tris/recoll
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Fixes issue with pdftohtml

Browse Source
This commit is contained in:
Tris 2021-07-02 18:31:12 +10:00
parent 2a92200012
commit 551bc3df5e
1 changed files with 36 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
src/filters/rclpdf.py
View File

@ -469,7 +469,40 @@ class PDFExtractor:
if annotsfield: if annotsfield:
self.em.setfield("pdfannot", annotsfield) self.em.setfield("pdfannot", annotsfield)
return html return html
def _patch_meta(self, html):
'''This fixes https://gitlab.freedesktop.org/poppler/poppler/-/issues/136'''
if not _mswindows:
pdfinfo = rclexecm.which("pdfinfo")
if not pdfinfo:
pdfinfo = rclexecm.which("poppler/pdfinfo")
if not pdfinfo:
return html
info = subprocess.check_output([pdfinfo, '-isodates', self.filename])
meta = {}
for line in info.split(b'\n'):
try:
key, value = line.strip().split(b':', 1)
meta[key.strip()] = value.strip().replace(b'"', b'\"')
except ValueError:
pass
if b'CreationDate' in meta:
meta[b'date'] = meta[b'CreationDate']
title = meta.get('Title')
head = [b'<title>%s</title>' % title] if title else []
head += [ b'<meta name="%s" content="%s"/>' % x for x in meta.items() ]
start = html.index(b'<head>') + 6
end = html.index(b'</head>')
return html[:start] + b"".join(head) + html[end:]
def _selfdoc(self): def _selfdoc(self):
'''Extract the text from the pdf doc (as opposed to attachment)''' '''Extract the text from the pdf doc (as opposed to attachment)'''
self.em.setmimetype('text/html') self.em.setmimetype('text/html')
@ -483,6 +516,8 @@ class PDFExtractor:
"UTF-8", "-eol", "unix", "-q", "UTF-8", "-eol", "unix", "-q",
self.filename, "-"]) self.filename, "-"])
html = self._patch_meta(html)
html, isempty = self._fixhtml(html) html, isempty = self._fixhtml(html)
#self.em.rclog("ISEMPTY: %d : data: \n%s" % (isempty, html)) #self.em.rclog("ISEMPTY: %d : data: \n%s" % (isempty, html))