This commit is contained in:
2026-09-14 11:47:29 +03:00
parent 971f60652d
commit 0dad80bb17
3 changed files with 301 additions and 17 deletions

View File

@@ -985,8 +985,69 @@ def parse_html(path: Path) -> list[Section]:
return sections
def _file_url_to_path(url: str) -> Optional[Path]:
"""file:///… или обикновен път → Path (вкл. UNC от Word r:link)."""
from urllib.parse import unquote
s = unquote((url or "").strip())
if not s:
return None
if s.lower().startswith("file:"):
s = s[5:]
while s.startswith("/"):
s = s[1:]
if not s:
return None
return Path(s)
def _load_external_image_bytes(url: str) -> Optional[tuple[bytes, str]]:
"""Чете външна картинка (r:link / TargetMode=External); връща (data, ext)."""
path = _file_url_to_path(url)
if path is None:
return None
try:
if not path.is_file():
return None
data = path.read_bytes()
except OSError:
return None
if not data:
return None
ext = (path.suffix or "").lstrip(".").lower() or "png"
if ext == "jpeg":
ext = "jpg"
return data, ext
def _resolve_docx_image_rid(doc, rId: str) -> Optional[tuple[bytes, str]]:
"""r:embed → related_parts; r:link → външен файл. Връща (data, ext)."""
if not rId:
return None
try:
part = doc.part.related_parts[rId]
data = part.blob
ct = getattr(part, "content_type", "") or ""
if data:
return data, _ext_from_content_type(ct)
except Exception:
pass
try:
rel = doc.part.rels[rId]
except Exception:
return None
if not getattr(rel, "is_external", False):
return None
target = getattr(rel, "target_ref", None) or ""
loaded = _load_external_image_bytes(target)
return loaded
def _extract_docx_paragraph_images(para, doc) -> list[ImageRef]:
"""Намира drawing-и в параграф; връща ImageRef-и за филтрираните по размер."""
"""Намира drawing-и в параграф; връща ImageRef-и за филтрираните по размер.
Поддържа r:embed (вградени) и r:link (външни file:/// / UNC пътища).
"""
from docx.oxml.ns import qn
imgs: list[ImageRef] = []
try:
@@ -995,19 +1056,17 @@ def _extract_docx_paragraph_images(para, doc) -> list[ImageRef]:
return imgs
embed_attr = qn("r:embed")
link_attr = qn("r:link")
for blip in blips:
rId = blip.get(embed_attr)
rId = blip.get(embed_attr) or blip.get(link_attr)
if not rId:
continue
try:
part = doc.part.related_parts[rId]
data = part.blob
ct = getattr(part, "content_type", "") or ""
except Exception:
resolved = _resolve_docx_image_rid(doc, rId)
if not resolved:
continue
data, ext = resolved
if not _should_keep_image(data):
continue
ext = _ext_from_content_type(ct)
imgs.append(ImageRef(placeholder=f"__IMG_{len(imgs)+1}__", data=data, ext=ext))
return imgs
@@ -1023,17 +1082,72 @@ def _docx_run_color_css(run) -> Optional[str]:
return None
def _docx_run_to_html(run) -> str:
def _docx_style_font_flag(style, attr: str) -> Optional[bool]:
"""True/False от style.font.<attr>; None ако стилът/атрибутът липсва."""
if style is None:
return None
try:
font = style.font
if font is None:
return None
val = getattr(font, attr, None)
if val is None:
return None
return bool(val)
except Exception:
return None
def _docx_run_effective_flag(run, para, attr: str) -> bool:
"""Ефективен bold/italic/underline: изричен run → char style → paragraph style."""
try:
direct = getattr(run, attr, None)
except Exception:
direct = None
if direct is True:
return True
if direct is False:
return False
# None = наследяване
try:
char_style = run.style
except Exception:
char_style = None
flag = _docx_style_font_flag(char_style, attr)
if flag is not None:
return flag
try:
p_style = para.style if para is not None else None
except Exception:
p_style = None
flag = _docx_style_font_flag(p_style, attr)
return bool(flag)
def _iter_docx_para_runs(para):
"""Runs в параграф, вкл. вътре в w:hyperlink (python-docx.para.runs ги пропуска)."""
from docx.oxml.ns import qn
from docx.text.run import Run
for child in para._element.iterchildren():
if child.tag == qn("w:r"):
yield Run(child, para)
elif child.tag == qn("w:hyperlink"):
for r_elem in child.findall(qn("w:r")):
yield Run(r_elem, para)
def _docx_run_to_html(run, para=None) -> str:
"""Един Word run → HTML с <b>/<i>/<u> и color span. Спец. символи се escape-ват."""
text = run.text or ""
if not text:
return ""
s = html_lib.escape(text, quote=False).replace("\n", "<br>")
if run.bold:
if _docx_run_effective_flag(run, para, "bold"):
s = f"<b>{s}</b>"
if run.italic:
if _docx_run_effective_flag(run, para, "italic"):
s = f"<i>{s}</i>"
if run.underline:
if _docx_run_effective_flag(run, para, "underline"):
s = f"<u>{s}</u>"
css_color = _docx_run_color_css(run)
if css_color:
@@ -1042,11 +1156,15 @@ def _docx_run_to_html(run) -> str:
def _docx_para_to_html(para) -> str:
"""Параграф → <p>…</p> с inline форматиране от runs."""
parts = [_docx_run_to_html(r) for r in para.runs]
"""Параграф → <p>…</p> с inline форматиране от runs (вкл. hyperlink TOC)."""
parts = [_docx_run_to_html(r, para) for r in _iter_docx_para_runs(para)]
inner = "".join(parts)
if not inner.strip():
return ""
# Fallback: para.text вижда hyperlink текст, но без runs в .runs
plain = (para.text or "").strip()
if not plain:
return ""
inner = html_lib.escape(plain, quote=False).replace("\n", "<br>")
return f"<p>{inner}</p>"