||||
def download(self) -> None:
|
||||
if not self.cached:
|
||||
log.info(
|
||||
f"Downloading {self.model_type.replace('-', ' ')} model '{self.model_name}'. This may take a while."
|
||||
)
|
||||
model_type = self.model_type.replace("-", " ")
|
||||
log.info(f"Downloading {model_type} model '{self.model_name}' to {self.model_path}. This may take a while.")
|
||||
self._download()
|
||||
|
||||
def load(self) -> None:
|
||||
@ -58,7 +57,7 @@ class InferenceModel(ABC):
|
||||
self.load()
|
||||
if model_kwargs:
|
||||
self.configure(**model_kwargs)
|
||||
return self._predict(*inputs, **model_kwargs)
|
||||
return self._predict(*inputs)
|
||||
|
||||
@abstractmethod
|
||||
def _predict(self, *inputs: Any, **model_kwargs: Any) -> Any: ...
|
||||
|
||||
@ -19,7 +19,7 @@ class BaseCLIPTextualEncoder(InferenceModel):
|
||||
depends = []
|
||||
identity = (ModelType.TEXTUAL, ModelTask.SEARCH)
|
||||
|
||||
def _predict(self, inputs: str, language: str | None = None, **kwargs: Any) -> str:
|
||||
def _predict(self, inputs: str, language: str | None = None) -> str:
|
||||
tokens = self.tokenize(inputs, language=language)
|
||||
res: NDArray[np.float32] = self.session.run(None, tokens)[0][0]
|
||||
return serialize_np_array(res)
|
||||
|
||||
@ -26,7 +26,7 @@ class BaseCLIPVisualEncoder(InferenceModel):
|
||||
depends = []
|
||||
identity = (ModelType.VISUAL, ModelTask.SEARCH)
|
||||
|
||||
def _predict(self, inputs: Image.Image | bytes, **kwargs: Any) -> str:
|
||||
def _predict(self, inputs: Image.Image | bytes) -> str:
|
||||
image = decode_pil(inputs)
|
||||
res: NDArray[np.float32] = self.session.run(None, self.transform(image))[0][0]
|
||||
return serialize_np_array(res)
|
||||
|
||||
@ -75,6 +75,11 @@ _INSIGHTFACE_MODELS = {
|
||||
}
|
||||
|
||||
|
||||
_PADDLE_MODELS = {
|
||||
"PP-OCRv5_server",
|
||||
"PP-OCRv5_mobile",
|
||||
}
|
||||
|
||||
SUPPORTED_PROVIDERS = [
|
||||
"CUDAExecutionProvider",
|
||||
"ROCMExecutionProvider",
|
||||
@ -159,4 +164,7 @@ def get_model_source(model_name: str) -> ModelSource | None:
|
||||
if cleaned_name in _OPENCLIP_MODELS:
|
||||
return ModelSource.OPENCLIP
|
||||
|
||||
if cleaned_name in _PADDLE_MODELS:
|
||||
return ModelSource.PADDLE
|
||||
|
||||
return None
|
||||
|
||||
@ -24,7 +24,7 @@ class FaceDetector(InferenceModel):
|
||||
|
||||
return session
|
||||
|
||||
def _predict(self, inputs: NDArray[np.uint8] | bytes, **kwargs: Any) -> FaceDetectionOutput:
|
||||
def _predict(self, inputs: NDArray[np.uint8] | bytes) -> FaceDetectionOutput:
|
||||
inputs = decode_cv2(inputs)
|
||||
|
||||
bboxes, landmarks = self._detect(inputs)
|
||||
|
||||
@ -44,7 +44,7 @@ class FaceRecognizer(InferenceModel):
|
||||
return session
|
||||
|
||||
def _predict(
|
||||
self, inputs: NDArray[np.uint8] | bytes | Image.Image, faces: FaceDetectionOutput, **kwargs: Any
|
||||
self, inputs: NDArray[np.uint8] | bytes | Image.Image, faces: FaceDetectionOutput
|
||||
) -> FacialRecognitionOutput:
|
||||
if faces["boxes"].shape[0] == 0:
|
||||
return []
|
||||
|
||||
86
machine-learning/immich_ml/models/ocr/detection.py
Normal file
86
machine-learning/immich_ml/models/ocr/detection.py
Normal file
@ -0,0 +1,86 @@
|
||||
from typing import Any
|
||||
|
||||
import numpy as np
|
||||
from PIL import Image
|
||||
from rapidocr.ch_ppocr_det import TextDetector as RapidTextDetector
|
||||
from rapidocr.inference_engine.base import FileInfo, InferSession
|
||||
from rapidocr.utils import DownloadFile, DownloadFileInput
|
||||
from rapidocr.utils.typings import EngineType, LangDet, OCRVersion, TaskType
|
||||
from rapidocr.utils.typings import ModelType as RapidModelType
|
||||
|
||||
from immich_ml.config import log
|
||||
from immich_ml.models.base import InferenceModel
|
||||
from immich_ml.models.transforms import decode_cv2
|
||||
from immich_ml.schemas import ModelFormat, ModelSession, ModelTask, ModelType
|
||||
from immich_ml.sessions.ort import OrtSession
|
||||
|
||||
from .schemas import OcrOptions, TextDetectionOutput
|
||||
|
||||
|
||||
class TextDetector(InferenceModel):
|
||||
depends = []
|
||||
identity = (ModelType.DETECTION, ModelTask.OCR)
|
||||
|
||||
def __init__(self, model_name: str, **model_kwargs: Any) -> None:
|
||||
super().__init__(model_name, **model_kwargs, model_format=ModelFormat.ONNX)
|
||||
self.max_resolution = 736
|
||||
self.min_score = 0.5
|
||||
self.score_mode = "fast"
|
||||
self._empty: TextDetectionOutput = {
|
||||
"image": np.empty(0, dtype=np.float32),
|
||||
"boxes": np.empty(0, dtype=np.float32),
|
||||
"scores": np.empty(0, dtype=np.float32),
|
||||
}
|
||||
|
||||
def _download(self) -> None:
|
||||
model_info = InferSession.get_model_url(
|
||||
FileInfo(
|
||||
engine_type=EngineType.ONNXRUNTIME,
|
||||
ocr_version=OCRVersion.PPOCRV5,
|
||||
task_type=TaskType.DET,
|
||||
lang_type=LangDet.CH,
|
||||
model_type=RapidModelType.MOBILE if "mobile" in self.model_name else RapidModelType.SERVER,
|
||||
)
|
||||
)
|
||||
download_params = DownloadFileInput(
|
||||
file_url=model_info["model_dir"],
|
||||
sha256=model_info["SHA256"],
|
||||
save_path=self.model_path,
|
||||
logger=log,
|
||||
)
|
||||
DownloadFile.run(download_params)
|
||||
|
||||
def _load(self) -> ModelSession:
|
||||
# TODO: support other runtime sessions
|
||||
session = OrtSession(self.model_path)
|
||||
self.model = RapidTextDetector(
|
||||
OcrOptions(
|
||||
session=session.session,
|
||||
limit_side_len=self.max_resolution,
|
||||
limit_type="min",
|
||||
box_thresh=self.min_score,
|
||||
score_mode=self.score_mode,
|
||||
)
|
||||
)
|
||||
return session
|
||||
|
||||
def _predict(self, inputs: bytes | Image.Image) -> TextDetectionOutput:
|
||||
results = self.model(decode_cv2(inputs))
|
||||
if results.boxes is None or results.scores is None or results.img is None:
|
||||
return self._empty
|
||||
return {
|
||||
"image": results.img,
|
||||
"boxes": np.array(results.boxes, dtype=np.float32),
|
||||
"scores": np.array(results.scores, dtype=np.float32),
|
||||
}
|
||||
|
||||
def configure(self, **kwargs: Any) -> None:
|
||||
if (max_resolution := kwargs.get("maxResolution")) is not None:
|
||||
self.max_resolution = max_resolution
|
||||
self.model.limit_side_len = max_resolution
|
||||
if (min_score := kwargs.get("minScore")) is not None:
|
||||
self.min_score = min_score
|
||||
self.model.postprocess_op.box_thresh = min_score
|
||||
if (score_mode := kwargs.get("scoreMode")) is not None:
|
||||
self.score_mode = score_mode
|
||||
self.model.postprocess_op.score_mode = score_mode
|
||||
117
machine-learning/immich_ml/models/ocr/recognition.py
Normal file
117
machine-learning/immich_ml/models/ocr/recognition.py
Normal file
@ -0,0 +1,117 @@
|
||||
from typing import Any
|
||||
|
||||
import cv2
|
||||
import numpy as np
|
||||
from numpy.typing import NDArray
|
||||
from PIL.Image import Image
|
||||
from rapidocr.ch_ppocr_rec import TextRecInput
|
||||
from rapidocr.ch_ppocr_rec import TextRecognizer as RapidTextRecognizer
|
||||
from rapidocr.inference_engine.base import FileInfo, InferSession
|
||||
from rapidocr.utils import DownloadFile, DownloadFileInput
|
||||
from rapidocr.utils.typings import EngineType, LangRec, OCRVersion, TaskType
|
||||
from rapidocr.utils.typings import ModelType as RapidModelType
|
||||
|
||||
from immich_ml.config import log, settings
|
||||
from immich_ml.models.base import InferenceModel
|
||||
from immich_ml.schemas import ModelFormat, ModelSession, ModelTask, ModelType
|
||||
from immich_ml.sessions.ort import OrtSession
|
||||
|
||||
from .schemas import OcrOptions, TextDetectionOutput, TextRecognitionOutput
|
||||
|
||||
|
||||
class TextRecognizer(InferenceModel):
|
||||
depends = [(ModelType.DETECTION, ModelTask.OCR)]
|
||||
identity = (ModelType.RECOGNITION, ModelTask.OCR)
|
||||
|
||||
def __init__(self, model_name: str, **model_kwargs: Any) -> None:
|
||||
self.min_score = model_kwargs.get("minScore", 0.9)
|
||||
self._empty: TextRecognitionOutput = {
|
||||
"box": np.empty(0, dtype=np.float32),
|
||||
"boxScore": np.empty(0, dtype=np.float32),
|
||||
"text": [],
|
||||
"textScore": np.empty(0, dtype=np.float32),
|
||||
}
|
||||
super().__init__(model_name, **model_kwargs, model_format=ModelFormat.ONNX)
|
||||
|
||||
def _download(self) -> None:
|
||||
model_info = InferSession.get_model_url(
|
||||
FileInfo(
|
||||
engine_type=EngineType.ONNXRUNTIME,
|
||||
ocr_version=OCRVersion.PPOCRV5,
|
||||
task_type=TaskType.REC,
|
||||
lang_type=LangRec.CH,
|
||||
model_type=RapidModelType.MOBILE if "mobile" in self.model_name else RapidModelType.SERVER,
|
||||
)
|
||||
)
|
||||
download_params = DownloadFileInput(
|
||||
file_url=model_info["model_dir"],
|
||||
sha256=model_info["SHA256"],
|
||||
save_path=self.model_path,
|
||||
logger=log,
|
||||
)
|
||||
DownloadFile.run(download_params)
|
||||
|
||||
def _load(self) -> ModelSession:
|
||||
# TODO: support other runtimes
|
||||
session = OrtSession(self.model_path)
|
||||
self.model = RapidTextRecognizer(
|
||||
OcrOptions(
|
||||
session=session.session,
|
||||
rec_batch_num=settings.max_batch_size.text_recognition if settings.max_batch_size is not None else 6,
|
||||
rec_img_shape=(3, 48, 320),
|
||||
)
|
||||
)
|
||||
return session
|
||||
|
||||
def _predict(self, _: Image, texts: TextDetectionOutput) -> TextRecognitionOutput:
|
||||
boxes, img, box_scores = texts["boxes"], texts["image"], texts["scores"]
|
||||
if boxes.shape[0] == 0:
|
||||
return self._empty
|
||||
rec = self.model(TextRecInput(img=self.get_crop_img_list(img, boxes)))
|
||||
if rec.txts is None:
|
||||
return self._empty
|
||||
|
||||
height, width = img.shape[0:2]
|
||||
boxes[:, :, 0] /= width
|
||||
boxes[:, :, 1] /= height
|
||||
|
||||
text_scores = np.array(rec.scores)
|
||||
valid_text_score_idx = text_scores > self.min_score
|
||||
valid_score_idx_list = valid_text_score_idx.tolist()
|
||||
return {
|
||||
"box": boxes.reshape(-1, 8)[valid_text_score_idx].reshape(-1),
|
||||
"text": [rec.txts[i] for i in range(len(rec.txts)) if valid_score_idx_list[i]],
|
||||
"boxScore": box_scores[valid_text_score_idx],
|
||||
"textScore": text_scores[valid_text_score_idx],
|
||||
}
|
||||
|
||||
def get_crop_img_list(self, img: NDArray[np.float32], boxes: NDArray[np.float32]) -> list[NDArray[np.float32]]:
|
||||
img_crop_width = np.maximum(
|
||||
np.linalg.norm(boxes[:, 1] - boxes[:, 0], axis=1), np.linalg.norm(boxes[:, 2] - boxes[:, 3], axis=1)
|
||||
).astype(np.int32)
|
||||
img_crop_height = np.maximum(
|
||||
np.linalg.norm(boxes[:, 0] - boxes[:, 3], axis=1), np.linalg.norm(boxes[:, 1] - boxes[:, 2], axis=1)
|
||||
).astype(np.int32)
|
||||
pts_std = np.zeros((img_crop_width.shape[0], 4, 2), dtype=np.float32)
|
||||
pts_std[:, 1:3, 0] = img_crop_width[:, None]
|
||||
pts_std[:, 2:4, 1] = img_crop_height[:, None]
|
||||
|
||||
img_crop_sizes = np.stack([img_crop_width, img_crop_height], axis=1).tolist()
|
||||
imgs: list[NDArray[np.float32]] = []
|
||||
for box, pts_std, dst_size in zip(list(boxes), list(pts_std), img_crop_sizes):
|
||||
M = cv2.getPerspectiveTransform(box, pts_std)
|
||||
dst_img: NDArray[np.float32] = cv2.warpPerspective(
|
||||
img,
|
||||
M,
|
||||
dst_size,
|
||||
borderMode=cv2.BORDER_REPLICATE,
|
||||
flags=cv2.INTER_CUBIC,
|
||||
) # type: ignore
|
||||
dst_height, dst_width = dst_img.shape[0:2]
|
||||
if dst_height * 1.0 / dst_width >= 1.5:
|
||||
dst_img = np.rot90(dst_img)
|
||||
imgs.append(dst_img)
|
||||
return imgs
|
||||
|
||||
def configure(self, **kwargs: Any) -> None:
|
||||
self.min_score = kwargs.get("minScore", self.min_score)
|
||||
28
machine-learning/immich_ml/models/ocr/schemas.py
Normal file
28
machine-learning/immich_ml/models/ocr/schemas.py
Normal file
@ -0,0 +1,28 @@
|
||||
from typing import Any, Iterable
|
||||
|
||||
import numpy as np
|
||||
import numpy.typing as npt
|
||||
from rapidocr.utils.typings import EngineType, LangRec
|
||||
from typing_extensions import TypedDict
|
||||
|
||||
|
||||
class TextDetectionOutput(TypedDict):
|
||||
image: npt.NDArray[np.float32]
|
||||
boxes: npt.NDArray[np.float32]
|
||||
scores: npt.NDArray[np.float32]
|
||||
|
||||
|
||||
class TextRecognitionOutput(TypedDict):
|
||||
box: npt.NDArray[np.float32]
|
||||
boxScore: npt.NDArray[np.float32]
|
||||
text: Iterable[str]
|
||||
textScore: npt.NDArray[np.float32]
|
||||
|
||||
|
||||
# RapidOCR expects `engine_type`, `lang_type`, and `font_path` to be attributes
|
||||
class OcrOptions(dict[str, Any]):
|
||||
def __init__(self, **options: Any) -> None:
|
||||
super().__init__(**options)
|
||||
self.engine_type = EngineType.ONNXRUNTIME
|
||||
self.lang_type = LangRec.CH
|
||||
self.font_path = None
|
||||
@ -23,6 +23,7 @@ class BoundingBox(TypedDict):
|
||||
class ModelTask(StrEnum):
|
||||
FACIAL_RECOGNITION = "facial-recognition"
|
||||
SEARCH = "clip"
|
||||
OCR = "ocr"
|
||||
|
||||
|
||||
class ModelType(StrEnum):
|
||||
@ -42,6 +43,7 @@ class ModelSource(StrEnum):
|
||||
INSIGHTFACE = "insightface"
|
||||
MCLIP = "mclip"
|
||||
OPENCLIP = "openclip"
|
||||
PADDLE = "paddle"
|
||||
|
||||
|
||||
ModelIdentity = tuple[ModelType, ModelTask]
|
||||
|
||||
@ -14,6 +14,8 @@ from ..config import log, settings
|
||||
|
||||
|
||||
class OrtSession:
|
||||
session: ort.InferenceSession
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
model_path: Path | str,
|
||||
|
||||
@ -22,6 +22,8 @@ dependencies = [
|
||||
"rich>=13.4.2",
|
||||
"tokenizers>=0.15.0,<1.0",
|
||||
"uvicorn[standard]>=0.22.0,<1.0",
|
||||
"setuptools>=78.1.0",
|
||||
"rapidocr>=3.1.0",
|
||||
]
|
||||
|
||||
[dependency-groups]
|
||||
|
||||
3470
machine-learning/uv.lock
generated
3470
machine-learning/uv.lock
generated
File diff suppressed because it is too large
Load Diff
@ -1,6 +1,6 @@
|
||||
enum SortOrder { asc, desc }
|
||||
|
||||
enum TextSearchType { context, filename, description }
|
||||
enum TextSearchType { context, filename, description, ocr }
|
||||
|
||||
enum AssetVisibilityEnum { timeline, hidden, archive, locked }
|
||||
|
||||
|
||||
@ -43,6 +43,7 @@ class SearchApiRepository extends ApiRepository {
|
||||
originalFileName: filter.filename != null && filter.filename!.isNotEmpty ? filter.filename : null,
|
||||
country: filter.location.country,
|
||||
description: filter.description != null && filter.description!.isNotEmpty ? filter.description : null,
|
||||
ocr: filter.ocr != null && filter.ocr!.isNotEmpty ? filter.ocr : null,
|
||||
state: filter.location.state,
|
||||
city: filter.location.city,
|
||||
make: filter.camera.make,
|
||||
|
||||
@ -176,6 +176,7 @@ class SearchFilter {
|
||||
String? context;
|
||||
String? filename;
|
||||
String? description;
|
||||
String? ocr;
|
||||
String? language;
|
||||
Set<PersonDto> people;
|
||||
SearchLocationFilter location;
|
||||
@ -190,6 +191,7 @@ class SearchFilter {
|
||||
this.context,
|
||||
this.filename,
|
||||
this.description,
|
||||
this.ocr,
|
||||
this.language,
|
||||
required this.people,
|
||||
required this.location,
|
||||
@ -203,6 +205,7 @@ class SearchFilter {
|
||||
return (context == null || (context != null && context!.isEmpty)) &&
|
||||
(filename == null || (filename!.isEmpty)) &&
|
||||
(description == null || (description!.isEmpty)) &&
|
||||
(ocr == null || (ocr!.isEmpty)) &&
|
||||
people.isEmpty &&
|
||||
location.country == null &&
|
||||
location.state == null &&
|
||||
@ -222,6 +225,7 @@ class SearchFilter {
|
||||
String? filename,
|
||||
String? description,
|
||||
String? language,
|
||||
String? ocr,
|
||||
Set<PersonDto>? people,
|
||||
SearchLocationFilter? location,
|
||||
SearchCameraFilter? camera,
|
||||
@ -234,6 +238,7 @@ class SearchFilter {
|
||||
filename: filename ?? this.filename,
|
||||
description: description ?? this.description,
|
||||
language: language ?? this.language,
|
||||
ocr: ocr ?? this.ocr,
|
||||
people: people ?? this.people,
|
||||
location: location ?? this.location,
|
||||
camera: camera ?? this.camera,
|
||||
@ -245,7 +250,7 @@ class SearchFilter {
|
||||
|
||||
@override
|
||||
String toString() {
|
||||
return 'SearchFilter(context: $context, filename: $filename, description: $description, language: $language, people: $people, location: $location, camera: $camera, date: $date, display: $display, mediaType: $mediaType)';
|
||||
return 'SearchFilter(context: $context, filename: $filename, description: $description, language: $language, ocr: $ocr, people: $people, location: $location, camera: $camera, date: $date, display: $display, mediaType: $mediaType)';
|
||||
}
|
||||
|
||||
@override
|
||||
@ -256,6 +261,7 @@ class SearchFilter {
|
||||
other.filename == filename &&
|
||||
other.description == description &&
|
||||
other.language == language &&
|
||||
other.ocr == ocr &&
|
||||
other.people == people &&
|
||||
other.location == location &&
|
||||
other.camera == camera &&
|
||||
@ -270,6 +276,7 @@ class SearchFilter {
|
||||
filename.hashCode ^
|
||||
description.hashCode ^
|
||||
language.hashCode ^
|
||||
ocr.hashCode ^
|
||||
people.hashCode ^
|
||||
location.hashCode ^
|
||||
camera.hashCode ^
|
||||
|
||||
@ -389,15 +389,18 @@ class SearchPage extends HookConsumerWidget {
|
||||
handleTextSubmitted(String value) {
|
||||
switch (textSearchType.value) {
|
||||
case TextSearchType.context:
|
||||
filter.value = filter.value.copyWith(filename: '', context: value, description: '');
|
||||
filter.value = filter.value.copyWith(filename: '', context: value, description: '', ocr: '');
|
||||
|
||||
break;
|
||||
case TextSearchType.filename:
|
||||
filter.value = filter.value.copyWith(filename: value, context: '', description: '');
|
||||
filter.value = filter.value.copyWith(filename: value, context: '', description: '', ocr: '');
|
||||
|
||||
break;
|
||||
case TextSearchType.description:
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: value);
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: value, ocr: '');
|
||||
break;
|
||||
case TextSearchType.ocr:
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: '', ocr: value);
|
||||
break;
|
||||
}
|
||||
|
||||
@ -408,6 +411,7 @@ class SearchPage extends HookConsumerWidget {
|
||||
TextSearchType.context => Icons.image_search_rounded,
|
||||
TextSearchType.filename => Icons.abc_rounded,
|
||||
TextSearchType.description => Icons.text_snippet_outlined,
|
||||
TextSearchType.ocr => Icons.document_scanner_outlined,
|
||||
};
|
||||
|
||||
return Scaffold(
|
||||
@ -493,6 +497,24 @@ class SearchPage extends HookConsumerWidget {
|
||||
searchHintText.value = 'search_by_description_example'.tr();
|
||||
},
|
||||
),
|
||||
MenuItemButton(
|
||||
child: ListTile(
|
||||
leading: const Icon(Icons.document_scanner_outlined),
|
||||
title: Text(
|
||||
'search_filter_ocr'.tr(),
|
||||
style: context.textTheme.bodyLarge?.copyWith(
|
||||
fontWeight: FontWeight.w500,
|
||||
color: textSearchType.value == TextSearchType.ocr ? context.colorScheme.primary : null,
|
||||
),
|
||||
),
|
||||
selectedColor: context.colorScheme.primary,
|
||||
selected: textSearchType.value == TextSearchType.ocr,
|
||||
),
|
||||
onPressed: () {
|
||||
textSearchType.value = TextSearchType.ocr;
|
||||
searchHintText.value = 'search_by_ocr_example'.tr();
|
||||
},
|
||||
),
|
||||
],
|
||||
),
|
||||
),
|
||||
|
||||
@ -395,15 +395,18 @@ class DriftSearchPage extends HookConsumerWidget {
|
||||
handleTextSubmitted(String value) {
|
||||
switch (textSearchType.value) {
|
||||
case TextSearchType.context:
|
||||
filter.value = filter.value.copyWith(filename: '', context: value, description: '');
|
||||
filter.value = filter.value.copyWith(filename: '', context: value, description: '', ocr: '');
|
||||
|
||||
break;
|
||||
case TextSearchType.filename:
|
||||
filter.value = filter.value.copyWith(filename: value, context: '', description: '');
|
||||
filter.value = filter.value.copyWith(filename: value, context: '', description: '', ocr: '');
|
||||
|
||||
break;
|
||||
case TextSearchType.description:
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: value);
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: value, ocr: '');
|
||||
break;
|
||||
case TextSearchType.ocr:
|
||||
filter.value = filter.value.copyWith(filename: '', context: '', description: '', ocr: value);
|
||||
break;
|
||||
}
|
||||
|
||||
@ -414,6 +417,7 @@ class DriftSearchPage extends HookConsumerWidget {
|
||||
TextSearchType.context => Icons.image_search_rounded,
|
||||
TextSearchType.filename => Icons.abc_rounded,
|
||||
TextSearchType.description => Icons.text_snippet_outlined,
|
||||
TextSearchType.ocr => Icons.document_scanner_outlined,
|
||||
};
|
||||
|
||||
return Scaffold(
|
||||
@ -499,6 +503,24 @@ class DriftSearchPage extends HookConsumerWidget {
|
||||
searchHintText.value = 'search_by_description_example'.t(context: context);
|
||||
},
|
||||
),
|
||||
MenuItemButton(
|
||||
child: ListTile(
|
||||
leading: const Icon(Icons.document_scanner_outlined),
|
||||
title: Text(
|
||||
'search_by_ocr'.t(context: context),
|
||||
style: context.textTheme.bodyLarge?.copyWith(
|
||||
fontWeight: FontWeight.w500,
|
||||
color: textSearchType.value == TextSearchType.ocr ? context.colorScheme.primary : null,
|
||||
),
|
||||
),
|
||||
selectedColor: context.colorScheme.primary,
|
||||
selected: textSearchType.value == TextSearchType.ocr,
|
||||
),
|
||||
onPressed: () {
|
||||
textSearchType.value = TextSearchType.ocr;
|
||||
searchHintText.value = 'search_by_ocr_example'.t(context: context);
|
||||
},
|
||||
),
|
||||
],
|
||||
),
|
||||
),
|
||||
|
||||
1
mobile/openapi/README.md
generated
1
mobile/openapi/README.md
generated
@ -419,6 +419,7 @@ Class | Method | HTTP request | Description
|
||||
- [OAuthCallbackDto](doc//OAuthCallbackDto.md)
|
||||
- [OAuthConfigDto](doc//OAuthConfigDto.md)
|
||||
- [OAuthTokenEndpointAuthMethod](doc//OAuthTokenEndpointAuthMethod.md)
|
||||
- [OcrConfig](doc//OcrConfig.md)
|
||||
- [OnThisDayDto](doc//OnThisDayDto.md)
|
||||
- [OnboardingDto](doc//OnboardingDto.md)
|
||||
- [OnboardingResponseDto](doc//OnboardingResponseDto.md)
|
||||
|
||||
1
mobile/openapi/lib/api.dart
generated
1
mobile/openapi/lib/api.dart
generated
@ -189,6 +189,7 @@ part 'model/o_auth_authorize_response_dto.dart';
|
||||
part 'model/o_auth_callback_dto.dart';
|
||||
part 'model/o_auth_config_dto.dart';
|
||||
part 'model/o_auth_token_endpoint_auth_method.dart';
|
||||
part 'model/ocr_config.dart';
|
||||
part 'model/on_this_day_dto.dart';
|
||||
part 'model/onboarding_dto.dart';
|
||||
part 'model/onboarding_response_dto.dart';
|
||||
|
||||
13
mobile/openapi/lib/api/search_api.dart
generated
13
mobile/openapi/lib/api/search_api.dart
generated
@ -353,6 +353,8 @@ class SearchApi {
|
||||
///
|
||||
/// * [String] model:
|
||||
///
|
||||
/// * [String] ocr:
|
||||
///
|
||||
/// * [List<String>] personIds:
|
||||
///
|
||||
/// * [num] rating:
|
||||
@ -382,7 +384,7 @@ class SearchApi {
|
||||
/// * [bool] withDeleted:
|
||||
///
|
||||
/// * [bool] withExif:
|
||||
Future<Response> searchLargeAssetsWithHttpInfo({ List<String>? albumIds, String? city, String? country, DateTime? createdAfter, DateTime? createdBefore, String? deviceId, bool? isEncoded, bool? isFavorite, bool? isMotion, bool? isNotInAlbum, bool? isOffline, String? lensModel, String? libraryId, String? make, int? minFileSize, String? model, List<String>? personIds, num? rating, num? size, String? state, List<String>? tagIds, DateTime? takenAfter, DateTime? takenBefore, DateTime? trashedAfter, DateTime? trashedBefore, AssetTypeEnum? type, DateTime? updatedAfter, DateTime? updatedBefore, AssetVisibility? visibility, bool? withDeleted, bool? withExif, }) async {
|
||||
Future<Response> searchLargeAssetsWithHttpInfo({ List<String>? albumIds, String? city, String? country, DateTime? createdAfter, DateTime? createdBefore, String? deviceId, bool? isEncoded, bool? isFavorite, bool? isMotion, bool? isNotInAlbum, bool? isOffline, String? lensModel, String? libraryId, String? make, int? minFileSize, String? model, String? ocr, List<String>? personIds, num? rating, num? size, String? state, List<String>? tagIds, DateTime? takenAfter, DateTime? takenBefore, DateTime? trashedAfter, DateTime? trashedBefore, AssetTypeEnum? type, DateTime? updatedAfter, DateTime? updatedBefore, AssetVisibility? visibility, bool? withDeleted, bool? withExif, }) async {
|
||||
// ignore: prefer_const_declarations
|
||||
final apiPath = r'/search/large-assets';
|
||||
|
||||
@ -441,6 +443,9 @@ class SearchApi {
|
||||
if (model != null) {
|
||||
queryParams.addAll(_queryParams('', 'model', model));
|
||||
}
|
||||
if (ocr != null) {
|
||||
queryParams.addAll(_queryParams('', 'ocr', ocr));
|
||||
}
|
||||
if (personIds != null) {
|
||||
queryParams.addAll(_queryParams('multi', 'personIds', personIds));
|
||||
}
|
||||
@ -537,6 +542,8 @@ class SearchApi {
|
||||
///
|
||||
/// * [String] model:
|
||||
///
|
||||
/// * [String] ocr:
|
||||
///
|
||||
/// * [List<String>] personIds:
|
||||
///
|
||||
/// * [num] rating:
|
||||
@ -566,8 +573,8 @@ class SearchApi {
|
||||
/// * [bool] withDeleted:
|
||||
///
|
||||
/// * [bool] withExif:
|
||||
Future<List<AssetResponseDto>?> searchLargeAssets({ List<String>? albumIds, String? city, String? country, DateTime? createdAfter, DateTime? createdBefore, String? deviceId, bool? isEncoded, bool? isFavorite, bool? isMotion, bool? isNotInAlbum, bool? isOffline, String? lensModel, String? libraryId, String? make, int? minFileSize, String? model, List<String>? personIds, num? rating, num? size, String? state, List<String>? tagIds, DateTime? takenAfter, DateTime? takenBefore, DateTime? trashedAfter, DateTime? trashedBefore, AssetTypeEnum? type, DateTime? updatedAfter, DateTime? updatedBefore, AssetVisibility? visibility, bool? withDeleted, bool? withExif, }) async {
|
||||
final response = await searchLargeAssetsWithHttpInfo( albumIds: albumIds, city: city, country: country, createdAfter: createdAfter, createdBefore: createdBefore, deviceId: deviceId, isEncoded: isEncoded, isFavorite: isFavorite, isMotion: isMotion, isNotInAlbum: isNotInAlbum, isOffline: isOffline, lensModel: lensModel, libraryId: libraryId, make: make, minFileSize: minFileSize, model: model, personIds: personIds, rating: rating, size: size, state: state, tagIds: tagIds, takenAfter: takenAfter, takenBefore: takenBefore, trashedAfter: trashedAfter, trashedBefore: trashedBefore, type: type, updatedAfter: updatedAfter, updatedBefore: updatedBefore, visibility: visibility, withDeleted: withDeleted, withExif: withExif, );
|
||||
Future<List<AssetResponseDto>?> searchLargeAssets({ List<String>? albumIds, String? city, String? country, DateTime? createdAfter, DateTime? createdBefore, String? deviceId, bool? isEncoded, bool? isFavorite, bool? isMotion, bool? isNotInAlbum, bool? isOffline, String? lensModel, String? libraryId, String? make, int? minFileSize, String? model, String? ocr, List<String>? personIds, num? rating, num? size, String? state, List<String>? tagIds, DateTime? takenAfter, DateTime? takenBefore, DateTime? trashedAfter, DateTime? trashedBefore, AssetTypeEnum? type, DateTime? updatedAfter, DateTime? updatedBefore, AssetVisibility? visibility, bool? withDeleted, bool? withExif, }) async {
|
||||
final response = await searchLargeAssetsWithHttpInfo( albumIds: albumIds, city: city, country: country, createdAfter: createdAfter, createdBefore: createdBefore, deviceId: deviceId, isEncoded: isEncoded, isFavorite: isFavorite, isMotion: isMotion, isNotInAlbum: isNotInAlbum, isOffline: isOffline, lensModel: lensModel, libraryId: libraryId, make: make, minFileSize: minFileSize, model: model, ocr: ocr, personIds: personIds, rating: rating, size: size, state: state, tagIds: tagIds, takenAfter: takenAfter, takenBefore: takenBefore, trashedAfter: trashedAfter, trashedBefore: trashedBefore, type: type, updatedAfter: updatedAfter, updatedBefore: updatedBefore, visibility: visibility, withDeleted: withDeleted, withExif: withExif, );
|
||||
if (response.statusCode >= HttpStatus.badRequest) {
|
||||
throw ApiException(response.statusCode, await _decodeBodyBytes(response));
|
||||
}
|
||||
|
||||
2
mobile/openapi/lib/api_client.dart
generated
2
mobile/openapi/lib/api_client.dart
generated
@ -432,6 +432,8 @@ class ApiClient {
|
||||
return OAuthConfigDto.fromJson(value);
|
||||
case 'OAuthTokenEndpointAuthMethod':
|
||||
return OAuthTokenEndpointAuthMethodTypeTransformer().decode(value);
|
||||
case 'OcrConfig':
|
||||
return OcrConfig.fromJson(value);
|
||||
case 'OnThisDayDto':
|
||||
return OnThisDayDto.fromJson(value);
|
||||
case 'OnboardingDto':
|
||||
|
||||
@ -22,6 +22,7 @@ class AllJobStatusResponseDto {
|
||||
required this.metadataExtraction,
|
||||
required this.migration,
|
||||
required this.notifications,
|
||||
required this.ocr,
|
||||
required this.search,
|
||||
required this.sidecar,
|
||||
required this.smartSearch,
|
||||
@ -48,6 +49,8 @@ class AllJobStatusResponseDto {
|
||||
|
||||
JobStatusDto notifications;
|
||||
|
||||
JobStatusDto ocr;
|
||||
|
||||
JobStatusDto search;
|
||||
|
||||
JobStatusDto sidecar;
|
||||
@ -71,6 +74,7 @@ class AllJobStatusResponseDto {
|
||||
other.metadataExtraction == metadataExtraction &&
|
||||
other.migration == migration &&
|
||||
other.notifications == notifications &&
|
||||
other.ocr == ocr &&
|
||||
other.search == search &&
|
||||
other.sidecar == sidecar &&
|
||||
other.smartSearch == smartSearch &&
|
||||
@ -90,6 +94,7 @@ class AllJobStatusResponseDto {
|
||||
(metadataExtraction.hashCode) +
|
||||
(migration.hashCode) +
|
||||
(notifications.hashCode) +
|
||||
(ocr.hashCode) +
|
||||
(search.hashCode) +
|
||||
(sidecar.hashCode) +
|
||||
(smartSearch.hashCode) +
|
||||
@ -98,7 +103,7 @@ class AllJobStatusResponseDto {
|
||||
(videoConversion.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'AllJobStatusResponseDto[backgroundTask=$backgroundTask, backupDatabase=$backupDatabase, duplicateDetection=$duplicateDetection, faceDetection=$faceDetection, facialRecognition=$facialRecognition, library_=$library_, metadataExtraction=$metadataExtraction, migration=$migration, notifications=$notifications, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, storageTemplateMigration=$storageTemplateMigration, thumbnailGeneration=$thumbnailGeneration, videoConversion=$videoConversion]';
|
||||
String toString() => 'AllJobStatusResponseDto[backgroundTask=$backgroundTask, backupDatabase=$backupDatabase, duplicateDetection=$duplicateDetection, faceDetection=$faceDetection, facialRecognition=$facialRecognition, library_=$library_, metadataExtraction=$metadataExtraction, migration=$migration, notifications=$notifications, ocr=$ocr, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, storageTemplateMigration=$storageTemplateMigration, thumbnailGeneration=$thumbnailGeneration, videoConversion=$videoConversion]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -111,6 +116,7 @@ class AllJobStatusResponseDto {
|
||||
json[r'metadataExtraction'] = this.metadataExtraction;
|
||||
json[r'migration'] = this.migration;
|
||||
json[r'notifications'] = this.notifications;
|
||||
json[r'ocr'] = this.ocr;
|
||||
json[r'search'] = this.search;
|
||||
json[r'sidecar'] = this.sidecar;
|
||||
json[r'smartSearch'] = this.smartSearch;
|
||||
@ -138,6 +144,7 @@ class AllJobStatusResponseDto {
|
||||
metadataExtraction: JobStatusDto.fromJson(json[r'metadataExtraction'])!,
|
||||
migration: JobStatusDto.fromJson(json[r'migration'])!,
|
||||
notifications: JobStatusDto.fromJson(json[r'notifications'])!,
|
||||
ocr: JobStatusDto.fromJson(json[r'ocr'])!,
|
||||
search: JobStatusDto.fromJson(json[r'search'])!,
|
||||
sidecar: JobStatusDto.fromJson(json[r'sidecar'])!,
|
||||
smartSearch: JobStatusDto.fromJson(json[r'smartSearch'])!,
|
||||
@ -200,6 +207,7 @@ class AllJobStatusResponseDto {
|
||||
'metadataExtraction',
|
||||
'migration',
|
||||
'notifications',
|
||||
'ocr',
|
||||
'search',
|
||||
'sidecar',
|
||||
'smartSearch',
|
||||
|
||||
3
mobile/openapi/lib/model/job_name.dart
generated
3
mobile/openapi/lib/model/job_name.dart
generated
@ -38,6 +38,7 @@ class JobName {
|
||||
static const library_ = JobName._(r'library');
|
||||
static const notifications = JobName._(r'notifications');
|
||||
static const backupDatabase = JobName._(r'backupDatabase');
|
||||
static const ocr = JobName._(r'ocr');
|
||||
|
||||
/// List of all possible values in this [enum][JobName].
|
||||
static const values = <JobName>[
|
||||
@ -56,6 +57,7 @@ class JobName {
|
||||
library_,
|
||||
notifications,
|
||||
backupDatabase,
|
||||
ocr,
|
||||
];
|
||||
|
||||
static JobName? fromJson(dynamic value) => JobNameTypeTransformer().decode(value);
|
||||
@ -109,6 +111,7 @@ class JobNameTypeTransformer {
|
||||
case r'library': return JobName.library_;
|
||||
case r'notifications': return JobName.notifications;
|
||||
case r'backupDatabase': return JobName.backupDatabase;
|
||||
case r'ocr': return JobName.ocr;
|
||||
default:
|
||||
if (!allowNull) {
|
||||
throw ArgumentError('Unknown enum value to decode: $data');
|
||||
|
||||
19
mobile/openapi/lib/model/metadata_search_dto.dart
generated
19
mobile/openapi/lib/model/metadata_search_dto.dart
generated
@ -33,6 +33,7 @@ class MetadataSearchDto {
|
||||
this.libraryId,
|
||||
this.make,
|
||||
this.model,
|
||||
this.ocr,
|
||||
this.order = AssetOrder.desc,
|
||||
this.originalFileName,
|
||||
this.originalPath,
|
||||
@ -182,6 +183,14 @@ class MetadataSearchDto {
|
||||
|
||||
String? model;
|
||||
|
||||
///
|
||||
/// Please note: This property should have been non-nullable! Since the specification file
|
||||
/// does not include a default value (using the "default:" property), however, the generated
|
||||
/// source code must fall back to having a nullable type.
|
||||
/// Consider adding a "default:" property in the specification file to hide this note.
|
||||
///
|
||||
String? ocr;
|
||||
|
||||
AssetOrder order;
|
||||
|
||||
///
|
||||
@ -369,6 +378,7 @@ class MetadataSearchDto {
|
||||
other.libraryId == libraryId &&
|
||||
other.make == make &&
|
||||
other.model == model &&
|
||||
other.ocr == ocr &&
|
||||
other.order == order &&
|
||||
other.originalFileName == originalFileName &&
|
||||
other.originalPath == originalPath &&
|
||||
@ -416,6 +426,7 @@ class MetadataSearchDto {
|
||||
(libraryId == null ? 0 : libraryId!.hashCode) +
|
||||
(make == null ? 0 : make!.hashCode) +
|
||||
(model == null ? 0 : model!.hashCode) +
|
||||
(ocr == null ? 0 : ocr!.hashCode) +
|
||||
(order.hashCode) +
|
||||
(originalFileName == null ? 0 : originalFileName!.hashCode) +
|
||||
(originalPath == null ? 0 : originalPath!.hashCode) +
|
||||
@ -441,7 +452,7 @@ class MetadataSearchDto {
|
||||
(withStacked == null ? 0 : withStacked!.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'MetadataSearchDto[albumIds=$albumIds, checksum=$checksum, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, description=$description, deviceAssetId=$deviceAssetId, deviceId=$deviceId, encodedVideoPath=$encodedVideoPath, id=$id, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, order=$order, originalFileName=$originalFileName, originalPath=$originalPath, page=$page, personIds=$personIds, previewPath=$previewPath, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, thumbnailPath=$thumbnailPath, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif, withPeople=$withPeople, withStacked=$withStacked]';
|
||||
String toString() => 'MetadataSearchDto[albumIds=$albumIds, checksum=$checksum, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, description=$description, deviceAssetId=$deviceAssetId, deviceId=$deviceId, encodedVideoPath=$encodedVideoPath, id=$id, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, ocr=$ocr, order=$order, originalFileName=$originalFileName, originalPath=$originalPath, page=$page, personIds=$personIds, previewPath=$previewPath, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, thumbnailPath=$thumbnailPath, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif, withPeople=$withPeople, withStacked=$withStacked]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -540,6 +551,11 @@ class MetadataSearchDto {
|
||||
json[r'model'] = this.model;
|
||||
} else {
|
||||
// json[r'model'] = null;
|
||||
}
|
||||
if (this.ocr != null) {
|
||||
json[r'ocr'] = this.ocr;
|
||||
} else {
|
||||
// json[r'ocr'] = null;
|
||||
}
|
||||
json[r'order'] = this.order;
|
||||
if (this.originalFileName != null) {
|
||||
@ -682,6 +698,7 @@ class MetadataSearchDto {
|
||||
libraryId: mapValueOfType<String>(json, r'libraryId'),
|
||||
make: mapValueOfType<String>(json, r'make'),
|
||||
model: mapValueOfType<String>(json, r'model'),
|
||||
ocr: mapValueOfType<String>(json, r'ocr'),
|
||||
order: AssetOrder.fromJson(json[r'order']) ?? AssetOrder.desc,
|
||||
originalFileName: mapValueOfType<String>(json, r'originalFileName'),
|
||||
originalPath: mapValueOfType<String>(json, r'originalPath'),
|
||||
|
||||
136
mobile/openapi/lib/model/ocr_config.dart
generated
Normal file
136
mobile/openapi/lib/model/ocr_config.dart
generated
Normal file
@ -0,0 +1,136 @@
|
||||
//
|
||||
// AUTO-GENERATED FILE, DO NOT MODIFY!
|
||||
//
|
||||
// @dart=2.18
|
||||
|
||||
// ignore_for_file: unused_element, unused_import
|
||||
// ignore_for_file: always_put_required_named_parameters_first
|
||||
// ignore_for_file: constant_identifier_names
|
||||
// ignore_for_file: lines_longer_than_80_chars
|
||||
|
||||
part of openapi.api;
|
||||
|
||||
class OcrConfig {
|
||||
/// Returns a new [OcrConfig] instance.
|
||||
OcrConfig({
|
||||
required this.enabled,
|
||||
required this.maxResolution,
|
||||
required this.minDetectionScore,
|
||||
required this.minRecognitionScore,
|
||||
required this.modelName,
|
||||
});
|
||||
|
||||
bool enabled;
|
||||
|
||||
/// Minimum value: 1
|
||||
int maxResolution;
|
||||
|
||||
/// Minimum value: 0.1
|
||||
/// Maximum value: 1
|
||||
double minDetectionScore;
|
||||
|
||||
/// Minimum value: 0.1
|
||||
/// Maximum value: 1
|
||||
double minRecognitionScore;
|
||||
|
||||
String modelName;
|
||||
|
||||
@override
|
||||
bool operator ==(Object other) => identical(this, other) || other is OcrConfig &&
|
||||
other.enabled == enabled &&
|
||||
other.maxResolution == maxResolution &&
|
||||
other.minDetectionScore == minDetectionScore &&
|
||||
other.minRecognitionScore == minRecognitionScore &&
|
||||
other.modelName == modelName;
|
||||
|
||||
@override
|
||||
int get hashCode =>
|
||||
// ignore: unnecessary_parenthesis
|
||||
(enabled.hashCode) +
|
||||
(maxResolution.hashCode) +
|
||||
(minDetectionScore.hashCode) +
|
||||
(minRecognitionScore.hashCode) +
|
||||
(modelName.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'OcrConfig[enabled=$enabled, maxResolution=$maxResolution, minDetectionScore=$minDetectionScore, minRecognitionScore=$minRecognitionScore, modelName=$modelName]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
json[r'enabled'] = this.enabled;
|
||||
json[r'maxResolution'] = this.maxResolution;
|
||||
json[r'minDetectionScore'] = this.minDetectionScore;
|
||||
json[r'minRecognitionScore'] = this.minRecognitionScore;
|
||||
json[r'modelName'] = this.modelName;
|
||||
return json;
|
||||
}
|
||||
|
||||
/// Returns a new [OcrConfig] instance and imports its values from
|
||||
/// [value] if it's a [Map], null otherwise.
|
||||
// ignore: prefer_constructors_over_static_methods
|
||||
static OcrConfig? fromJson(dynamic value) {
|
||||
upgradeDto(value, "OcrConfig");
|
||||
if (value is Map) {
|
||||
final json = value.cast<String, dynamic>();
|
||||
|
||||
return OcrConfig(
|
||||
enabled: mapValueOfType<bool>(json, r'enabled')!,
|
||||
maxResolution: mapValueOfType<int>(json, r'maxResolution')!,
|
||||
minDetectionScore: (mapValueOfType<num>(json, r'minDetectionScore')!).toDouble(),
|
||||
minRecognitionScore: (mapValueOfType<num>(json, r'minRecognitionScore')!).toDouble(),
|
||||
modelName: mapValueOfType<String>(json, r'modelName')!,
|
||||
);
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
static List<OcrConfig> listFromJson(dynamic json, {bool growable = false,}) {
|
||||
final result = <OcrConfig>[];
|
||||
if (json is List && json.isNotEmpty) {
|
||||
for (final row in json) {
|
||||
final value = OcrConfig.fromJson(row);
|
||||
if (value != null) {
|
||||
result.add(value);
|
||||
}
|
||||
}
|
||||
}
|
||||
return result.toList(growable: growable);
|
||||
}
|
||||
|
||||
static Map<String, OcrConfig> mapFromJson(dynamic json) {
|
||||
final map = <String, OcrConfig>{};
|
||||
if (json is Map && json.isNotEmpty) {
|
||||
json = json.cast<String, dynamic>(); // ignore: parameter_assignments
|
||||
for (final entry in json.entries) {
|
||||
final value = OcrConfig.fromJson(entry.value);
|
||||
if (value != null) {
|
||||
map[entry.key] = value;
|
||||
}
|
||||
}
|
||||
}
|
||||
return map;
|
||||
}
|
||||
|
||||
// maps a json object with a list of OcrConfig-objects as value to a dart map
|
||||
static Map<String, List<OcrConfig>> mapListFromJson(dynamic json, {bool growable = false,}) {
|
||||
final map = <String, List<OcrConfig>>{};
|
||||
if (json is Map && json.isNotEmpty) {
|
||||
// ignore: parameter_assignments
|
||||
json = json.cast<String, dynamic>();
|
||||
for (final entry in json.entries) {
|
||||
map[entry.key] = OcrConfig.listFromJson(entry.value, growable: growable,);
|
||||
}
|
||||
}
|
||||
return map;
|
||||
}
|
||||
|
||||
/// The list of required keys that must be present in a JSON.
|
||||
static const requiredKeys = <String>{
|
||||
'enabled',
|
||||
'maxResolution',
|
||||
'minDetectionScore',
|
||||
'minRecognitionScore',
|
||||
'modelName',
|
||||
};
|
||||
}
|
||||
|
||||
19
mobile/openapi/lib/model/random_search_dto.dart
generated
19
mobile/openapi/lib/model/random_search_dto.dart
generated
@ -28,6 +28,7 @@ class RandomSearchDto {
|
||||
this.libraryId,
|
||||
this.make,
|
||||
this.model,
|
||||
this.ocr,
|
||||
this.personIds = const [],
|
||||
this.rating,
|
||||
this.size,
|
||||
@ -131,6 +132,14 @@ class RandomSearchDto {
|
||||
|
||||
String? model;
|
||||
|
||||
///
|
||||
/// Please note: This property should have been non-nullable! Since the specification file
|
||||
/// does not include a default value (using the "default:" property), however, the generated
|
||||
/// source code must fall back to having a nullable type.
|
||||
/// Consider adding a "default:" property in the specification file to hide this note.
|
||||
///
|
||||
String? ocr;
|
||||
|
||||
List<String> personIds;
|
||||
|
||||
/// Minimum value: -1
|
||||
@ -270,6 +279,7 @@ class RandomSearchDto {
|
||||
other.libraryId == libraryId &&
|
||||
other.make == make &&
|
||||
other.model == model &&
|
||||
other.ocr == ocr &&
|
||||
_deepEquality.equals(other.personIds, personIds) &&
|
||||
other.rating == rating &&
|
||||
other.size == size &&
|
||||
@ -306,6 +316,7 @@ class RandomSearchDto {
|
||||
(libraryId == null ? 0 : libraryId!.hashCode) +
|
||||
(make == null ? 0 : make!.hashCode) +
|
||||
(model == null ? 0 : model!.hashCode) +
|
||||
(ocr == null ? 0 : ocr!.hashCode) +
|
||||
(personIds.hashCode) +
|
||||
(rating == null ? 0 : rating!.hashCode) +
|
||||
(size == null ? 0 : size!.hashCode) +
|
||||
@ -325,7 +336,7 @@ class RandomSearchDto {
|
||||
(withStacked == null ? 0 : withStacked!.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'RandomSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, personIds=$personIds, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif, withPeople=$withPeople, withStacked=$withStacked]';
|
||||
String toString() => 'RandomSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, ocr=$ocr, personIds=$personIds, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif, withPeople=$withPeople, withStacked=$withStacked]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -399,6 +410,11 @@ class RandomSearchDto {
|
||||
json[r'model'] = this.model;
|
||||
} else {
|
||||
// json[r'model'] = null;
|
||||
}
|
||||
if (this.ocr != null) {
|
||||
json[r'ocr'] = this.ocr;
|
||||
} else {
|
||||
// json[r'ocr'] = null;
|
||||
}
|
||||
json[r'personIds'] = this.personIds;
|
||||
if (this.rating != null) {
|
||||
@ -510,6 +526,7 @@ class RandomSearchDto {
|
||||
libraryId: mapValueOfType<String>(json, r'libraryId'),
|
||||
make: mapValueOfType<String>(json, r'make'),
|
||||
model: mapValueOfType<String>(json, r'model'),
|
||||
ocr: mapValueOfType<String>(json, r'ocr'),
|
||||
personIds: json[r'personIds'] is Iterable
|
||||
? (json[r'personIds'] as Iterable).cast<String>().toList(growable: false)
|
||||
: const [],
|
||||
|
||||
10
mobile/openapi/lib/model/server_features_dto.dart
generated
10
mobile/openapi/lib/model/server_features_dto.dart
generated
@ -21,6 +21,7 @@ class ServerFeaturesDto {
|
||||
required this.map,
|
||||
required this.oauth,
|
||||
required this.oauthAutoLaunch,
|
||||
required this.ocr,
|
||||
required this.passwordLogin,
|
||||
required this.reverseGeocoding,
|
||||
required this.search,
|
||||
@ -45,6 +46,8 @@ class ServerFeaturesDto {
|
||||
|
||||
bool oauthAutoLaunch;
|
||||
|
||||
bool ocr;
|
||||
|
||||
bool passwordLogin;
|
||||
|
||||
bool reverseGeocoding;
|
||||
@ -67,6 +70,7 @@ class ServerFeaturesDto {
|
||||
other.map == map &&
|
||||
other.oauth == oauth &&
|
||||
other.oauthAutoLaunch == oauthAutoLaunch &&
|
||||
other.ocr == ocr &&
|
||||
other.passwordLogin == passwordLogin &&
|
||||
other.reverseGeocoding == reverseGeocoding &&
|
||||
other.search == search &&
|
||||
@ -85,6 +89,7 @@ class ServerFeaturesDto {
|
||||
(map.hashCode) +
|
||||
(oauth.hashCode) +
|
||||
(oauthAutoLaunch.hashCode) +
|
||||
(ocr.hashCode) +
|
||||
(passwordLogin.hashCode) +
|
||||
(reverseGeocoding.hashCode) +
|
||||
(search.hashCode) +
|
||||
@ -93,7 +98,7 @@ class ServerFeaturesDto {
|
||||
(trash.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'ServerFeaturesDto[configFile=$configFile, duplicateDetection=$duplicateDetection, email=$email, facialRecognition=$facialRecognition, importFaces=$importFaces, map=$map, oauth=$oauth, oauthAutoLaunch=$oauthAutoLaunch, passwordLogin=$passwordLogin, reverseGeocoding=$reverseGeocoding, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, trash=$trash]';
|
||||
String toString() => 'ServerFeaturesDto[configFile=$configFile, duplicateDetection=$duplicateDetection, email=$email, facialRecognition=$facialRecognition, importFaces=$importFaces, map=$map, oauth=$oauth, oauthAutoLaunch=$oauthAutoLaunch, ocr=$ocr, passwordLogin=$passwordLogin, reverseGeocoding=$reverseGeocoding, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, trash=$trash]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -105,6 +110,7 @@ class ServerFeaturesDto {
|
||||
json[r'map'] = this.map;
|
||||
json[r'oauth'] = this.oauth;
|
||||
json[r'oauthAutoLaunch'] = this.oauthAutoLaunch;
|
||||
json[r'ocr'] = this.ocr;
|
||||
json[r'passwordLogin'] = this.passwordLogin;
|
||||
json[r'reverseGeocoding'] = this.reverseGeocoding;
|
||||
json[r'search'] = this.search;
|
||||
@ -131,6 +137,7 @@ class ServerFeaturesDto {
|
||||
map: mapValueOfType<bool>(json, r'map')!,
|
||||
oauth: mapValueOfType<bool>(json, r'oauth')!,
|
||||
oauthAutoLaunch: mapValueOfType<bool>(json, r'oauthAutoLaunch')!,
|
||||
ocr: mapValueOfType<bool>(json, r'ocr')!,
|
||||
passwordLogin: mapValueOfType<bool>(json, r'passwordLogin')!,
|
||||
reverseGeocoding: mapValueOfType<bool>(json, r'reverseGeocoding')!,
|
||||
search: mapValueOfType<bool>(json, r'search')!,
|
||||
@ -192,6 +199,7 @@ class ServerFeaturesDto {
|
||||
'map',
|
||||
'oauth',
|
||||
'oauthAutoLaunch',
|
||||
'ocr',
|
||||
'passwordLogin',
|
||||
'reverseGeocoding',
|
||||
'search',
|
||||
|
||||
19
mobile/openapi/lib/model/smart_search_dto.dart
generated
19
mobile/openapi/lib/model/smart_search_dto.dart
generated
@ -29,6 +29,7 @@ class SmartSearchDto {
|
||||
this.libraryId,
|
||||
this.make,
|
||||
this.model,
|
||||
this.ocr,
|
||||
this.page,
|
||||
this.personIds = const [],
|
||||
this.query,
|
||||
@ -141,6 +142,14 @@ class SmartSearchDto {
|
||||
|
||||
String? model;
|
||||
|
||||
///
|
||||
/// Please note: This property should have been non-nullable! Since the specification file
|
||||
/// does not include a default value (using the "default:" property), however, the generated
|
||||
/// source code must fall back to having a nullable type.
|
||||
/// Consider adding a "default:" property in the specification file to hide this note.
|
||||
///
|
||||
String? ocr;
|
||||
|
||||
/// Minimum value: 1
|
||||
///
|
||||
/// Please note: This property should have been non-nullable! Since the specification file
|
||||
@ -290,6 +299,7 @@ class SmartSearchDto {
|
||||
other.libraryId == libraryId &&
|
||||
other.make == make &&
|
||||
other.model == model &&
|
||||
other.ocr == ocr &&
|
||||
other.page == page &&
|
||||
_deepEquality.equals(other.personIds, personIds) &&
|
||||
other.query == query &&
|
||||
@ -328,6 +338,7 @@ class SmartSearchDto {
|
||||
(libraryId == null ? 0 : libraryId!.hashCode) +
|
||||
(make == null ? 0 : make!.hashCode) +
|
||||
(model == null ? 0 : model!.hashCode) +
|
||||
(ocr == null ? 0 : ocr!.hashCode) +
|
||||
(page == null ? 0 : page!.hashCode) +
|
||||
(personIds.hashCode) +
|
||||
(query == null ? 0 : query!.hashCode) +
|
||||
@ -348,7 +359,7 @@ class SmartSearchDto {
|
||||
(withExif == null ? 0 : withExif!.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'SmartSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, language=$language, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, page=$page, personIds=$personIds, query=$query, queryAssetId=$queryAssetId, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif]';
|
||||
String toString() => 'SmartSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, language=$language, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, ocr=$ocr, page=$page, personIds=$personIds, query=$query, queryAssetId=$queryAssetId, rating=$rating, size=$size, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility, withDeleted=$withDeleted, withExif=$withExif]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -428,6 +439,11 @@ class SmartSearchDto {
|
||||
} else {
|
||||
// json[r'model'] = null;
|
||||
}
|
||||
if (this.ocr != null) {
|
||||
json[r'ocr'] = this.ocr;
|
||||
} else {
|
||||
// json[r'ocr'] = null;
|
||||
}
|
||||
if (this.page != null) {
|
||||
json[r'page'] = this.page;
|
||||
} else {
|
||||
@ -544,6 +560,7 @@ class SmartSearchDto {
|
||||
libraryId: mapValueOfType<String>(json, r'libraryId'),
|
||||
make: mapValueOfType<String>(json, r'make'),
|
||||
model: mapValueOfType<String>(json, r'model'),
|
||||
ocr: mapValueOfType<String>(json, r'ocr'),
|
||||
page: num.parse('${json[r'page']}'),
|
||||
personIds: json[r'personIds'] is Iterable
|
||||
? (json[r'personIds'] as Iterable).cast<String>().toList(growable: false)
|
||||
|
||||
19
mobile/openapi/lib/model/statistics_search_dto.dart
generated
19
mobile/openapi/lib/model/statistics_search_dto.dart
generated
@ -29,6 +29,7 @@ class StatisticsSearchDto {
|
||||
this.libraryId,
|
||||
this.make,
|
||||
this.model,
|
||||
this.ocr,
|
||||
this.personIds = const [],
|
||||
this.rating,
|
||||
this.state,
|
||||
@ -135,6 +136,14 @@ class StatisticsSearchDto {
|
||||
|
||||
String? model;
|
||||
|
||||
///
|
||||
/// Please note: This property should have been non-nullable! Since the specification file
|
||||
/// does not include a default value (using the "default:" property), however, the generated
|
||||
/// source code must fall back to having a nullable type.
|
||||
/// Consider adding a "default:" property in the specification file to hide this note.
|
||||
///
|
||||
String? ocr;
|
||||
|
||||
List<String> personIds;
|
||||
|
||||
/// Minimum value: -1
|
||||
@ -233,6 +242,7 @@ class StatisticsSearchDto {
|
||||
other.libraryId == libraryId &&
|
||||
other.make == make &&
|
||||
other.model == model &&
|
||||
other.ocr == ocr &&
|
||||
_deepEquality.equals(other.personIds, personIds) &&
|
||||
other.rating == rating &&
|
||||
other.state == state &&
|
||||
@ -265,6 +275,7 @@ class StatisticsSearchDto {
|
||||
(libraryId == null ? 0 : libraryId!.hashCode) +
|
||||
(make == null ? 0 : make!.hashCode) +
|
||||
(model == null ? 0 : model!.hashCode) +
|
||||
(ocr == null ? 0 : ocr!.hashCode) +
|
||||
(personIds.hashCode) +
|
||||
(rating == null ? 0 : rating!.hashCode) +
|
||||
(state == null ? 0 : state!.hashCode) +
|
||||
@ -279,7 +290,7 @@ class StatisticsSearchDto {
|
||||
(visibility == null ? 0 : visibility!.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'StatisticsSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, description=$description, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, personIds=$personIds, rating=$rating, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility]';
|
||||
String toString() => 'StatisticsSearchDto[albumIds=$albumIds, city=$city, country=$country, createdAfter=$createdAfter, createdBefore=$createdBefore, description=$description, deviceId=$deviceId, isEncoded=$isEncoded, isFavorite=$isFavorite, isMotion=$isMotion, isNotInAlbum=$isNotInAlbum, isOffline=$isOffline, lensModel=$lensModel, libraryId=$libraryId, make=$make, model=$model, ocr=$ocr, personIds=$personIds, rating=$rating, state=$state, tagIds=$tagIds, takenAfter=$takenAfter, takenBefore=$takenBefore, trashedAfter=$trashedAfter, trashedBefore=$trashedBefore, type=$type, updatedAfter=$updatedAfter, updatedBefore=$updatedBefore, visibility=$visibility]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -358,6 +369,11 @@ class StatisticsSearchDto {
|
||||
json[r'model'] = this.model;
|
||||
} else {
|
||||
// json[r'model'] = null;
|
||||
}
|
||||
if (this.ocr != null) {
|
||||
json[r'ocr'] = this.ocr;
|
||||
} else {
|
||||
// json[r'ocr'] = null;
|
||||
}
|
||||
json[r'personIds'] = this.personIds;
|
||||
if (this.rating != null) {
|
||||
@ -445,6 +461,7 @@ class StatisticsSearchDto {
|
||||
libraryId: mapValueOfType<String>(json, r'libraryId'),
|
||||
make: mapValueOfType<String>(json, r'make'),
|
||||
model: mapValueOfType<String>(json, r'model'),
|
||||
ocr: mapValueOfType<String>(json, r'ocr'),
|
||||
personIds: json[r'personIds'] is Iterable
|
||||
? (json[r'personIds'] as Iterable).cast<String>().toList(growable: false)
|
||||
: const [],
|
||||
|
||||
10
mobile/openapi/lib/model/system_config_job_dto.dart
generated
10
mobile/openapi/lib/model/system_config_job_dto.dart
generated
@ -19,6 +19,7 @@ class SystemConfigJobDto {
|
||||
required this.metadataExtraction,
|
||||
required this.migration,
|
||||
required this.notifications,
|
||||
required this.ocr,
|
||||
required this.search,
|
||||
required this.sidecar,
|
||||
required this.smartSearch,
|
||||
@ -38,6 +39,8 @@ class SystemConfigJobDto {
|
||||
|
||||
JobSettingsDto notifications;
|
||||
|
||||
JobSettingsDto ocr;
|
||||
|
||||
JobSettingsDto search;
|
||||
|
||||
JobSettingsDto sidecar;
|
||||
@ -56,6 +59,7 @@ class SystemConfigJobDto {
|
||||
other.metadataExtraction == metadataExtraction &&
|
||||
other.migration == migration &&
|
||||
other.notifications == notifications &&
|
||||
other.ocr == ocr &&
|
||||
other.search == search &&
|
||||
other.sidecar == sidecar &&
|
||||
other.smartSearch == smartSearch &&
|
||||
@ -71,6 +75,7 @@ class SystemConfigJobDto {
|
||||
(metadataExtraction.hashCode) +
|
||||
(migration.hashCode) +
|
||||
(notifications.hashCode) +
|
||||
(ocr.hashCode) +
|
||||
(search.hashCode) +
|
||||
(sidecar.hashCode) +
|
||||
(smartSearch.hashCode) +
|
||||
@ -78,7 +83,7 @@ class SystemConfigJobDto {
|
||||
(videoConversion.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'SystemConfigJobDto[backgroundTask=$backgroundTask, faceDetection=$faceDetection, library_=$library_, metadataExtraction=$metadataExtraction, migration=$migration, notifications=$notifications, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, thumbnailGeneration=$thumbnailGeneration, videoConversion=$videoConversion]';
|
||||
String toString() => 'SystemConfigJobDto[backgroundTask=$backgroundTask, faceDetection=$faceDetection, library_=$library_, metadataExtraction=$metadataExtraction, migration=$migration, notifications=$notifications, ocr=$ocr, search=$search, sidecar=$sidecar, smartSearch=$smartSearch, thumbnailGeneration=$thumbnailGeneration, videoConversion=$videoConversion]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -88,6 +93,7 @@ class SystemConfigJobDto {
|
||||
json[r'metadataExtraction'] = this.metadataExtraction;
|
||||
json[r'migration'] = this.migration;
|
||||
json[r'notifications'] = this.notifications;
|
||||
json[r'ocr'] = this.ocr;
|
||||
json[r'search'] = this.search;
|
||||
json[r'sidecar'] = this.sidecar;
|
||||
json[r'smartSearch'] = this.smartSearch;
|
||||
@ -111,6 +117,7 @@ class SystemConfigJobDto {
|
||||
metadataExtraction: JobSettingsDto.fromJson(json[r'metadataExtraction'])!,
|
||||
migration: JobSettingsDto.fromJson(json[r'migration'])!,
|
||||
notifications: JobSettingsDto.fromJson(json[r'notifications'])!,
|
||||
ocr: JobSettingsDto.fromJson(json[r'ocr'])!,
|
||||
search: JobSettingsDto.fromJson(json[r'search'])!,
|
||||
sidecar: JobSettingsDto.fromJson(json[r'sidecar'])!,
|
||||
smartSearch: JobSettingsDto.fromJson(json[r'smartSearch'])!,
|
||||
@ -169,6 +176,7 @@ class SystemConfigJobDto {
|
||||
'metadataExtraction',
|
||||
'migration',
|
||||
'notifications',
|
||||
'ocr',
|
||||
'search',
|
||||
'sidecar',
|
||||
'smartSearch',
|
||||
|
||||
@ -18,6 +18,7 @@ class SystemConfigMachineLearningDto {
|
||||
required this.duplicateDetection,
|
||||
required this.enabled,
|
||||
required this.facialRecognition,
|
||||
required this.ocr,
|
||||
this.urls = const [],
|
||||
});
|
||||
|
||||
@ -31,6 +32,8 @@ class SystemConfigMachineLearningDto {
|
||||
|
||||
FacialRecognitionConfig facialRecognition;
|
||||
|
||||
OcrConfig ocr;
|
||||
|
||||
List<String> urls;
|
||||
|
||||
@override
|
||||
@ -40,6 +43,7 @@ class SystemConfigMachineLearningDto {
|
||||
other.duplicateDetection == duplicateDetection &&
|
||||
other.enabled == enabled &&
|
||||
other.facialRecognition == facialRecognition &&
|
||||
other.ocr == ocr &&
|
||||
_deepEquality.equals(other.urls, urls);
|
||||
|
||||
@override
|
||||
@ -50,10 +54,11 @@ class SystemConfigMachineLearningDto {
|
||||
(duplicateDetection.hashCode) +
|
||||
(enabled.hashCode) +
|
||||
(facialRecognition.hashCode) +
|
||||
(ocr.hashCode) +
|
||||
(urls.hashCode);
|
||||
|
||||
@override
|
||||
String toString() => 'SystemConfigMachineLearningDto[availabilityChecks=$availabilityChecks, clip=$clip, duplicateDetection=$duplicateDetection, enabled=$enabled, facialRecognition=$facialRecognition, urls=$urls]';
|
||||
String toString() => 'SystemConfigMachineLearningDto[availabilityChecks=$availabilityChecks, clip=$clip, duplicateDetection=$duplicateDetection, enabled=$enabled, facialRecognition=$facialRecognition, ocr=$ocr, urls=$urls]';
|
||||
|
||||
Map<String, dynamic> toJson() {
|
||||
final json = <String, dynamic>{};
|
||||
@ -62,6 +67,7 @@ class SystemConfigMachineLearningDto {
|
||||
json[r'duplicateDetection'] = this.duplicateDetection;
|
||||
json[r'enabled'] = this.enabled;
|
||||
json[r'facialRecognition'] = this.facialRecognition;
|
||||
json[r'ocr'] = this.ocr;
|
||||
json[r'urls'] = this.urls;
|
||||
return json;
|
||||
}
|
||||
@ -80,6 +86,7 @@ class SystemConfigMachineLearningDto {
|
||||
duplicateDetection: DuplicateDetectionConfig.fromJson(json[r'duplicateDetection'])!,
|
||||
enabled: mapValueOfType<bool>(json, r'enabled')!,
|
||||
facialRecognition: FacialRecognitionConfig.fromJson(json[r'facialRecognition'])!,
|
||||
ocr: OcrConfig.fromJson(json[r'ocr'])!,
|
||||
urls: json[r'urls'] is Iterable
|
||||
? (json[r'urls'] as Iterable).cast<String>().toList(growable: false)
|
||||
: const [],
|
||||
@ -135,6 +142,7 @@ class SystemConfigMachineLearningDto {
|
||||
'duplicateDetection',
|
||||
'enabled',
|
||||
'facialRecognition',
|
||||
'ocr',
|
||||
'urls',
|
||||
};
|
||||
}
|
||||
|
||||
@ -5984,6 +5984,14 @@
|
||||
"type": "string"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "ocr",
|
||||
"required": false,
|
||||
"in": "query",
|
||||
"schema": {
|
||||
"type": "string"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "personIds",
|
||||
"required": false,
|
||||
@ -10408,6 +10416,9 @@
|
||||
"notifications": {
|
||||
"$ref": "#/components/schemas/JobStatusDto"
|
||||
},
|
||||
"ocr": {
|
||||
"$ref": "#/components/schemas/JobStatusDto"
|
||||
},
|
||||
"search": {
|
||||
"$ref": "#/components/schemas/JobStatusDto"
|
||||
},
|
||||
@ -10437,6 +10448,7 @@
|
||||
"metadataExtraction",
|
||||
"migration",
|
||||
"notifications",
|
||||
"ocr",
|
||||
"search",
|
||||
"sidecar",
|
||||
"smartSearch",
|
||||
@ -12108,7 +12120,8 @@
|
||||
"sidecar",
|
||||
"library",
|
||||
"notifications",
|
||||
"backupDatabase"
|
||||
"backupDatabase",
|
||||
"ocr"
|
||||
],
|
||||
"type": "string"
|
||||
},
|
||||
@ -12678,6 +12691,9 @@
|
||||
"nullable": true,
|
||||
"type": "string"
|
||||
},
|
||||
"ocr": {
|
||||
"type": "string"
|
||||
},
|
||||
"order": {
|
||||
"allOf": [
|
||||
{
|
||||
@ -12989,6 +13005,40 @@
|
||||
],
|
||||
"type": "string"
|
||||
},
|
||||
"OcrConfig": {
|
||||
"properties": {
|
||||
"enabled": {
|
||||
"type": "boolean"
|
||||
},
|
||||
"maxResolution": {
|
||||
"minimum": 1,
|
||||
"type": "integer"
|
||||
},
|
||||
"minDetectionScore": {
|
||||
"format": "double",
|
||||
"maximum": 1,
|
||||
"minimum": 0.1,
|
||||
"type": "number"
|
||||
},
|
||||
"minRecognitionScore": {
|
||||
"format": "double",
|
||||
"maximum": 1,
|
||||
"minimum": 0.1,
|
||||
"type": "number"
|
||||
},
|
||||
"modelName": {
|
||||
"type": "string"
|
||||
}
|
||||
},
|
||||
"required": [
|
||||
"enabled",
|
||||
"maxResolution",
|
||||
"minDetectionScore",
|
||||
"minRecognitionScore",
|
||||
"modelName"
|
||||
],
|
||||
"type": "object"
|
||||
},
|
||||
"OnThisDayDto": {
|
||||
"properties": {
|
||||
"year": {
|
||||
@ -13659,6 +13709,9 @@
|
||||
"nullable": true,
|
||||
"type": "string"
|
||||
},
|
||||
"ocr": {
|
||||
"type": "string"
|
||||
},
|
||||
"personIds": {
|
||||
"items": {
|
||||
"format": "uuid",
|
||||
@ -14127,6 +14180,9 @@
|
||||
"oauthAutoLaunch": {
|
||||
"type": "boolean"
|
||||
},
|
||||
"ocr": {
|
||||
"type": "boolean"
|
||||
},
|
||||
"passwordLogin": {
|
||||
"type": "boolean"
|
||||
},
|
||||
@ -14155,6 +14211,7 @@
|
||||
"map",
|
||||
"oauth",
|
||||
"oauthAutoLaunch",
|
||||
"ocr",
|
||||
"passwordLogin",
|
||||
"reverseGeocoding",
|
||||
"search",
|
||||
@ -14762,6 +14819,9 @@
|
||||
"nullable": true,
|
||||
"type": "string"
|
||||
},
|
||||
"ocr": {
|
||||
"type": "string"
|
||||
},
|
||||
"page": {
|
||||
"minimum": 1,
|
||||
"type": "number"
|
||||
@ -14967,6 +15027,9 @@
|
||||
"nullable": true,
|
||||
"type": "string"
|
||||
},
|
||||
"ocr": {
|
||||
"type": "string"
|
||||
},
|
||||
"personIds": {
|
||||
"items": {
|
||||
"format": "uuid",
|
||||
@ -16416,6 +16479,9 @@
|
||||
"notifications": {
|
||||
"$ref": "#/components/schemas/JobSettingsDto"
|
||||
},
|
||||
"ocr": {
|
||||
"$ref": "#/components/schemas/JobSettingsDto"
|
||||
},
|
||||
"search": {
|
||||
"$ref": "#/components/schemas/JobSettingsDto"
|
||||
},
|
||||
@ -16439,6 +16505,7 @@
|
||||
"metadataExtraction",
|
||||
"migration",
|
||||
"notifications",
|
||||
"ocr",
|
||||
"search",
|
||||
"sidecar",
|
||||
"smartSearch",
|
||||
@ -16524,6 +16591,9 @@
|
||||
"facialRecognition": {
|
||||
"$ref": "#/components/schemas/FacialRecognitionConfig"
|
||||
},
|
||||
"ocr": {
|
||||
"$ref": "#/components/schemas/OcrConfig"
|
||||
},
|
||||
"urls": {
|
||||
"format": "uri",
|
||||
"items": {
|
||||
@ -16540,6 +16610,7 @@
|
||||
"duplicateDetection",
|
||||
"enabled",
|
||||
"facialRecognition",
|
||||
"ocr",
|
||||
"urls"
|
||||
],
|
||||
"type": "object"
|
||||
|
||||
@ -696,6 +696,7 @@ export type AllJobStatusResponseDto = {
|
||||
metadataExtraction: JobStatusDto;
|
||||
migration: JobStatusDto;
|
||||
notifications: JobStatusDto;
|
||||
ocr: JobStatusDto;
|
||||
search: JobStatusDto;
|
||||
sidecar: JobStatusDto;
|
||||
smartSearch: JobStatusDto;
|
||||
@ -926,6 +927,7 @@ export type MetadataSearchDto = {
|
||||
libraryId?: string | null;
|
||||
make?: string;
|
||||
model?: string | null;
|
||||
ocr?: string;
|
||||
order?: AssetOrder;
|
||||
originalFileName?: string;
|
||||
originalPath?: string;
|
||||
@ -998,6 +1000,7 @@ export type RandomSearchDto = {
|
||||
libraryId?: string | null;
|
||||
make?: string;
|
||||
model?: string | null;
|
||||
ocr?: string;
|
||||
personIds?: string[];
|
||||
rating?: number;
|
||||
size?: number;
|
||||
@ -1033,6 +1036,7 @@ export type SmartSearchDto = {
|
||||
libraryId?: string | null;
|
||||
make?: string;
|
||||
model?: string | null;
|
||||
ocr?: string;
|
||||
page?: number;
|
||||
personIds?: string[];
|
||||
query?: string;
|
||||
@ -1069,6 +1073,7 @@ export type StatisticsSearchDto = {
|
||||
libraryId?: string | null;
|
||||
make?: string;
|
||||
model?: string | null;
|
||||
ocr?: string;
|
||||
personIds?: string[];
|
||||
rating?: number;
|
||||
state?: string | null;
|
||||
@ -1135,6 +1140,7 @@ export type ServerFeaturesDto = {
|
||||
map: boolean;
|
||||
oauth: boolean;
|
||||
oauthAutoLaunch: boolean;
|
||||
ocr: boolean;
|
||||
passwordLogin: boolean;
|
||||
reverseGeocoding: boolean;
|
||||
search: boolean;
|
||||
@ -1371,6 +1377,7 @@ export type SystemConfigJobDto = {
|
||||
metadataExtraction: JobSettingsDto;
|
||||
migration: JobSettingsDto;
|
||||
notifications: JobSettingsDto;
|
||||
ocr: JobSettingsDto;
|
||||
search: JobSettingsDto;
|
||||
sidecar: JobSettingsDto;
|
||||
smartSearch: JobSettingsDto;
|
||||
@ -1412,12 +1419,20 @@ export type FacialRecognitionConfig = {
|
||||
minScore: number;
|
||||
modelName: string;
|
||||
};
|
||||
export type OcrConfig = {
|
||||
enabled: boolean;
|
||||
maxResolution: number;
|
||||
minDetectionScore: number;
|
||||
minRecognitionScore: number;
|
||||
modelName: string;
|
||||
};
|
||||
export type SystemConfigMachineLearningDto = {
|
||||
availabilityChecks: MachineLearningAvailabilityChecksDto;
|
||||
clip: ClipConfig;
|
||||
duplicateDetection: DuplicateDetectionConfig;
|
||||
enabled: boolean;
|
||||
facialRecognition: FacialRecognitionConfig;
|
||||
ocr: OcrConfig;
|
||||
urls: string[];
|
||||
};
|
||||
export type SystemConfigMapDto = {
|
||||
@ -3399,7 +3414,7 @@ export function getExploreData(opts?: Oazapfts.RequestOpts) {
|
||||
/**
|
||||
* This endpoint requires the `asset.read` permission.
|
||||
*/
|
||||
export function searchLargeAssets({ albumIds, city, country, createdAfter, createdBefore, deviceId, isEncoded, isFavorite, isMotion, isNotInAlbum, isOffline, lensModel, libraryId, make, minFileSize, model, personIds, rating, size, state, tagIds, takenAfter, takenBefore, trashedAfter, trashedBefore, $type, updatedAfter, updatedBefore, visibility, withDeleted, withExif }: {
|
||||
export function searchLargeAssets({ albumIds, city, country, createdAfter, createdBefore, deviceId, isEncoded, isFavorite, isMotion, isNotInAlbum, isOffline, lensModel, libraryId, make, minFileSize, model, ocr, personIds, rating, size, state, tagIds, takenAfter, takenBefore, trashedAfter, trashedBefore, $type, updatedAfter, updatedBefore, visibility, withDeleted, withExif }: {
|
||||
albumIds?: string[];
|
||||
city?: string | null;
|
||||
country?: string | null;
|
||||
@ -3416,6 +3431,7 @@ export function searchLargeAssets({ albumIds, city, country, createdAfter, creat
|
||||
make?: string;
|
||||
minFileSize?: number;
|
||||
model?: string | null;
|
||||
ocr?: string;
|
||||
personIds?: string[];
|
||||
rating?: number;
|
||||
size?: number;
|
||||
@ -3452,6 +3468,7 @@ export function searchLargeAssets({ albumIds, city, country, createdAfter, creat
|
||||
make,
|
||||
minFileSize,
|
||||
model,
|
||||
ocr,
|
||||
personIds,
|
||||
rating,
|
||||
size,
|
||||
@ -4901,7 +4918,8 @@ export enum JobName {
|
||||
Sidecar = "sidecar",
|
||||
Library = "library",
|
||||
Notifications = "notifications",
|
||||
BackupDatabase = "backupDatabase"
|
||||
BackupDatabase = "backupDatabase",
|
||||
Ocr = "ocr"
|
||||
}
|
||||
export enum JobCommand {
|
||||
Start = "start",
|
||||
|
||||
@ -74,6 +74,13 @@ export interface SystemConfig {
|
||||
minFaces: number;
|
||||
maxDistance: number;
|
||||
};
|
||||
ocr: {
|
||||
enabled: boolean;
|
||||
modelName: string;
|
||||
minDetectionScore: number;
|
||||
minRecognitionScore: number;
|
||||
maxResolution: number;
|
||||
};
|
||||
};
|
||||
map: {
|
||||
enabled: boolean;
|
||||
@ -227,6 +234,7 @@ export const defaults = Object.freeze<SystemConfig>({
|
||||
[QueueName.ThumbnailGeneration]: { concurrency: 3 },
|
||||
[QueueName.VideoConversion]: { concurrency: 1 },
|
||||
[QueueName.Notification]: { concurrency: 5 },
|
||||
[QueueName.Ocr]: { concurrency: 1 },
|
||||
},
|
||||
logging: {
|
||||
enabled: true,
|
||||
@ -255,6 +263,13 @@ export const defaults = Object.freeze<SystemConfig>({
|
||||
maxDistance: 0.5,
|
||||
minFaces: 3,
|
||||
},
|
||||
ocr: {
|
||||
enabled: true,
|
||||
modelName: 'PP-OCRv5_mobile',
|
||||
minDetectionScore: 0.5,
|
||||
minRecognitionScore: 0.8,
|
||||
maxResolution: 736,
|
||||
},
|
||||
},
|
||||
map: {
|
||||
enabled: true,
|
||||
|
||||
@ -93,4 +93,7 @@ export class AllJobStatusResponseDto implements Record<QueueName, JobStatusDto>
|
||||
|
||||
@ApiProperty({ type: JobStatusDto })
|
||||
[QueueName.BackupDatabase]!: JobStatusDto;
|
||||
|
||||
@ApiProperty({ type: JobStatusDto })
|
||||
[QueueName.Ocr]!: JobStatusDto;
|
||||
}
|
||||
|
||||
@ -46,3 +46,25 @@ export class FacialRecognitionConfig extends ModelConfig {
|
||||
@ApiProperty({ type: 'integer' })
|
||||
minFaces!: number;
|
||||
}
|
||||
|
||||
export class OcrConfig extends ModelConfig {
|
||||
@IsNumber()
|
||||
@Min(1)
|
||||
@Type(() => Number)
|
||||
@ApiProperty({ type: 'integer' })
|
||||
maxResolution!: number;
|
||||
|
||||
@IsNumber()
|
||||
@Min(0.1)
|
||||
@Max(1)
|
||||
@Type(() => Number)
|
||||
@ApiProperty({ type: 'number', format: 'double' })
|
||||
minDetectionScore!: number;
|
||||
|
||||
@IsNumber()
|
||||
@Min(0.1)
|
||||
@Max(1)
|
||||
@Type(() => Number)
|
||||
@ApiProperty({ type: 'number', format: 'double' })
|
||||
minRecognitionScore!: number;
|
||||
}
|
||||
|
||||
@ -101,6 +101,11 @@ class BaseSearchDto {
|
||||
@Max(5)
|
||||
@Min(-1)
|
||||
rating?: number;
|
||||
|
||||
@IsString()
|
||||
@IsNotEmpty()
|
||||
@Optional()
|
||||
ocr?: string;
|
||||
}
|
||||
|
||||
class BaseSearchWithResultsDto extends BaseSearchDto {
|
||||
|
||||
@ -171,6 +171,7 @@ export class ServerFeaturesDto {
|
||||
sidecar!: boolean;
|
||||
search!: boolean;
|
||||
email!: boolean;
|
||||
ocr!: boolean;
|
||||
}
|
||||
|
||||
export interface ReleaseNotification {
|
||||
|
||||
@ -15,7 +15,7 @@ import {
|
||||
ValidateNested,
|
||||
} from 'class-validator';
|
||||
import { SystemConfig } from 'src/config';
|
||||
import { CLIPConfig, DuplicateDetectionConfig, FacialRecognitionConfig } from 'src/dtos/model-config.dto';
|
||||
import { CLIPConfig, DuplicateDetectionConfig, FacialRecognitionConfig, OcrConfig } from 'src/dtos/model-config.dto';
|
||||
import {
|
||||
AudioCodec,
|
||||
CQMode,
|
||||
@ -201,6 +201,12 @@ class SystemConfigJobDto implements Record<ConcurrentQueueName, JobSettingsDto>
|
||||
@Type(() => JobSettingsDto)
|
||||
[QueueName.FaceDetection]!: JobSettingsDto;
|
||||
|
||||
@ApiProperty({ type: JobSettingsDto })
|
||||
@ValidateNested()
|
||||
@IsObject()
|
||||
@Type(() => JobSettingsDto)
|
||||
[QueueName.Ocr]!: JobSettingsDto;
|
||||
|
||||
@ApiProperty({ type: JobSettingsDto })
|
||||
@ValidateNested()
|
||||
@IsObject()
|
||||
@ -296,6 +302,11 @@ class SystemConfigMachineLearningDto {
|
||||
@ValidateNested()
|
||||
@IsObject()
|
||||
facialRecognition!: FacialRecognitionConfig;
|
||||
|
||||
@Type(() => OcrConfig)
|
||||
@ValidateNested()
|
||||
@IsObject()
|
||||
ocr!: OcrConfig;
|
||||
}
|
||||
|
||||
enum MapTheme {
|
||||
|
||||
@ -513,6 +513,7 @@ export enum QueueName {
|
||||
Library = 'library',
|
||||
Notification = 'notifications',
|
||||
BackupDatabase = 'backupDatabase',
|
||||
Ocr = 'ocr',
|
||||
}
|
||||
|
||||
export enum JobName {
|
||||
@ -585,6 +586,10 @@ export enum JobName {
|
||||
TagCleanup = 'TagCleanup',
|
||||
|
||||
VersionCheck = 'VersionCheck',
|
||||
|
||||
// OCR
|
||||
OcrQueueAll = 'OcrQueueAll',
|
||||
Ocr = 'Ocr',
|
||||
}
|
||||
|
||||
export enum JobCommand {
|
||||
|
||||
@ -285,6 +285,23 @@ from
|
||||
where
|
||||
"asset"."id" = $2
|
||||
|
||||
-- AssetJobRepository.getForOcr
|
||||
select
|
||||
"asset"."visibility",
|
||||
(
|
||||
select
|
||||
"asset_file"."path"
|
||||
from
|
||||
"asset_file"
|
||||
where
|
||||
"asset_file"."assetId" = "asset"."id"
|
||||
and "asset_file"."type" = $1
|
||||
) as "previewFile"
|
||||
from
|
||||
"asset"
|
||||
where
|
||||
"asset"."id" = $2
|
||||
|
||||
-- AssetJobRepository.getForSyncAssets
|
||||
select
|
||||
"asset"."id",
|
||||
@ -483,6 +500,17 @@ where
|
||||
order by
|
||||
"asset"."fileCreatedAt" desc
|
||||
|
||||
-- AssetJobRepository.streamForOcrJob
|
||||
select
|
||||
"asset"."id"
|
||||
from
|
||||
"asset"
|
||||
inner join "asset_job_status" on "asset_job_status"."assetId" = "asset"."id"
|
||||
where
|
||||
"asset_job_status"."ocrAt" is null
|
||||
and "asset"."deletedAt" is null
|
||||
and "asset"."visibility" != $1
|
||||
|
||||
-- AssetJobRepository.streamForMigrationJob
|
||||
select
|
||||
"id"
|
||||
|
||||
68
server/src/queries/ocr.repository.sql
Normal file
68
server/src/queries/ocr.repository.sql
Normal file
@ -0,0 +1,68 @@
|
||||
-- NOTE: This file is auto generated by ./sql-generator
|
||||
|
||||
-- OcrRepository.getById
|
||||
select
|
||||
"asset_ocr".*
|
||||
from
|
||||
"asset_ocr"
|
||||
where
|
||||
"asset_ocr"."id" = $1
|
||||
|
||||
-- OcrRepository.getByAssetId
|
||||
select
|
||||
"asset_ocr".*
|
||||
from
|
||||
"asset_ocr"
|
||||
where
|
||||
"asset_ocr"."assetId" = $1
|
||||
|
||||
-- OcrRepository.upsert
|
||||
with
|
||||
"deleted_ocr" as (
|
||||
delete from "asset_ocr"
|
||||
where
|
||||
"assetId" = $1
|
||||
),
|
||||
"inserted_ocr" as (
|
||||
insert into
|
||||
"asset_ocr" (
|
||||
"assetId",
|
||||
"x1",
|
||||
"y1",
|
||||
"x2",
|
||||
"y2",
|
||||
"x3",
|
||||
"y3",
|
||||
"x4",
|
||||
"y4",
|
||||
"text",
|
||||
"boxScore",
|
||||
"textScore"
|
||||
)
|
||||
values
|
||||
(
|
||||
$2,
|
||||
$3,
|
||||
$4,
|
||||
$5,
|
||||
$6,
|
||||
$7,
|
||||
$8,
|
||||
$9,
|
||||
$10,
|
||||
$11,
|
||||
$12,
|
||||
$13
|
||||
)
|
||||
),
|
||||
"inserted_search" as (
|
||||
insert into
|
||||
"ocr_search" ("assetId", "text")
|
||||
values
|
||||
($14, $15)
|
||||
on conflict ("assetId") do update
|
||||
set
|
||||
"text" = "excluded"."text"
|
||||
)
|
||||
select
|
||||
1 as "dummy"
|
||||
@ -16,6 +16,7 @@ import {
|
||||
withExifInner,
|
||||
withFaces,
|
||||
withFacesAndPeople,
|
||||
withFilePath,
|
||||
withFiles,
|
||||
} from 'src/utils/database';
|
||||
|
||||
@ -192,6 +193,15 @@ export class AssetJobRepository {
|
||||
.executeTakeFirst();
|
||||
}
|
||||
|
||||
@GenerateSql({ params: [DummyValue.UUID] })
|
||||
getForOcr(id: string) {
|
||||
return this.db
|
||||
.selectFrom('asset')
|
||||
.select((eb) => ['asset.visibility', withFilePath(eb, AssetFileType.Preview).as('previewFile')])
|
||||
.where('asset.id', '=', id)
|
||||
.executeTakeFirst();
|
||||
}
|
||||
|
||||
@GenerateSql({ params: [[DummyValue.UUID]] })
|
||||
getForSyncAssets(ids: string[]) {
|
||||
return this.db
|
||||
@ -348,6 +358,21 @@ export class AssetJobRepository {
|
||||
.stream();
|
||||
}
|
||||
|
||||
@GenerateSql({ params: [], stream: true })
|
||||
streamForOcrJob(force?: boolean) {
|
||||
return this.db
|
||||
.selectFrom('asset')
|
||||
.select(['asset.id'])
|
||||
.$if(!force, (qb) =>
|
||||
qb
|
||||
.innerJoin('asset_job_status', 'asset_job_status.assetId', 'asset.id')
|
||||
.where('asset_job_status.ocrAt', 'is', null),
|
||||
)
|
||||
.where('asset.deletedAt', 'is', null)
|
||||
.where('asset.visibility', '!=', AssetVisibility.Hidden)
|
||||
.stream();
|
||||
}
|
||||
|
||||
@GenerateSql({ params: [DummyValue.DATE], stream: true })
|
||||
streamForMigrationJob() {
|
||||
return this.db.selectFrom('asset').select(['id']).where('asset.deletedAt', 'is', null).stream();
|
||||
|
||||
@ -205,6 +205,7 @@ export class AssetRepository {
|
||||
metadataExtractedAt: eb.ref('excluded.metadataExtractedAt'),
|
||||
previewAt: eb.ref('excluded.previewAt'),
|
||||
thumbnailAt: eb.ref('excluded.thumbnailAt'),
|
||||
ocrAt: eb.ref('excluded.ocrAt'),
|
||||
},
|
||||
values[0],
|
||||
),
|
||||
|
||||
@ -25,6 +25,7 @@ import { MetadataRepository } from 'src/repositories/metadata.repository';
|
||||
import { MoveRepository } from 'src/repositories/move.repository';
|
||||
import { NotificationRepository } from 'src/repositories/notification.repository';
|
||||
import { OAuthRepository } from 'src/repositories/oauth.repository';
|
||||
import { OcrRepository } from 'src/repositories/ocr.repository';
|
||||
import { PartnerRepository } from 'src/repositories/partner.repository';
|
||||
import { PersonRepository } from 'src/repositories/person.repository';
|
||||
import { ProcessRepository } from 'src/repositories/process.repository';
|
||||
@ -74,6 +75,7 @@ export const repositories = [
|
||||
MoveRepository,
|
||||
NotificationRepository,
|
||||
OAuthRepository,
|
||||
OcrRepository,
|
||||
PartnerRepository,
|
||||
PersonRepository,
|
||||
ProcessRepository,
|
||||
|
||||
@ -15,6 +15,7 @@ export interface BoundingBox {
|
||||
export enum ModelTask {
|
||||
FACIAL_RECOGNITION = 'facial-recognition',
|
||||
SEARCH = 'clip',
|
||||
OCR = 'ocr',
|
||||
}
|
||||
|
||||
export enum ModelType {
|
||||
@ -23,6 +24,7 @@ export enum ModelType {
|
||||
RECOGNITION = 'recognition',
|
||||
TEXTUAL = 'textual',
|
||||
VISUAL = 'visual',
|
||||
OCR = 'ocr',
|
||||
}
|
||||
|
||||
export type ModelPayload = { imagePath: string } | { text: string };
|
||||
@ -30,7 +32,11 @@ export type ModelPayload = { imagePath: string } | { text: string };
|
||||
type ModelOptions = { modelName: string };
|
||||
|
||||
export type FaceDetectionOptions = ModelOptions & { minScore: number };
|
||||
|
||||
export type OcrOptions = ModelOptions & {
|
||||
minDetectionScore: number;
|
||||
minRecognitionScore: number;
|
||||
maxResolution: number;
|
||||
};
|
||||
type VisualResponse = { imageHeight: number; imageWidth: number };
|
||||
export type ClipVisualRequest = { [ModelTask.SEARCH]: { [ModelType.VISUAL]: ModelOptions } };
|
||||
export type ClipVisualResponse = { [ModelTask.SEARCH]: string } & VisualResponse;
|
||||
@ -38,6 +44,21 @@ export type ClipVisualResponse = { [ModelTask.SEARCH]: string } & VisualResponse
|
||||
export type ClipTextualRequest = { [ModelTask.SEARCH]: { [ModelType.TEXTUAL]: ModelOptions } };
|
||||
export type ClipTextualResponse = { [ModelTask.SEARCH]: string };
|
||||
|
||||
export type OCR = {
|
||||
text: string[];
|
||||
box: number[];
|
||||
boxScore: number[];
|
||||
textScore: number[];
|
||||
};
|
||||
|
||||
export type OcrRequest = {
|
||||
[ModelTask.OCR]: {
|
||||
[ModelType.DETECTION]: ModelOptions & { options: { minScore: number; maxResolution: number } };
|
||||
[ModelType.RECOGNITION]: ModelOptions & { options: { minScore: number } };
|
||||
};
|
||||
};
|
||||
export type OcrResponse = { [ModelTask.OCR]: OCR } & VisualResponse;
|
||||
|
||||
export type FacialRecognitionRequest = {
|
||||
[ModelTask.FACIAL_RECOGNITION]: {
|
||||
[ModelType.DETECTION]: ModelOptions & { options: { minScore: number } };
|
||||
@ -53,7 +74,7 @@ export interface Face {
|
||||
|
||||
export type FacialRecognitionResponse = { [ModelTask.FACIAL_RECOGNITION]: Face[] } & VisualResponse;
|
||||
export type DetectedFaces = { faces: Face[] } & VisualResponse;
|
||||
export type MachineLearningRequest = ClipVisualRequest | ClipTextualRequest | FacialRecognitionRequest;
|
||||
export type MachineLearningRequest = ClipVisualRequest | ClipTextualRequest | FacialRecognitionRequest | OcrRequest;
|
||||
export type TextEncodingOptions = ModelOptions & { language?: string };
|
||||
|
||||
@Injectable()
|
||||
@ -197,6 +218,17 @@ export class MachineLearningRepository {
|
||||
return response[ModelTask.SEARCH];
|
||||
}
|
||||
|
||||
async ocr(imagePath: string, { modelName, minDetectionScore, minRecognitionScore, maxResolution }: OcrOptions) {
|
||||
const request = {
|
||||
[ModelTask.OCR]: {
|
||||
[ModelType.DETECTION]: { modelName, options: { minScore: minDetectionScore, maxResolution } },
|
||||
[ModelType.RECOGNITION]: { modelName, options: { minScore: minRecognitionScore } },
|
||||
},
|
||||
};
|
||||
const response = await this.predict<OcrResponse>({ imagePath }, request);
|
||||
return response[ModelTask.OCR];
|
||||
}
|
||||
|
||||
private async getFormData(payload: ModelPayload, config: MachineLearningRequest): Promise<FormData> {
|
||||
const formData = new FormData();
|
||||
formData.append('entries', JSON.stringify(config));
|
||||
|
||||
68
server/src/repositories/ocr.repository.ts
Normal file
68
server/src/repositories/ocr.repository.ts
Normal file
@ -0,0 +1,68 @@
|
||||
import { Injectable } from '@nestjs/common';
|
||||
import { Insertable, Kysely, sql } from 'kysely';
|
||||
import { InjectKysely } from 'nestjs-kysely';
|
||||
import { DummyValue, GenerateSql } from 'src/decorators';
|
||||
import { DB } from 'src/schema';
|
||||
import { AssetOcrTable } from 'src/schema/tables/asset-ocr.table';
|
||||
|
||||
@Injectable()
|
||||
export class OcrRepository {
|
||||
constructor(@InjectKysely() private db: Kysely<DB>) {}
|
||||
|
||||
@GenerateSql({ params: [DummyValue.UUID] })
|
||||
getById(id: string) {
|
||||
return this.db.selectFrom('asset_ocr').selectAll('asset_ocr').where('asset_ocr.id', '=', id).executeTakeFirst();
|
||||
}
|
||||
|
||||
@GenerateSql({ params: [DummyValue.UUID] })
|
||||
getByAssetId(id: string) {
|
||||
return this.db.selectFrom('asset_ocr').selectAll('asset_ocr').where('asset_ocr.assetId', '=', id).execute();
|
||||
}
|
||||
|
||||
deleteAll() {
|
||||
return this.db.transaction().execute(async (trx: Kysely<DB>) => {
|
||||
await sql`truncate ${sql.table('asset_ocr')}`.execute(trx);
|
||||
await sql`truncate ${sql.table('ocr_search')}`.execute(trx);
|
||||
});
|
||||
}
|
||||
|
||||
@GenerateSql({
|
||||
params: [
|
||||
DummyValue.UUID,
|
||||
[
|
||||
{
|
||||
assetId: DummyValue.UUID,
|
||||
x1: DummyValue.NUMBER,
|
||||
y1: DummyValue.NUMBER,
|
||||
x2: DummyValue.NUMBER,
|
||||
y2: DummyValue.NUMBER,
|
||||
x3: DummyValue.NUMBER,
|
||||
y3: DummyValue.NUMBER,
|
||||
x4: DummyValue.NUMBER,
|
||||
y4: DummyValue.NUMBER,
|
||||
text: DummyValue.STRING,
|
||||
boxScore: DummyValue.NUMBER,
|
||||
textScore: DummyValue.NUMBER,
|
||||
},
|
||||
],
|
||||
],
|
||||
})
|
||||
upsert(assetId: string, ocrDataList: Insertable<AssetOcrTable>[]) {
|
||||
let query = this.db.with('deleted_ocr', (db) => db.deleteFrom('asset_ocr').where('assetId', '=', assetId));
|
||||
if (ocrDataList.length > 0) {
|
||||
const searchText = ocrDataList.map((item) => item.text.trim()).join(' ');
|
||||
(query as any) = query
|
||||
.with('inserted_ocr', (db) => db.insertInto('asset_ocr').values(ocrDataList))
|
||||
.with('inserted_search', (db) =>
|
||||
db
|
||||
.insertInto('ocr_search')
|
||||
.values({ assetId, text: searchText })
|
||||
.onConflict((oc) => oc.column('assetId').doUpdateSet((eb) => ({ text: eb.ref('excluded.text') }))),
|
||||
);
|
||||
} else {
|
||||
(query as any) = query.with('deleted_search', (db) => db.deleteFrom('ocr_search').where('assetId', '=', assetId));
|
||||
}
|
||||
|
||||
return query.selectNoFrom(sql`1`.as('dummy')).execute();
|
||||
}
|
||||
}
|
||||
@ -84,6 +84,10 @@ export interface SearchEmbeddingOptions {
|
||||
userIds: string[];
|
||||
}
|
||||
|
||||
export interface SearchOcrOptions {
|
||||
ocr?: string;
|
||||
}
|
||||
|
||||
export interface SearchPeopleOptions {
|
||||
personIds?: string[];
|
||||
}
|
||||
@ -114,7 +118,8 @@ type BaseAssetSearchOptions = SearchDateOptions &
|
||||
SearchUserIdOptions &
|
||||
SearchPeopleOptions &
|
||||
SearchTagOptions &
|
||||
SearchAlbumOptions;
|
||||
SearchAlbumOptions &
|
||||
SearchOcrOptions;
|
||||
|
||||
export type AssetSearchOptions = BaseAssetSearchOptions & SearchRelationOptions;
|
||||
|
||||
@ -127,7 +132,10 @@ export type SmartSearchOptions = SearchDateOptions &
|
||||
SearchStatusOptions &
|
||||
SearchUserIdOptions &
|
||||
SearchPeopleOptions &
|
||||
SearchTagOptions;
|
||||
SearchTagOptions &
|
||||
SearchOcrOptions;
|
||||
|
||||
export type OcrSearchOptions = SearchDateOptions & SearchOcrOptions;
|
||||
|
||||
export type LargeAssetSearchOptions = AssetSearchOptions & { minFileSize?: number };
|
||||
|
||||
|
||||
@ -35,6 +35,7 @@ import { AssetFileTable } from 'src/schema/tables/asset-file.table';
|
||||
import { AssetJobStatusTable } from 'src/schema/tables/asset-job-status.table';
|
||||
import { AssetMetadataAuditTable } from 'src/schema/tables/asset-metadata-audit.table';
|
||||
import { AssetMetadataTable } from 'src/schema/tables/asset-metadata.table';
|
||||
import { AssetOcrTable } from 'src/schema/tables/asset-ocr.table';
|
||||
import { AssetTable } from 'src/schema/tables/asset.table';
|
||||
import { AuditTable } from 'src/schema/tables/audit.table';
|
||||
import { FaceSearchTable } from 'src/schema/tables/face-search.table';
|
||||
@ -47,6 +48,7 @@ import { MemoryTable } from 'src/schema/tables/memory.table';
|
||||
import { MoveTable } from 'src/schema/tables/move.table';
|
||||
import { NaturalEarthCountriesTable } from 'src/schema/tables/natural-earth-countries.table';
|
||||
import { NotificationTable } from 'src/schema/tables/notification.table';
|
||||
import { OcrSearchTable } from 'src/schema/tables/ocr-search.table';
|
||||
import { PartnerAuditTable } from 'src/schema/tables/partner-audit.table';
|
||||
import { PartnerTable } from 'src/schema/tables/partner.table';
|
||||
import { PersonAuditTable } from 'src/schema/tables/person-audit.table';
|
||||
@ -87,6 +89,7 @@ export class ImmichDatabase {
|
||||
AssetMetadataTable,
|
||||
AssetMetadataAuditTable,
|
||||
AssetJobStatusTable,
|
||||
AssetOcrTable,
|
||||
AssetTable,
|
||||
AssetFileTable,
|
||||
AuditTable,
|
||||
@ -101,6 +104,7 @@ export class ImmichDatabase {
|
||||
MoveTable,
|
||||
NaturalEarthCountriesTable,
|
||||
NotificationTable,
|
||||
OcrSearchTable,
|
||||
PartnerAuditTable,
|
||||
PartnerTable,
|
||||
PersonTable,
|
||||
@ -174,6 +178,8 @@ export interface DB {
|
||||
asset_metadata: AssetMetadataTable;
|
||||
asset_metadata_audit: AssetMetadataAuditTable;
|
||||
asset_job_status: AssetJobStatusTable;
|
||||
asset_ocr: AssetOcrTable;
|
||||
ocr_search: OcrSearchTable;
|
||||
|
||||
audit: AuditTable;
|
||||
|
||||
|
||||
@ -0,0 +1,16 @@
|
||||
import { Kysely, sql } from 'kysely';
|
||||
|
||||
export async function up(db: Kysely<any>): Promise<void> {
|
||||
await sql`CREATE TABLE "asset_ocr" ("id" uuid NOT NULL DEFAULT uuid_generate_v4(), "assetId" uuid NOT NULL, "x1" real NOT NULL, "y1" real NOT NULL, "x2" real NOT NULL, "y2" real NOT NULL, "x3" real NOT NULL, "y3" real NOT NULL, "x4" real NOT NULL, "y4" real NOT NULL, "boxScore" real NOT NULL, "textScore" real NOT NULL, "text" text NOT NULL);`.execute(
|
||||
db,
|
||||
);
|
||||
await sql`ALTER TABLE "asset_ocr" ADD CONSTRAINT "asset_ocr_pkey" PRIMARY KEY ("id");`.execute(db);
|
||||
await sql`ALTER TABLE "asset_ocr" ADD CONSTRAINT "asset_ocr_assetId_fkey" FOREIGN KEY ("assetId") REFERENCES "asset" ("id") ON UPDATE CASCADE ON DELETE CASCADE;`.execute(
|
||||
db,
|
||||
);
|
||||
await sql`CREATE INDEX "asset_ocr_assetId_idx" ON "asset_ocr" ("assetId")`.execute(db);
|
||||
}
|
||||
|
||||
export async function down(db: Kysely<any>): Promise<void> {
|
||||
await sql`DROP TABLE "asset_ocr";`.execute(db);
|
||||
}
|
||||
@ -0,0 +1,20 @@
|
||||
import { Kysely, sql } from 'kysely';
|
||||
|
||||
export async function up(db: Kysely<any>): Promise<void> {
|
||||
await sql`CREATE TABLE "ocr_search" ("assetId" uuid NOT NULL, "text" text NOT NULL);`.execute(db);
|
||||
await sql`ALTER TABLE "ocr_search" ADD CONSTRAINT "ocr_search_pkey" PRIMARY KEY ("assetId");`.execute(db);
|
||||
await sql`ALTER TABLE "ocr_search" ADD CONSTRAINT "ocr_search_assetId_fkey" FOREIGN KEY ("assetId") REFERENCES "asset" ("id") ON UPDATE CASCADE ON DELETE CASCADE;`.execute(
|
||||
db,
|
||||
);
|
||||
await sql`CREATE INDEX "idx_ocr_search_text" ON "ocr_search" USING gin (f_unaccent("text") gin_trgm_ops);`.execute(
|
||||
db,
|
||||
);
|
||||
await sql`INSERT INTO "migration_overrides" ("name", "value") VALUES ('index_idx_ocr_search_text', '{"type":"index","name":"idx_ocr_search_text","sql":"CREATE INDEX \\"idx_ocr_search_text\\" ON \\"ocr_search\\" USING gin (f_unaccent(\\"text\\") gin_trgm_ops);"}'::jsonb);`.execute(
|
||||
db,
|
||||
);
|
||||
}
|
||||
|
||||
export async function down(db: Kysely<any>): Promise<void> {
|
||||
await sql`DROP TABLE "ocr_search";`.execute(db);
|
||||
await sql`DELETE FROM "migration_overrides" WHERE "name" = 'index_idx_ocr_search_text';`.execute(db);
|
||||
}
|
||||
@ -0,0 +1,9 @@
|
||||
import { Kysely, sql } from 'kysely';
|
||||
|
||||
export async function up(db: Kysely<any>): Promise<void> {
|
||||
await sql`ALTER TABLE "asset_job_status" ADD "ocrAt" timestamp with time zone;`.execute(db);
|
||||
}
|
||||
|
||||
export async function down(db: Kysely<any>): Promise<void> {
|
||||
await sql`ALTER TABLE "asset_job_status" DROP COLUMN "ocrAt";`.execute(db);
|
||||
}
|
||||
@ -20,4 +20,7 @@ export class AssetJobStatusTable {
|
||||
|
||||
@Column({ type: 'timestamp with time zone', nullable: true })
|
||||
thumbnailAt!: Timestamp | null;
|
||||
|
||||
@Column({ type: 'timestamp with time zone', nullable: true })
|
||||
ocrAt!: Timestamp | null;
|
||||
}
|
||||
|
||||
45
server/src/schema/tables/asset-ocr.table.ts
Normal file
45
server/src/schema/tables/asset-ocr.table.ts
Normal file
@ -0,0 +1,45 @@
|
||||
import { AssetTable } from 'src/schema/tables/asset.table';
|
||||
import { Column, ForeignKeyColumn, Generated, PrimaryGeneratedColumn, Table } from 'src/sql-tools';
|
||||
|
||||
@Table('asset_ocr')
|
||||
export class AssetOcrTable {
|
||||
@PrimaryGeneratedColumn()
|
||||
id!: Generated<string>;
|
||||
|
||||
@ForeignKeyColumn(() => AssetTable, { onDelete: 'CASCADE', onUpdate: 'CASCADE' })
|
||||
assetId!: string;
|
||||
|
||||
// box positions are normalized, with values between 0 and 1
|
||||
@Column({ type: 'real' })
|
||||
x1!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
y1!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
x2!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
y2!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
x3!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
y3!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
x4!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
y4!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
boxScore!: number;
|
||||
|
||||
@Column({ type: 'real' })
|
||||
textScore!: number;
|
||||
|
||||
@Column({ type: 'text' })
|
||||
text!: string;
|
||||
}
|
||||
20
server/src/schema/tables/ocr-search.table.ts
Normal file
20
server/src/schema/tables/ocr-search.table.ts
Normal file
@ -0,0 +1,20 @@
|
||||
import { AssetTable } from 'src/schema/tables/asset.table';
|
||||
import { Column, ForeignKeyColumn, Index, Table } from 'src/sql-tools';
|
||||
|
||||
@Table('ocr_search')
|
||||
@Index({
|
||||
name: 'idx_ocr_search_text',
|
||||
using: 'gin',
|
||||
expression: 'f_unaccent("text") gin_trgm_ops',
|
||||
})
|
||||
export class OcrSearchTable {
|
||||
@ForeignKeyColumn(() => AssetTable, {
|
||||
onDelete: 'CASCADE',
|
||||
onUpdate: 'CASCADE',
|
||||
primary: true,
|
||||
})
|
||||
assetId!: string;
|
||||
|
||||
@Column({ type: 'text' })
|
||||
text!: string;
|
||||
}
|
||||
@ -32,6 +32,7 @@ import { MetadataRepository } from 'src/repositories/metadata.repository';
|
||||
import { MoveRepository } from 'src/repositories/move.repository';
|
||||
import { NotificationRepository } from 'src/repositories/notification.repository';
|
||||
import { OAuthRepository } from 'src/repositories/oauth.repository';
|
||||
import { OcrRepository } from 'src/repositories/ocr.repository';
|
||||
import { PartnerRepository } from 'src/repositories/partner.repository';
|
||||
import { PersonRepository } from 'src/repositories/person.repository';
|
||||
import { ProcessRepository } from 'src/repositories/process.repository';
|
||||
@ -84,6 +85,7 @@ export const BASE_SERVICE_DEPENDENCIES = [
|
||||
MoveRepository,
|
||||
NotificationRepository,
|
||||
OAuthRepository,
|
||||
OcrRepository,
|
||||
PartnerRepository,
|
||||
PersonRepository,
|
||||
ProcessRepository,
|
||||
@ -137,6 +139,7 @@ export class BaseService {
|
||||
protected moveRepository: MoveRepository,
|
||||
protected notificationRepository: NotificationRepository,
|
||||
protected oauthRepository: OAuthRepository,
|
||||
protected ocrRepository: OcrRepository,
|
||||
protected partnerRepository: PartnerRepository,
|
||||
protected personRepository: PersonRepository,
|
||||
protected processRepository: ProcessRepository,
|
||||
|
||||
@ -20,6 +20,7 @@ import { MemoryService } from 'src/services/memory.service';
|
||||
import { MetadataService } from 'src/services/metadata.service';
|
||||
import { NotificationAdminService } from 'src/services/notification-admin.service';
|
||||
import { NotificationService } from 'src/services/notification.service';
|
||||
import { OcrService } from 'src/services/ocr.service';
|
||||
import { PartnerService } from 'src/services/partner.service';
|
||||
import { PersonService } from 'src/services/person.service';
|
||||
import { SearchService } from 'src/services/search.service';
|
||||
@ -65,6 +66,7 @@ export const services = [
|
||||
MetadataService,
|
||||
NotificationService,
|
||||
NotificationAdminService,
|
||||
OcrService,
|
||||
PartnerService,
|
||||
PersonService,
|
||||
SearchService,
|
||||
|
||||
@ -24,7 +24,7 @@ describe(JobService.name, () => {
|
||||
it('should update concurrency', () => {
|
||||
sut.onConfigUpdate({ newConfig: defaults, oldConfig: {} as SystemConfig });
|
||||
|
||||
expect(mocks.job.setConcurrency).toHaveBeenCalledTimes(15);
|
||||
expect(mocks.job.setConcurrency).toHaveBeenCalledTimes(16);
|
||||
expect(mocks.job.setConcurrency).toHaveBeenNthCalledWith(5, QueueName.FacialRecognition, 1);
|
||||
expect(mocks.job.setConcurrency).toHaveBeenNthCalledWith(7, QueueName.DuplicateDetection, 1);
|
||||
expect(mocks.job.setConcurrency).toHaveBeenNthCalledWith(8, QueueName.BackgroundTask, 5);
|
||||
@ -98,6 +98,7 @@ describe(JobService.name, () => {
|
||||
[QueueName.Library]: expectedJobStatus,
|
||||
[QueueName.Notification]: expectedJobStatus,
|
||||
[QueueName.BackupDatabase]: expectedJobStatus,
|
||||
[QueueName.Ocr]: expectedJobStatus,
|
||||
});
|
||||
});
|
||||
});
|
||||
@ -268,12 +269,12 @@ describe(JobService.name, () => {
|
||||
},
|
||||
{
|
||||
item: { name: JobName.AssetGenerateThumbnails, data: { id: 'asset-1', source: 'upload' } },
|
||||
jobs: [JobName.SmartSearch, JobName.AssetDetectFaces],
|
||||
jobs: [JobName.SmartSearch, JobName.AssetDetectFaces, JobName.Ocr],
|
||||
stub: [assetStub.livePhotoStillAsset],
|
||||
},
|
||||
{
|
||||
item: { name: JobName.AssetGenerateThumbnails, data: { id: 'asset-1', source: 'upload' } },
|
||||
jobs: [JobName.SmartSearch, JobName.AssetDetectFaces, JobName.AssetEncodeVideo],
|
||||
jobs: [JobName.SmartSearch, JobName.AssetDetectFaces, JobName.Ocr, JobName.AssetEncodeVideo],
|
||||
stub: [assetStub.video],
|
||||
},
|
||||
{
|
||||
|
||||
@ -236,6 +236,10 @@ export class JobService extends BaseService {
|
||||
return this.jobRepository.queue({ name: JobName.DatabaseBackup, data: { force } });
|
||||
}
|
||||
|
||||
case QueueName.Ocr: {
|
||||
return this.jobRepository.queue({ name: JobName.OcrQueueAll, data: { force } });
|
||||
}
|
||||
|
||||
default: {
|
||||
throw new BadRequestException(`Invalid job name: ${name}`);
|
||||
}
|
||||
@ -350,6 +354,7 @@ export class JobService extends BaseService {
|
||||
const jobs: JobItem[] = [
|
||||
{ name: JobName.SmartSearch, data: item.data },
|
||||
{ name: JobName.AssetDetectFaces, data: item.data },
|
||||
{ name: JobName.Ocr, data: item.data },
|
||||
];
|
||||
|
||||
if (asset.type === AssetType.Video) {
|
||||
|
||||
177
server/src/services/ocr.service.spec.ts
Normal file
177
server/src/services/ocr.service.spec.ts
Normal file
@ -0,0 +1,177 @@
|
||||
import { AssetVisibility, ImmichWorker, JobName, JobStatus } from 'src/enum';
|
||||
import { OcrService } from 'src/services/ocr.service';
|
||||
import { assetStub } from 'test/fixtures/asset.stub';
|
||||
import { systemConfigStub } from 'test/fixtures/system-config.stub';
|
||||
import { makeStream, newTestService, ServiceMocks } from 'test/utils';
|
||||
|
||||
describe(OcrService.name, () => {
|
||||
let sut: OcrService;
|
||||
let mocks: ServiceMocks;
|
||||
|
||||
beforeEach(() => {
|
||||
({ sut, mocks } = newTestService(OcrService));
|
||||
|
||||
mocks.config.getWorker.mockReturnValue(ImmichWorker.Microservices);
|
||||
});
|
||||
|
||||
it('should work', () => {
|
||||
expect(sut).toBeDefined();
|
||||
});
|
||||
|
||||
describe('handleQueueOcr', () => {
|
||||
it('should do nothing if machine learning is disabled', async () => {
|
||||
mocks.systemMetadata.get.mockResolvedValue(systemConfigStub.machineLearningDisabled);
|
||||
|
||||
await sut.handleQueueOcr({ force: false });
|
||||
|
||||
expect(mocks.database.setDimensionSize).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it('should queue the assets without ocr', async () => {
|
||||
mocks.assetJob.streamForOcrJob.mockReturnValue(makeStream([assetStub.image]));
|
||||
|
||||
await sut.handleQueueOcr({ force: false });
|
||||
|
||||
expect(mocks.job.queueAll).toHaveBeenCalledWith([{ name: JobName.Ocr, data: { id: assetStub.image.id } }]);
|
||||
expect(mocks.assetJob.streamForOcrJob).toHaveBeenCalledWith(false);
|
||||
});
|
||||
|
||||
it('should queue all the assets', async () => {
|
||||
mocks.assetJob.streamForOcrJob.mockReturnValue(makeStream([assetStub.image]));
|
||||
|
||||
await sut.handleQueueOcr({ force: true });
|
||||
|
||||
expect(mocks.job.queueAll).toHaveBeenCalledWith([{ name: JobName.Ocr, data: { id: assetStub.image.id } }]);
|
||||
expect(mocks.assetJob.streamForOcrJob).toHaveBeenCalledWith(true);
|
||||
});
|
||||
});
|
||||
|
||||
describe('handleOcr', () => {
|
||||
it('should do nothing if machine learning is disabled', async () => {
|
||||
mocks.systemMetadata.get.mockResolvedValue(systemConfigStub.machineLearningDisabled);
|
||||
|
||||
expect(await sut.handleOcr({ id: '123' })).toEqual(JobStatus.Skipped);
|
||||
|
||||
expect(mocks.asset.getByIds).not.toHaveBeenCalled();
|
||||
expect(mocks.machineLearning.encodeImage).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it('should skip assets without a resize path', async () => {
|
||||
mocks.assetJob.getForOcr.mockResolvedValue({ visibility: AssetVisibility.Timeline, previewFile: null });
|
||||
|
||||
expect(await sut.handleOcr({ id: assetStub.noResizePath.id })).toEqual(JobStatus.Failed);
|
||||
|
||||
expect(mocks.ocr.upsert).not.toHaveBeenCalled();
|
||||
expect(mocks.machineLearning.ocr).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it('should save the returned objects', async () => {
|
||||
mocks.machineLearning.ocr.mockResolvedValue({
|
||||
box: [10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160],
|
||||
boxScore: [0.9, 0.8],
|
||||
text: ['One Two Three', 'Four Five'],
|
||||
textScore: [0.95, 0.85],
|
||||
});
|
||||
mocks.assetJob.getForOcr.mockResolvedValue({
|
||||
visibility: AssetVisibility.Timeline,
|
||||
previewFile: assetStub.image.files[1].path,
|
||||
});
|
||||
|
||||
expect(await sut.handleOcr({ id: assetStub.image.id })).toEqual(JobStatus.Success);
|
||||
|
||||
expect(mocks.machineLearning.ocr).toHaveBeenCalledWith(
|
||||
'/uploads/user-id/thumbs/path.jpg',
|
||||
expect.objectContaining({
|
||||
modelName: 'PP-OCRv5_mobile',
|
||||
minDetectionScore: 0.5,
|
||||
minRecognitionScore: 0.8,
|
||||
maxResolution: 736,
|
||||
}),
|
||||
);
|
||||
expect(mocks.ocr.upsert).toHaveBeenCalledWith(assetStub.image.id, [
|
||||
{
|
||||
assetId: assetStub.image.id,
|
||||
boxScore: 0.9,
|
||||
text: 'One Two Three',
|
||||
textScore: 0.95,
|
||||
x1: 10,
|
||||
y1: 20,
|
||||
x2: 30,
|
||||
y2: 40,
|
||||
x3: 50,
|
||||
y3: 60,
|
||||
x4: 70,
|
||||
y4: 80,
|
||||
},
|
||||
{
|
||||
assetId: assetStub.image.id,
|
||||
boxScore: 0.8,
|
||||
text: 'Four Five',
|
||||
textScore: 0.85,
|
||||
x1: 90,
|
||||
y1: 100,
|
||||
x2: 110,
|
||||
y2: 120,
|
||||
x3: 130,
|
||||
y3: 140,
|
||||
x4: 150,
|
||||
y4: 160,
|
||||
},
|
||||
]);
|
||||
});
|
||||
|
||||
it('should apply config settings', async () => {
|
||||
mocks.systemMetadata.get.mockResolvedValue({
|
||||
machineLearning: {
|
||||
enabled: true,
|
||||
ocr: {
|
||||
modelName: 'PP-OCRv5_server',
|
||||
enabled: true,
|
||||
minDetectionScore: 0.8,
|
||||
minRecognitionScore: 0.9,
|
||||
maxResolution: 1500,
|
||||
},
|
||||
},
|
||||
});
|
||||
mocks.machineLearning.ocr.mockResolvedValue({ box: [], boxScore: [], text: [], textScore: [] });
|
||||
mocks.assetJob.getForOcr.mockResolvedValue({
|
||||
visibility: AssetVisibility.Timeline,
|
||||
previewFile: assetStub.image.files[1].path,
|
||||
});
|
||||
|
||||
expect(await sut.handleOcr({ id: assetStub.image.id })).toEqual(JobStatus.Success);
|
||||
|
||||
expect(mocks.machineLearning.ocr).toHaveBeenCalledWith(
|
||||
'/uploads/user-id/thumbs/path.jpg',
|
||||
expect.objectContaining({
|
||||
modelName: 'PP-OCRv5_server',
|
||||
minDetectionScore: 0.8,
|
||||
minRecognitionScore: 0.9,
|
||||
maxResolution: 1500,
|
||||
}),
|
||||
);
|
||||
expect(mocks.ocr.upsert).toHaveBeenCalledWith(assetStub.image.id, []);
|
||||
});
|
||||
|
||||
it('should skip invisible assets', async () => {
|
||||
mocks.assetJob.getForOcr.mockResolvedValue({
|
||||
visibility: AssetVisibility.Hidden,
|
||||
previewFile: assetStub.image.files[1].path,
|
||||
});
|
||||
|
||||
expect(await sut.handleOcr({ id: assetStub.livePhotoMotionAsset.id })).toEqual(JobStatus.Skipped);
|
||||
|
||||
expect(mocks.machineLearning.ocr).not.toHaveBeenCalled();
|
||||
expect(mocks.ocr.upsert).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
it('should fail if asset could not be found', async () => {
|
||||
mocks.assetJob.getForOcr.mockResolvedValue(void 0);
|
||||
|
||||
expect(await sut.handleOcr({ id: assetStub.image.id })).toEqual(JobStatus.Failed);
|
||||
|
||||
expect(mocks.machineLearning.ocr).not.toHaveBeenCalled();
|
||||
expect(mocks.ocr.upsert).not.toHaveBeenCalled();
|
||||
});
|
||||
});
|
||||
});
|
||||
86
server/src/services/ocr.service.ts
Normal file
86
server/src/services/ocr.service.ts
Normal file
@ -0,0 +1,86 @@
|
||||
import { Injectable } from '@nestjs/common';
|
||||
import { JOBS_ASSET_PAGINATION_SIZE } from 'src/constants';
|
||||
import { OnJob } from 'src/decorators';
|
||||
import { AssetVisibility, JobName, JobStatus, QueueName } from 'src/enum';
|
||||
import { OCR } from 'src/repositories/machine-learning.repository';
|
||||
import { BaseService } from 'src/services/base.service';
|
||||
import { JobItem, JobOf } from 'src/types';
|
||||
import { isOcrEnabled } from 'src/utils/misc';
|
||||
|
||||
@Injectable()
|
||||
export class OcrService extends BaseService {
|
||||
@OnJob({ name: JobName.OcrQueueAll, queue: QueueName.Ocr })
|
||||
async handleQueueOcr({ force }: JobOf<JobName.OcrQueueAll>): Promise<JobStatus> {
|
||||
const { machineLearning } = await this.getConfig({ withCache: false });
|
||||
if (!isOcrEnabled(machineLearning)) {
|
||||
return JobStatus.Skipped;
|
||||
}
|
||||
|
||||
if (force) {
|
||||
await this.ocrRepository.deleteAll();
|
||||
}
|
||||
|
||||
let jobs: JobItem[] = [];
|
||||
const assets = this.assetJobRepository.streamForOcrJob(force);
|
||||
|
||||
for await (const asset of assets) {
|
||||
jobs.push({ name: JobName.Ocr, data: { id: asset.id } });
|
||||
|
||||
if (jobs.length >= JOBS_ASSET_PAGINATION_SIZE) {
|
||||
await this.jobRepository.queueAll(jobs);
|
||||
jobs = [];
|
||||
}
|
||||
}
|
||||
|
||||
await this.jobRepository.queueAll(jobs);
|
||||
return JobStatus.Success;
|
||||
}
|
||||
|
||||
@OnJob({ name: JobName.Ocr, queue: QueueName.Ocr })
|
||||
async handleOcr({ id }: JobOf<JobName.Ocr>): Promise<JobStatus> {
|
||||
const { machineLearning } = await this.getConfig({ withCache: true });
|
||||
if (!isOcrEnabled(machineLearning)) {
|
||||
return JobStatus.Skipped;
|
||||
}
|
||||
|
||||
const asset = await this.assetJobRepository.getForOcr(id);
|
||||
if (!asset || !asset.previewFile) {
|
||||
return JobStatus.Failed;
|
||||
}
|
||||
|
||||
if (asset.visibility === AssetVisibility.Hidden) {
|
||||
return JobStatus.Skipped;
|
||||
}
|
||||
|
||||
const ocrResults = await this.machineLearningRepository.ocr(asset.previewFile, machineLearning.ocr);
|
||||
|
||||
await this.ocrRepository.upsert(id, this.parseOcrResults(id, ocrResults));
|
||||
|
||||
await this.assetRepository.upsertJobStatus({ assetId: id, ocrAt: new Date() });
|
||||
|
||||
this.logger.debug(`Processed ${ocrResults.text.length} OCR result(s) for ${id}`);
|
||||
return JobStatus.Success;
|
||||
}
|
||||
|
||||
private parseOcrResults(id: string, { box, boxScore, text, textScore }: OCR) {
|
||||
const ocrDataList = [];
|
||||
for (let i = 0; i < text.length; i++) {
|
||||
const boxOffset = i * 8;
|
||||
ocrDataList.push({
|
||||
assetId: id,
|
||||
x1: box[boxOffset],
|
||||
y1: box[boxOffset + 1],
|
||||
x2: box[boxOffset + 2],
|
||||
y2: box[boxOffset + 3],
|
||||
x3: box[boxOffset + 4],
|
||||
y3: box[boxOffset + 5],
|
||||
x4: box[boxOffset + 6],
|
||||
y4: box[boxOffset + 7],
|
||||
boxScore: boxScore[i],
|
||||
textScore: textScore[i],
|
||||
text: text[i],
|
||||
});
|
||||
}
|
||||
return ocrDataList;
|
||||
}
|
||||
}
|
||||
@ -141,6 +141,7 @@ describe(ServerService.name, () => {
|
||||
reverseGeocoding: true,
|
||||
oauth: false,
|
||||
oauthAutoLaunch: false,
|
||||
ocr: true,
|
||||
passwordLogin: true,
|
||||
search: true,
|
||||
sidecar: true,
|
||||
|
||||
@ -19,7 +19,12 @@ import { UserStatsQueryResponse } from 'src/repositories/user.repository';
|
||||
import { BaseService } from 'src/services/base.service';
|
||||
import { asHumanReadable } from 'src/utils/bytes';
|
||||
import { mimeTypes } from 'src/utils/mime-types';
|
||||
import { isDuplicateDetectionEnabled, isFacialRecognitionEnabled, isSmartSearchEnabled } from 'src/utils/misc';
|
||||
import {
|
||||
isDuplicateDetectionEnabled,
|
||||
isFacialRecognitionEnabled,
|
||||
isOcrEnabled,
|
||||
isSmartSearchEnabled,
|
||||
} from 'src/utils/misc';
|
||||
|
||||
@Injectable()
|
||||
export class ServerService extends BaseService {
|
||||
@ -97,6 +102,7 @@ export class ServerService extends BaseService {
|
||||
trash: trash.enabled,
|
||||
oauth: oauth.enabled,
|
||||
oauthAutoLaunch: oauth.autoLaunch,
|
||||
ocr: isOcrEnabled(machineLearning),
|
||||
passwordLogin: passwordLogin.enabled,
|
||||
configFile: !!configFile,
|
||||
email: notifications.smtp.enabled,
|
||||
|
||||
@ -39,6 +39,7 @@ const updatedConfig = Object.freeze<SystemConfig>({
|
||||
[QueueName.ThumbnailGeneration]: { concurrency: 3 },
|
||||
[QueueName.VideoConversion]: { concurrency: 1 },
|
||||
[QueueName.Notification]: { concurrency: 5 },
|
||||
[QueueName.Ocr]: { concurrency: 1 },
|
||||
},
|
||||
backup: {
|
||||
database: {
|
||||
@ -102,6 +103,13 @@ const updatedConfig = Object.freeze<SystemConfig>({
|
||||
maxDistance: 0.5,
|
||||
minFaces: 3,
|
||||
},
|
||||
ocr: {
|
||||
enabled: true,
|
||||
modelName: 'PP-OCRv5_mobile',
|
||||
minDetectionScore: 0.5,
|
||||
minRecognitionScore: 0.8,
|
||||
maxResolution: 736,
|
||||
},
|
||||
},
|
||||
map: {
|
||||
enabled: true,
|
||||
|
||||
@ -322,7 +322,8 @@ export type ColumnType =
|
||||
| 'uuid'
|
||||
| 'vector'
|
||||
| 'enum'
|
||||
| 'serial';
|
||||
| 'serial'
|
||||
| 'real';
|
||||
|
||||
export type DatabaseSchema = {
|
||||
databaseName: string;
|
||||
|
||||
@ -370,7 +370,11 @@ export type JobItem =
|
||||
| { name: JobName.NotifyUserSignup; data: INotifySignupJob }
|
||||
|
||||
// Version check
|
||||
| { name: JobName.VersionCheck; data: IBaseJob };
|
||||
| { name: JobName.VersionCheck; data: IBaseJob }
|
||||
|
||||
// OCR
|
||||
| { name: JobName.OcrQueueAll; data: IBaseJob }
|
||||
| { name: JobName.Ocr; data: IEntityJob };
|
||||
|
||||
export type VectorExtension = (typeof VECTOR_EXTENSIONS)[number];
|
||||
|
||||
|
||||
@ -200,6 +200,14 @@ export function withFiles(eb: ExpressionBuilder<DB, 'asset'>, type?: AssetFileTy
|
||||
).as('files');
|
||||
}
|
||||
|
||||
export function withFilePath(eb: ExpressionBuilder<DB, 'asset'>, type: AssetFileType) {
|
||||
return eb
|
||||
.selectFrom('asset_file')
|
||||
.select('asset_file.path')
|
||||
.whereRef('asset_file.assetId', '=', 'asset.id')
|
||||
.where('asset_file.type', '=', type);
|
||||
}
|
||||
|
||||
export function withFacesAndPeople(eb: ExpressionBuilder<DB, 'asset'>, withDeletedFace?: boolean) {
|
||||
return jsonArrayFrom(
|
||||
eb
|
||||
@ -380,6 +388,11 @@ export function searchAssetBuilder(kysely: Kysely<DB>, options: AssetSearchBuild
|
||||
.innerJoin('asset_exif', 'asset.id', 'asset_exif.assetId')
|
||||
.where(sql`f_unaccent(asset_exif.description)`, 'ilike', sql`'%' || f_unaccent(${options.description}) || '%'`),
|
||||
)
|
||||
.$if(!!options.ocr, (qb) =>
|
||||
qb
|
||||
.innerJoin('ocr_search', 'asset.id', 'ocr_search.assetId')
|
||||
.where(() => sql`f_unaccent(ocr_search.text) %>> f_unaccent(${options.ocr!})`),
|
||||
)
|
||||
.$if(!!options.type, (qb) => qb.where('asset.type', '=', options.type!))
|
||||
.$if(options.isFavorite !== undefined, (qb) => qb.where('asset.isFavorite', '=', options.isFavorite!))
|
||||
.$if(options.isOffline !== undefined, (qb) => qb.where('asset.isOffline', '=', options.isOffline!))
|
||||
|
||||
@ -95,6 +95,8 @@ export const unsetDeep = (object: unknown, key: string) => {
|
||||
const isMachineLearningEnabled = (machineLearning: SystemConfig['machineLearning']) => machineLearning.enabled;
|
||||
export const isSmartSearchEnabled = (machineLearning: SystemConfig['machineLearning']) =>
|
||||
isMachineLearningEnabled(machineLearning) && machineLearning.clip.enabled;
|
||||
export const isOcrEnabled = (machineLearning: SystemConfig['machineLearning']) =>
|
||||
isMachineLearningEnabled(machineLearning) && machineLearning.ocr.enabled;
|
||||
export const isFacialRecognitionEnabled = (machineLearning: SystemConfig['machineLearning']) =>
|
||||
isMachineLearningEnabled(machineLearning) && machineLearning.facialRecognition.enabled;
|
||||
export const isDuplicateDetectionEnabled = (machineLearning: SystemConfig['machineLearning']) =>
|
||||
|
||||
@ -27,8 +27,10 @@ import { EmailRepository } from 'src/repositories/email.repository';
|
||||
import { EventRepository } from 'src/repositories/event.repository';
|
||||
import { JobRepository } from 'src/repositories/job.repository';
|
||||
import { LoggingRepository } from 'src/repositories/logging.repository';
|
||||
import { MachineLearningRepository } from 'src/repositories/machine-learning.repository';
|
||||
import { MemoryRepository } from 'src/repositories/memory.repository';
|
||||
import { NotificationRepository } from 'src/repositories/notification.repository';
|
||||
import { OcrRepository } from 'src/repositories/ocr.repository';
|
||||
import { PartnerRepository } from 'src/repositories/partner.repository';
|
||||
import { PersonRepository } from 'src/repositories/person.repository';
|
||||
import { SearchRepository } from 'src/repositories/search.repository';
|
||||
@ -47,6 +49,7 @@ import { VersionHistoryRepository } from 'src/repositories/version-history.repos
|
||||
import { DB } from 'src/schema';
|
||||
import { AlbumTable } from 'src/schema/tables/album.table';
|
||||
import { AssetExifTable } from 'src/schema/tables/asset-exif.table';
|
||||
import { AssetFileTable } from 'src/schema/tables/asset-file.table';
|
||||
import { AssetJobStatusTable } from 'src/schema/tables/asset-job-status.table';
|
||||
import { AssetTable } from 'src/schema/tables/asset.table';
|
||||
import { FaceSearchTable } from 'src/schema/tables/face-search.table';
|
||||
@ -169,6 +172,11 @@ export class MediumTestContext<S extends BaseService = BaseService> {
|
||||
return { asset, result };
|
||||
}
|
||||
|
||||
async newAssetFile(dto: Insertable<AssetFileTable>) {
|
||||
const result = await this.get(AssetRepository).upsertFile(dto);
|
||||
return { result };
|
||||
}
|
||||
|
||||
async newAssetFace(dto: Partial<Insertable<AssetFace>> & { assetId: string }) {
|
||||
const assetFace = mediumFactory.assetFaceInsert(dto);
|
||||
const result = await this.get(PersonRepository).createAssetFace(assetFace);
|
||||
@ -307,6 +315,7 @@ const newRealRepository = <T>(key: ClassConstructor<T>, db: Kysely<DB>): T => {
|
||||
case AssetJobRepository:
|
||||
case MemoryRepository:
|
||||
case NotificationRepository:
|
||||
case OcrRepository:
|
||||
case PartnerRepository:
|
||||
case PersonRepository:
|
||||
case SearchRepository:
|
||||
@ -359,6 +368,7 @@ const newMockRepository = <T>(key: ClassConstructor<T>) => {
|
||||
case CryptoRepository:
|
||||
case MemoryRepository:
|
||||
case NotificationRepository:
|
||||
case OcrRepository:
|
||||
case PartnerRepository:
|
||||
case PersonRepository:
|
||||
case SessionRepository:
|
||||
@ -407,6 +417,10 @@ const newMockRepository = <T>(key: ClassConstructor<T>) => {
|
||||
return automock(LoggingRepository, { args: [undefined, configMock], strict: false });
|
||||
}
|
||||
|
||||
case MachineLearningRepository: {
|
||||
return automock(MachineLearningRepository, { args: [{ setContext: () => {} }] });
|
||||
}
|
||||
|
||||
case StorageRepository: {
|
||||
return automock(StorageRepository, { args: [{ setContext: () => {} }] });
|
||||
}
|
||||
|
||||
243
server/test/medium/specs/services/ocr.service.spec.ts
Normal file
243
server/test/medium/specs/services/ocr.service.spec.ts
Normal file
@ -0,0 +1,243 @@
|
||||
import { Kysely } from 'kysely';
|
||||
import { AssetFileType, JobStatus } from 'src/enum';
|
||||
import { AssetJobRepository } from 'src/repositories/asset-job.repository';
|
||||
import { AssetRepository } from 'src/repositories/asset.repository';
|
||||
import { ConfigRepository } from 'src/repositories/config.repository';
|
||||
import { JobRepository } from 'src/repositories/job.repository';
|
||||
import { LoggingRepository } from 'src/repositories/logging.repository';
|
||||
import { MachineLearningRepository } from 'src/repositories/machine-learning.repository';
|
||||
import { OcrRepository } from 'src/repositories/ocr.repository';
|
||||
import { SystemMetadataRepository } from 'src/repositories/system-metadata.repository';
|
||||
import { DB } from 'src/schema';
|
||||
import { OcrService } from 'src/services/ocr.service';
|
||||
import { newMediumService } from 'test/medium.factory';
|
||||
import { getKyselyDB } from 'test/utils';
|
||||
|
||||
let defaultDatabase: Kysely<DB>;
|
||||
|
||||
const setup = (db?: Kysely<DB>) => {
|
||||
return newMediumService(OcrService, {
|
||||
database: db || defaultDatabase,
|
||||
real: [AssetRepository, AssetJobRepository, ConfigRepository, OcrRepository, SystemMetadataRepository],
|
||||
mock: [JobRepository, LoggingRepository, MachineLearningRepository],
|
||||
});
|
||||
};
|
||||
|
||||
beforeAll(async () => {
|
||||
defaultDatabase = await getKyselyDB();
|
||||
});
|
||||
|
||||
describe(OcrService.name, () => {
|
||||
it('should work', () => {
|
||||
const { sut } = setup();
|
||||
expect(sut).toBeDefined();
|
||||
});
|
||||
|
||||
it('should parse asset', async () => {
|
||||
const { sut, ctx } = setup();
|
||||
const { user } = await ctx.newUser();
|
||||
const { asset } = await ctx.newAsset({ ownerId: user.id });
|
||||
await ctx.newAssetFile({ assetId: asset.id, type: AssetFileType.Preview, path: 'preview.jpg' });
|
||||
|
||||
const machineLearningMock = ctx.getMock(MachineLearningRepository);
|
||||
machineLearningMock.ocr.mockResolvedValue({
|
||||
box: [10, 10, 50, 10, 50, 50, 10, 50],
|
||||
boxScore: [0.99],
|
||||
text: ['Test OCR'],
|
||||
textScore: [0.95],
|
||||
});
|
||||
|
||||
await expect(sut.handleOcr({ id: asset.id })).resolves.toBe(JobStatus.Success);
|
||||
|
||||
const ocrRepository = ctx.get(OcrRepository);
|
||||
await expect(ocrRepository.getByAssetId(asset.id)).resolves.toEqual([
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.99,
|
||||
id: expect.any(String),
|
||||
text: 'Test OCR',
|
||||
textScore: 0.95,
|
||||
x1: 10,
|
||||
y1: 10,
|
||||
x2: 50,
|
||||
y2: 10,
|
||||
x3: 50,
|
||||
y3: 50,
|
||||
x4: 10,
|
||||
y4: 50,
|
||||
},
|
||||
]);
|
||||
await expect(
|
||||
ctx.database.selectFrom('ocr_search').selectAll().where('assetId', '=', asset.id).executeTakeFirst(),
|
||||
).resolves.toEqual({
|
||||
assetId: asset.id,
|
||||
text: 'Test OCR',
|
||||
});
|
||||
await expect(
|
||||
ctx.database
|
||||
.selectFrom('asset_job_status')
|
||||
.select('asset_job_status.ocrAt')
|
||||
.where('assetId', '=', asset.id)
|
||||
.executeTakeFirst(),
|
||||
).resolves.toEqual({ ocrAt: expect.any(Date) });
|
||||
});
|
||||
|
||||
it('should handle multiple boxes', async () => {
|
||||
const { sut, ctx } = setup();
|
||||
const { user } = await ctx.newUser();
|
||||
const { asset } = await ctx.newAsset({ ownerId: user.id });
|
||||
await ctx.newAssetFile({ assetId: asset.id, type: AssetFileType.Preview, path: 'preview.jpg' });
|
||||
|
||||
const machineLearningMock = ctx.getMock(MachineLearningRepository);
|
||||
machineLearningMock.ocr.mockResolvedValue({
|
||||
box: Array.from({ length: 8 * 5 }, (_, i) => i),
|
||||
boxScore: [0.7, 0.67, 0.65, 0.62, 0.6],
|
||||
text: ['One', 'Two', 'Three', 'Four', 'Five'],
|
||||
textScore: [0.9, 0.89, 0.88, 0.87, 0.86],
|
||||
});
|
||||
|
||||
await expect(sut.handleOcr({ id: asset.id })).resolves.toBe(JobStatus.Success);
|
||||
|
||||
const ocrRepository = ctx.get(OcrRepository);
|
||||
await expect(ocrRepository.getByAssetId(asset.id)).resolves.toEqual([
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.7,
|
||||
id: expect.any(String),
|
||||
text: 'One',
|
||||
textScore: 0.9,
|
||||
x1: 0,
|
||||
y1: 1,
|
||||
x2: 2,
|
||||
y2: 3,
|
||||
x3: 4,
|
||||
y3: 5,
|
||||
x4: 6,
|
||||
y4: 7,
|
||||
},
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.67,
|
||||
id: expect.any(String),
|
||||
text: 'Two',
|
||||
textScore: 0.89,
|
||||
x1: 8,
|
||||
y1: 9,
|
||||
x2: 10,
|
||||
y2: 11,
|
||||
x3: 12,
|
||||
y3: 13,
|
||||
x4: 14,
|
||||
y4: 15,
|
||||
},
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.65,
|
||||
id: expect.any(String),
|
||||
text: 'Three',
|
||||
textScore: 0.88,
|
||||
x1: 16,
|
||||
y1: 17,
|
||||
x2: 18,
|
||||
y2: 19,
|
||||
x3: 20,
|
||||
y3: 21,
|
||||
x4: 22,
|
||||
y4: 23,
|
||||
},
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.62,
|
||||
id: expect.any(String),
|
||||
text: 'Four',
|
||||
textScore: 0.87,
|
||||
x1: 24,
|
||||
y1: 25,
|
||||
x2: 26,
|
||||
y2: 27,
|
||||
x3: 28,
|
||||
y3: 29,
|
||||
x4: 30,
|
||||
y4: 31,
|
||||
},
|
||||
{
|
||||
assetId: asset.id,
|
||||
boxScore: 0.6,
|
||||
id: expect.any(String),
|
||||
text: 'Five',
|
||||
textScore: 0.86,
|
||||
x1: 32,
|
||||
y1: 33,
|
||||
x2: 34,
|
||||
y2: 35,
|
||||
x3: 36,
|
||||
y3: 37,
|
||||
x4: 38,
|
||||
y4: 39,
|
||||
},
|
||||
]);
|
||||
await expect(
|
||||
ctx.database.selectFrom('ocr_search').selectAll().where('assetId', '=', asset.id).executeTakeFirst(),
|
||||
).resolves.toEqual({
|
||||
assetId: asset.id,
|
||||
text: 'One Two Three Four Five',
|
||||
});
|
||||
await expect(
|
||||
ctx.database
|
||||
.selectFrom('asset_job_status')
|
||||
.select('asset_job_status.ocrAt')
|
||||
.where('assetId', '=', asset.id)
|
||||
.executeTakeFirst(),
|
||||
).resolves.toEqual({ ocrAt: expect.any(Date) });
|
||||
});
|
||||
|
||||
it('should handle no boxes', async () => {
|
||||
const { sut, ctx } = setup();
|
||||
const { user } = await ctx.newUser();
|
||||
const { asset } = await ctx.newAsset({ ownerId: user.id });
|
||||
await ctx.newAssetFile({ assetId: asset.id, type: AssetFileType.Preview, path: 'preview.jpg' });
|
||||
|
||||
const machineLearningMock = ctx.getMock(MachineLearningRepository);
|
||||
machineLearningMock.ocr.mockResolvedValue({ box: [], boxScore: [], text: [], textScore: [] });
|
||||
|
||||
await expect(sut.handleOcr({ id: asset.id })).resolves.toBe(JobStatus.Success);
|
||||
|
||||
const ocrRepository = ctx.get(OcrRepository);
|
||||
await expect(ocrRepository.getByAssetId(asset.id)).resolves.toEqual([]);
|
||||
await expect(
|
||||
ctx.database.selectFrom('ocr_search').selectAll().where('assetId', '=', asset.id).executeTakeFirst(),
|
||||
).resolves.toBeUndefined();
|
||||
await expect(
|
||||
ctx.database
|
||||
.selectFrom('asset_job_status')
|
||||
.select('asset_job_status.ocrAt')
|
||||
.where('assetId', '=', asset.id)
|
||||
.executeTakeFirst(),
|
||||
).resolves.toEqual({ ocrAt: expect.any(Date) });
|
||||
});
|
||||
|
||||
it('should update existing results', async () => {
|
||||
const { sut, ctx } = setup();
|
||||
const { user } = await ctx.newUser();
|
||||
const { asset } = await ctx.newAsset({ ownerId: user.id });
|
||||
await ctx.newAssetFile({ assetId: asset.id, type: AssetFileType.Preview, path: 'preview.jpg' });
|
||||
|
||||
const machineLearningMock = ctx.getMock(MachineLearningRepository);
|
||||
machineLearningMock.ocr.mockResolvedValue({
|
||||
box: [10, 10, 50, 10, 50, 50, 10, 50],
|
||||
boxScore: [0.99],
|
||||
text: ['Test OCR'],
|
||||
textScore: [0.95],
|
||||
});
|
||||
await expect(sut.handleOcr({ id: asset.id })).resolves.toBe(JobStatus.Success);
|
||||
|
||||
machineLearningMock.ocr.mockResolvedValue({ box: [], boxScore: [], text: [], textScore: [] });
|
||||
await expect(sut.handleOcr({ id: asset.id })).resolves.toBe(JobStatus.Success);
|
||||
|
||||
const ocrRepository = ctx.get(OcrRepository);
|
||||
await expect(ocrRepository.getByAssetId(asset.id)).resolves.toEqual([]);
|
||||
await expect(
|
||||
ctx.database.selectFrom('ocr_search').selectAll().where('assetId', '=', asset.id).executeTakeFirst(),
|
||||
).resolves.toBeUndefined();
|
||||
});
|
||||
});
|
||||
@ -41,6 +41,7 @@ import { MetadataRepository } from 'src/repositories/metadata.repository';
|
||||
import { MoveRepository } from 'src/repositories/move.repository';
|
||||
import { NotificationRepository } from 'src/repositories/notification.repository';
|
||||
import { OAuthRepository } from 'src/repositories/oauth.repository';
|
||||
import { OcrRepository } from 'src/repositories/ocr.repository';
|
||||
import { PartnerRepository } from 'src/repositories/partner.repository';
|
||||
import { PersonRepository } from 'src/repositories/person.repository';
|
||||
import { ProcessRepository } from 'src/repositories/process.repository';
|
||||
@ -230,6 +231,7 @@ export type ServiceOverrides = {
|
||||
metadata: MetadataRepository;
|
||||
move: MoveRepository;
|
||||
notification: NotificationRepository;
|
||||
ocr: OcrRepository;
|
||||
oauth: OAuthRepository;
|
||||
partner: PartnerRepository;
|
||||
person: PersonRepository;
|
||||
@ -302,6 +304,7 @@ export const newTestService = <T extends BaseService>(
|
||||
metadata: newMetadataRepositoryMock(),
|
||||
move: automock(MoveRepository, { strict: false }),
|
||||
notification: automock(NotificationRepository),
|
||||
ocr: automock(OcrRepository, { strict: false }),
|
||||
oauth: automock(OAuthRepository, { args: [loggerMock] }),
|
||||
partner: automock(PartnerRepository, { strict: false }),
|
||||
person: automock(PersonRepository, { strict: false }),
|
||||
@ -357,6 +360,7 @@ export const newTestService = <T extends BaseService>(
|
||||
overrides.move || (mocks.move as As<MoveRepository>),
|
||||
overrides.notification || (mocks.notification as As<NotificationRepository>),
|
||||
overrides.oauth || (mocks.oauth as As<OAuthRepository>),
|
||||
overrides.ocr || (mocks.ocr as As<OcrRepository>),
|
||||
overrides.partner || (mocks.partner as As<PartnerRepository>),
|
||||
overrides.person || (mocks.person as As<PersonRepository>),
|
||||
overrides.process || (mocks.process as As<ProcessRepository>),
|
||||
|
||||
@ -31,6 +31,7 @@
|
||||
JobName.VideoConversion,
|
||||
JobName.StorageTemplateMigration,
|
||||
JobName.Migration,
|
||||
JobName.Ocr,
|
||||
];
|
||||
|
||||
// eslint-disable-next-line @typescript-eslint/no-explicit-any
|
||||
|
||||
@ -254,6 +254,71 @@
|
||||
</div>
|
||||
</SettingAccordion>
|
||||
|
||||
<SettingAccordion
|
||||
key="ocr"
|
||||
title={$t('admin.machine_learning_ocr')}
|
||||
subtitle={$t('admin.machine_learning_ocr_description')}
|
||||
>
|
||||
<div class="ml-4 mt-4 flex flex-col gap-4">
|
||||
<SettingSwitch
|
||||
title={$t('admin.machine_learning_ocr_enabled')}
|
||||
subtitle={$t('admin.machine_learning_ocr_enabled_description')}
|
||||
bind:checked={config.machineLearning.ocr.enabled}
|
||||
disabled={disabled || !config.machineLearning.enabled}
|
||||
/>
|
||||
|
||||
<hr />
|
||||
|
||||
<SettingSelect
|
||||
label={$t('admin.machine_learning_ocr_model')}
|
||||
desc={$t('admin.machine_learning_ocr_model_description')}
|
||||
name="ocr-model"
|
||||
bind:value={config.machineLearning.ocr.modelName}
|
||||
options={[
|
||||
{ value: 'PP-OCRv5_server', text: 'PP-OCRv5_server' },
|
||||
{ value: 'PP-OCRv5_mobile', text: 'PP-OCRv5_mobile' },
|
||||
]}
|
||||
disabled={disabled || !config.machineLearning.enabled || !config.machineLearning.ocr.enabled}
|
||||
isEdited={config.machineLearning.ocr.modelName !== savedConfig.machineLearning.ocr.modelName}
|
||||
/>
|
||||
|
||||
<SettingInputField
|
||||
inputType={SettingInputFieldType.NUMBER}
|
||||
label={$t('admin.machine_learning_ocr_min_detection_score')}
|
||||
description={$t('admin.machine_learning_ocr_min_detection_score_description')}
|
||||
bind:value={config.machineLearning.ocr.minDetectionScore}
|
||||
step="0.1"
|
||||
min={0.1}
|
||||
max={1}
|
||||
disabled={disabled || !config.machineLearning.enabled || !config.machineLearning.ocr.enabled}
|
||||
isEdited={config.machineLearning.ocr.minDetectionScore !==
|
||||
savedConfig.machineLearning.ocr.minDetectionScore}
|
||||
/>
|
||||
|
||||
<SettingInputField
|
||||
inputType={SettingInputFieldType.NUMBER}
|
||||
label={$t('admin.machine_learning_ocr_min_recognition_score')}
|
||||
description={$t('admin.machine_learning_ocr_min_score_recognition_description')}
|
||||
bind:value={config.machineLearning.ocr.minRecognitionScore}
|
||||
step="0.1"
|
||||
min={0.1}
|
||||
max={1}
|
||||
disabled={disabled || !config.machineLearning.enabled || !config.machineLearning.ocr.enabled}
|
||||
isEdited={config.machineLearning.ocr.minRecognitionScore !==
|
||||
savedConfig.machineLearning.ocr.minRecognitionScore}
|
||||
/>
|
||||
|
||||
<SettingInputField
|
||||
inputType={SettingInputFieldType.NUMBER}
|
||||
label={$t('admin.machine_learning_ocr_max_resolution')}
|
||||
description={$t('admin.machine_learning_ocr_max_resolution_description')}
|
||||
bind:value={config.machineLearning.ocr.maxResolution}
|
||||
min={1}
|
||||
disabled={disabled || !config.machineLearning.enabled || !config.machineLearning.ocr.enabled}
|
||||
isEdited={config.machineLearning.ocr.maxResolution !== savedConfig.machineLearning.ocr.maxResolution}
|
||||
/>
|
||||
</div>
|
||||
</SettingAccordion>
|
||||
<SettingButtonsRow
|
||||
onReset={(options) => onReset({ ...options, configKeys: ['machineLearning'] })}
|
||||
onSave={() => onSave({ machineLearning: config.machineLearning })}
|
||||
|
||||
@ -19,6 +19,7 @@
|
||||
mdiTable,
|
||||
mdiTagFaces,
|
||||
mdiVideo,
|
||||
mdiOcr,
|
||||
} from '@mdi/js';
|
||||
import type { Component } from 'svelte';
|
||||
import { t } from 'svelte-i18n';
|
||||
@ -124,6 +125,14 @@
|
||||
handleCommand: handleConfirmCommand,
|
||||
disabled: !$featureFlags.facialRecognition,
|
||||
},
|
||||
[JobName.Ocr]: {
|
||||
icon: mdiOcr,
|
||||
title: $getJobName(JobName.Ocr),
|
||||
subtitle: $t('admin.ocr_job_description'),
|
||||
allText: $t('all'),
|
||||
missingText: $t('missing'),
|
||||
disabled: !$featureFlags.ocr,
|
||||
},
|
||||
[JobName.VideoConversion]: {
|
||||
icon: mdiVideo,
|
||||
title: $getJobName(JobName.VideoConversion),
|
||||
|
||||
@ -107,7 +107,7 @@
|
||||
|
||||
const onSubmit = () => {
|
||||
const searchType = getSearchType();
|
||||
let payload: SmartSearchDto | MetadataSearchDto = {} as SmartSearchDto | MetadataSearchDto;
|
||||
let payload = {} as SmartSearchDto | MetadataSearchDto;
|
||||
|
||||
switch (searchType) {
|
||||
case 'smart': {
|
||||
@ -122,6 +122,10 @@
|
||||
payload = { description: value } as MetadataSearchDto;
|
||||
break;
|
||||
}
|
||||
case 'ocr': {
|
||||
payload = { ocr: value } as MetadataSearchDto;
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
handlePromiseError(handleSearch(payload));
|
||||
@ -171,17 +175,14 @@
|
||||
onSubmit();
|
||||
};
|
||||
|
||||
function getSearchType(): 'smart' | 'metadata' | 'description' {
|
||||
function getSearchType() {
|
||||
const searchType = localStorage.getItem('searchQueryType');
|
||||
switch (searchType) {
|
||||
case 'smart': {
|
||||
return 'smart';
|
||||
}
|
||||
case 'metadata': {
|
||||
return 'metadata';
|
||||
}
|
||||
case 'description': {
|
||||
return 'description';
|
||||
case 'smart':
|
||||
case 'metadata':
|
||||
case 'description':
|
||||
case 'ocr': {
|
||||
return searchType;
|
||||
}
|
||||
default: {
|
||||
return 'smart';
|
||||
@ -201,6 +202,9 @@
|
||||
case 'description': {
|
||||
return $t('description');
|
||||
}
|
||||
case 'ocr': {
|
||||
return $t('ocr');
|
||||
}
|
||||
}
|
||||
}
|
||||
</script>
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
|
||||
interface Props {
|
||||
query: string | undefined;
|
||||
queryType?: 'smart' | 'metadata' | 'description';
|
||||
queryType?: 'smart' | 'metadata' | 'description' | 'ocr';
|
||||
}
|
||||
|
||||
let { query = $bindable(), queryType = $bindable('smart') }: Props = $props();
|
||||
@ -28,6 +28,7 @@
|
||||
bind:group={queryType}
|
||||
value="description"
|
||||
/>
|
||||
<RadioButton name="query-type" id="ocr-radio" label={$t('ocr')} bind:group={queryType} value="ocr" />
|
||||
</div>
|
||||
</fieldset>
|
||||
|
||||
@ -63,4 +64,15 @@
|
||||
bind:value={query}
|
||||
aria-labelledby="description-label"
|
||||
/>
|
||||
{:else if queryType === 'ocr'}
|
||||
<label for="ocr-input" class="immich-form-label">{$t('search_by_ocr')}</label>
|
||||
<input
|
||||
class="immich-form-input hover:cursor-text w-full !mt-1"
|
||||
type="text"
|
||||
id="ocr-input"
|
||||
name="ocr"
|
||||
placeholder={$t('search_by_ocr_example')}
|
||||
bind:value={query}
|
||||
aria-labelledby="ocr-label"
|
||||
/>
|
||||
{/if}
|
||||
|
||||
@ -138,9 +138,10 @@ export enum QueryType {
|
||||
SMART = 'smart',
|
||||
METADATA = 'metadata',
|
||||
DESCRIPTION = 'description',
|
||||
OCR = 'ocr',
|
||||
}
|
||||
|
||||
export const validQueryTypes = new Set([QueryType.SMART, QueryType.METADATA, QueryType.DESCRIPTION]);
|
||||
export const validQueryTypes = new Set([QueryType.SMART, QueryType.METADATA, QueryType.DESCRIPTION, QueryType.OCR]);
|
||||
|
||||
export const locales = [
|
||||
{ code: 'af-ZA', name: 'Afrikaans (South Africa)' },
|
||||
|
||||
@ -6,7 +6,8 @@
|
||||
|
||||
export type SearchFilter = {
|
||||
query: string;
|
||||
queryType: 'smart' | 'metadata' | 'description';
|
||||
ocr?: string;
|
||||
queryType: 'smart' | 'metadata' | 'description' | 'ocr';
|
||||
personIds: SvelteSet<string>;
|
||||
tagIds: SvelteSet<string> | null;
|
||||
location: SearchLocationFilter;
|
||||
@ -74,6 +75,7 @@
|
||||
|
||||
let filter: SearchFilter = $state({
|
||||
query,
|
||||
ocr: searchQuery.ocr,
|
||||
queryType: defaultQueryType(),
|
||||
personIds: new SvelteSet('personIds' in searchQuery ? searchQuery.personIds : []),
|
||||
tagIds:
|
||||
@ -113,6 +115,7 @@
|
||||
const resetForm = () => {
|
||||
filter = {
|
||||
query: '',
|
||||
ocr: undefined,
|
||||
queryType: defaultQueryType(), // retain from localStorage or default
|
||||
personIds: new SvelteSet(),
|
||||
tagIds: new SvelteSet(),
|
||||
@ -141,6 +144,7 @@
|
||||
|
||||
let payload: SmartSearchDto | MetadataSearchDto = {
|
||||
query: filter.queryType === 'smart' ? query : undefined,
|
||||
ocr: filter.queryType === 'ocr' ? query : undefined,
|
||||
originalFileName: filter.queryType === 'metadata' ? query : undefined,
|
||||
description: filter.queryType === 'description' ? query : undefined,
|
||||
country: filter.location.country,
|
||||
|
||||
@ -26,6 +26,7 @@ export const featureFlags = writable<FeatureFlags>({
|
||||
configFile: false,
|
||||
trash: true,
|
||||
email: false,
|
||||
ocr: true,
|
||||
});
|
||||
|
||||
export type ServerConfig = ServerConfigDto & { loaded: boolean };
|
||||
|
||||
@ -162,6 +162,7 @@ export const getJobName = derived(t, ($t) => {
|
||||
[JobName.Library]: $t('external_libraries'),
|
||||
[JobName.Notifications]: $t('notifications'),
|
||||
[JobName.BackupDatabase]: $t('admin.backup_database'),
|
||||
[JobName.Ocr]: $t('admin.machine_learning_ocr'),
|
||||
};
|
||||
|
||||
return names[jobName];
|
||||
|
||||
@ -205,6 +205,7 @@
|
||||
originalFileName: $t('file_name'),
|
||||
description: $t('description'),
|
||||
queryAssetId: $t('query_asset_id'),
|
||||
ocr: $t('ocr'),
|
||||
};
|
||||
return keyMap[key] || key;
|
||||
}
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user