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

Do not generate smil wrappers for spine items that have no sentences

Browse Source
This commit is contained in:
Kovid Goyal 2024-10-25 15:27:05 +05:30
parent 2ea2460753
commit 80c74e54cc
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
2 changed files with 25 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/calibre/ebooks/oeb/polish/container.py
View File

@ -903,20 +903,21 @@ class Container(ContainerBase): # {{{
self.parsed_cache.pop(name, None) self.parsed_cache.pop(name, None)
self.dirtied.discard(name) self.dirtied.discard(name)
def set_media_overlay_durations(self, duration_map): def set_media_overlay_durations(self, duration_map=None):
self.dirty(self.opf_name) self.dirty(self.opf_name)
for meta in self.opf_xpath('//opf:meta[@property="media:duration"]'): for meta in self.opf_xpath('//opf:meta[@property="media:duration"]'):
self.remove_from_xml(meta) self.remove_from_xml(meta)
metadata = self.opf_xpath('//opf:metadata')[0] metadata = self.opf_xpath('//opf:metadata')[0]
total_duration = 0 total_duration = 0
for item_id, duration in duration_map.items(): for item_id, duration in (duration_map or {}).items():
meta = metadata.makeelement(OPF('meta'), property="media:duration", refines="#" + item_id) meta = metadata.makeelement(OPF('meta'), property="media:duration", refines="#" + item_id)
meta.text = seconds_to_timestamp(duration) meta.text = seconds_to_timestamp(duration)
self.insert_into_xml(metadata, meta) self.insert_into_xml(metadata, meta)
total_duration += duration total_duration += duration
meta = metadata.makeelement(OPF('meta'), property="media:duration") if duration_map:
meta.text = seconds_to_timestamp(total_duration) meta = metadata.makeelement(OPF('meta'), property="media:duration")
self.insert_into_xml(metadata, meta) meta.text = seconds_to_timestamp(total_duration)
self.insert_into_xml(metadata, meta)
def dirty(self, name): def dirty(self, name):
''' Mark the parsed object corresponding to name as dirty. See also: :meth:`parsed`. ''' ''' Mark the parsed object corresponding to name as dirty. See also: :meth:`parsed`. '''

28
src/calibre/ebooks/oeb/polish/tts.py
View File

@ -116,8 +116,10 @@ def mark_sentences_in_html(root, lang: str = '', voice: str = '') -> list[Senten
text = ''.join(c.text for c in self.texts) text = ''.join(c.text for c in self.texts)
self.pos = 0 self.pos = 0
for start, length in split_into_sentences_for_tts_embed(text, self.lang): for start, length in split_into_sentences_for_tts_embed(text, self.lang):
elem_id = self.wrap_sentence(start, length) stext = text[start:start+length]
ans.append(Sentence(elem_id, text[start:start+length], self.lang, self.voice)) if stext.strip():
elem_id = self.wrap_sentence(start, length)
ans.append(Sentence(elem_id, stext, self.lang, self.voice))
if self.has_tail: if self.has_tail:
p = self.elem.getparent() p = self.elem.getparent()
spans = [] spans = []
@ -125,6 +127,8 @@ def mark_sentences_in_html(root, lang: str = '', voice: str = '') -> list[Senten
for start, length in split_into_sentences_for_tts_embed(self.elem.tail, self.parent_lang): for start, length in split_into_sentences_for_tts_embed(self.elem.tail, self.parent_lang):
end = start + length end = start + length
text = self.elem.tail[start:end] text = self.elem.tail[start:end]
if not text.strip():
continue
if before is None: if before is None:
before = self.elem.tail[:start] before = self.elem.tail[:start]
span = self.make_wrapper(text, p) span = self.make_wrapper(text, p)
@ -422,7 +426,7 @@ def remove_embedded_tts(container):
container.set_media_overlay_durations({}) container.set_media_overlay_durations({})
media_files = set() media_files = set()
for item in manifest_items: for item in manifest_items:
smil_id = item.get('media-overlay') smil_id = item.attrib.pop('media-overlay', '')
href = item.get('href') href = item.get('href')
if href and smil_id: if href and smil_id:
name = container.href_to_name(href, container.opf_name) name = container.href_to_name(href, container.opf_name)
@ -469,17 +473,23 @@ def embed_tts(container, report_progress=None, callback_to_download_voices=None)
return False return False
all_voices = set() all_voices = set()
total_num_sentences = 0 total_num_sentences = 0
files_with_no_sentences = set()
for i, (name, pfd) in enumerate(name_map.items()): for i, (name, pfd) in enumerate(name_map.items()):
pfd.root = container.parsed(name) pfd.root = container.parsed(name)
pfd.sentences = mark_sentences_in_html(pfd.root, lang=language) pfd.sentences = mark_sentences_in_html(pfd.root, lang=language)
total_num_sentences += len(pfd.sentences) if not pfd.sentences:
for s in pfd.sentences: files_with_no_sentences.add(name)
key = s.lang, s.voice else:
pfd.key_map[key].append(s) total_num_sentences += len(pfd.sentences)
all_voices.add(key) for s in pfd.sentences:
container.dirty(name) key = s.lang, s.voice
pfd.key_map[key].append(s)
all_voices.add(key)
container.dirty(name)
if report_progress(stage, name, i+1, len(name_map)): if report_progress(stage, name, i+1, len(name_map)):
return False return False
for rname in files_with_no_sentences:
name_map.pop(rname)
if callback_to_download_voices is None: if callback_to_download_voices is None:
piper.ensure_voices_downloaded(iter(all_voices)) piper.ensure_voices_downloaded(iter(all_voices))
else: else: