As I understand, pdftable.py converts a pdf table to a .csv file by using html as an intermediary.
Since pdftable uses htmllib, which was deprecated in 2.6 in favor of the HTMLParser module, your problem is not with the transition from 2.x HTMLParser.HTMLParser to 3.x html.parser.HTMLParser, but with the transition from 2.x htmllib.HTMLParser to 2.x HTMLParser.HTMLParser. Even though the class name remained HTMLParser, the API is quite different for everything other than the .feed(text) method. In order to rewrite htmllib code, one must understand what it is doing as mechanical replacement is not possible.
For htmllib, the signature is HTMLParser(formatter), where formatter is expected to be one of the classes in the formatter module, or a subclass thereof. (The formatter module was deprecated in 3.4 since the removal of htmllib left is pretty much unused.) The intention is that one instantiate a subclass of HTMLParser with added tag methods. However, pdftable uses an empty parser.
self.html_parser = htmllib.HTMLParser(None)
In the first line of filter,
str = re.sub(r'<[^>]+>', '', str)
the regex appears to match tags and the substitution removes them. For the next three lines
self.set.html_parser.save_bgn()
self.set.html_parser.feed(str)
return self.set.html_parser.save_end()
save_bgn() says to begin "saving character data in a buffer instead of sending it to the formatter object." (Good thing, since there is no formatter.) With no tags and no tag methods, I don't know what feeding the string through the parser does. I would not be surprised if the answer is "Nothing". If so, your answer would be to remove the three lines, and possibly def filter, replacing the filter() call by the re.sub call.
To find out, I suggest adding some 2.x print statements to filter, and then run on 2.7 with your example pdf file.
def filter(self, str):
print 'Before replace:', str
str = re.sub(r'<[^>]+>', '', str)
print 'After replace:', str
self.set.html_parser.save_bgn()
print 'After parse:', str
self.set.html_parser.feed(str)
return self.set.html_parser.save_end()
