Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Implement highlight and snippet for FTS

Browse Source
This commit is contained in:
Kovid Goyal 2020-06-13 16:35:43 +05:30
parent 5916fd1f5d
commit 6092abe13c
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
3 changed files with 29 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
src/calibre/db/backend.py
View File

@ -1775,12 +1775,20 @@ class DB(object):
yield x
def search_annotations(self,
fts_engine_query, use_stemming, highlight_start, highlight_end, annotation_type,
fts_engine_query, use_stemming, highlight_start, highlight_end, snippet_size, annotation_type,
restrict_to_book_ids, restrict_to_user
):
fts_table = 'annotations_fts_stemmed' if use_stemming else 'annotations_fts'
query = 'SELECT {0}.id, {0}.book, {0}.format, {0}.user_type, {0}.user, {0}.annot_data FROM {0} '
query = query.format('annotations')
text = 'annotations.searchable_text'
if highlight_start is not None and highlight_end is not None:
if snippet_size is not None:
text = 'snippet({fts_table}, 0, "{highlight_start}", "{highlight_end}", "…", {snippet_size})'.format(
fts_table=fts_table, highlight_start=highlight_start, highlight_end=highlight_end,
snippet_size=max(1, min(snippet_size, 64)))
else:
text = 'highlight({}, 0, "{}", "{}")'.format(fts_table, highlight_start, highlight_end)
query = 'SELECT {0}.id, {0}.book, {0}.format, {0}.user_type, {0}.user, {0}.annot_data, {1} FROM {0} '
query = query.format('annotations', text)
query += ' JOIN {fts_table} ON annotations.id = {fts_table}.rowid'.format(fts_table=fts_table)
query += ' WHERE {fts_table} MATCH ?'.format(fts_table=fts_table)
data = [fts_engine_query]
@ -1790,8 +1798,16 @@ class DB(object):
if annotation_type:
query += ' AND annotations.annot_type = ? '
data.append(annotation_type)
for (rowid, book_id, fmt, user_type, user, annot_data) in self.execute(query, tuple(data)):
yield {'id': rowid, 'book_id': book_id, 'format': fmt, 'user_type': user_type, 'user': user, 'annotation': annot_data}
for (rowid, book_id, fmt, user_type, user, annot_data, text) in self.execute(query, tuple(data)):
yield {
'id': rowid,
'book_id': book_id,
'format': fmt,
'user_type': user_type,
'user': user,
'text': text,
'annotation': annot_data
}
def all_annotations_for_book(self, book_id):
for (fmt, user_type, user, data) in self.execute('SELECT format, user_type, user, annot_data FROM annotations WHERE book=?', (book_id,)):

3
src/calibre/db/cache.py
View File

@ -2308,13 +2308,14 @@ class Cache(object):
use_stemming=True,
highlight_start=None,
highlight_end=None,
snippet_size=None,
annotation_type=None,
restrict_to_book_ids=None,
restrict_to_user=None,
):
return tuple(self.backend.search_annotations(
fts_engine_query, use_stemming, highlight_start, highlight_end,
annotation_type, restrict_to_book_ids, restrict_to_user
snippet_size, annotation_type, restrict_to_book_ids, restrict_to_user
))
@write_api

8
src/calibre/db/tests/writing.py
View File

@ -778,7 +778,7 @@ class WritingTest(BaseTest):
annot_list = [
a(type='bookmark', title='bookmark1 changed', seq=1),
a(type='highlight', highlighted_text='text1', uuid='1', seq=2),
a(type='highlight', highlighted_text='text2', uuid='2', seq=3, notes='notes2 some word changed'),
a(type='highlight', highlighted_text='text2', uuid='2', seq=3, notes='notes2 some word changed again'),
]
def map_as_list(amap):
@ -806,7 +806,11 @@ class WritingTest(BaseTest):
results = cache.search_annotations('"Change"')
self.assertEqual([1, 3], [x['id'] for x in results])
results = cache.search_annotations('"change"', use_stemming=False)
self.assertEqual([], [x['id'] for x in results])
self.assertFalse(results)
results = cache.search_annotations('"bookmark1"', highlight_start='[', highlight_end=']')
self.assertEqual(results[0]['text'], '[bookmark1] changed')
results = cache.search_annotations('"word"', highlight_start='[', highlight_end=']', snippet_size=3)
self.assertEqual(results[0]['text'], '…some [word] changed…')
annot_list[0][0]['title'] = 'changed title'
cache.set_annotations_for_book(1, 'moo', annot_list)