Parse error for empty @string definitions

[dominique-unruh]

If test.bib contains:

@string{test = ""}

and we call

import bibtexparser
bibtexparser.load(open("test.bib"))

we get an exception (see below). The expected behavior would be to just define test as an empty string. (Bibtex accepts such entries.)

Affected version: PyPI version from May 26, 2020 (1.2.0)

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-78-f9b4a9ef7457> in <module>
      1 import bibtexparser
----> 2 bibtexparser.load(open("test.bib"))

/usr/local/lib/python3.8/dist-packages/bibtexparser/__init__.py in load(bibtex_file, parser)
     67     if parser is None:
     68         parser = bparser.BibTexParser()
---> 69     return parser.parse_file(bibtex_file)
     70 
     71 

/usr/local/lib/python3.8/dist-packages/bibtexparser/bparser.py in parse_file(self, file, partial)
    167         :rtype: BibDatabase
    168         """
--> 169         return self.parse(file.read(), partial=partial)
    170 
    171     def _init_expressions(self):

/usr/local/lib/python3.8/dist-packages/bibtexparser/bparser.py in parse(self, bibtex_str, partial)
    145         bibtex_file_obj = self._bibtex_file_obj(bibtex_str)
    146         try:
--> 147             self._expr.parseFile(bibtex_file_obj)
    148         except self._expr.ParseException as exc:
    149             logger.error("Could not parse properly, starting at %s", exc.line)

/usr/local/lib/python3.8/dist-packages/bibtexparser/bibtexexpression.py in parseFile(self, file_obj)
    276 
    277     def parseFile(self, file_obj):
--> 278         return self.main_expression.parseFile(file_obj, parseAll=True)

/usr/local/lib/python3.8/dist-packages/pyparsing.py in parseFile(self, file_or_filename, parseAll)
   2575                 file_contents = f.read()
   2576         try:
-> 2577             return self.parseString(file_contents, parseAll)
   2578         except ParseBaseException as exc:
   2579             if ParserElement.verbose_stacktrace:

/usr/local/lib/python3.8/dist-packages/pyparsing.py in parseString(self, instring, parseAll)
   1941             instring = instring.expandtabs()
   1942         try:
-> 1943             loc, tokens = self._parse(instring, 0)
   1944             if parseAll:
   1945                 loc = self.preParse(instring, loc)

/usr/local/lib/python3.8/dist-packages/pyparsing.py in _parseNoCache(self, instring, loc, doActions, callPreParse)
   1681             if self.mayIndexError or preloc >= len(instring):
   1682                 try:
-> 1683                     loc, tokens = self.parseImpl(instring, preloc, doActions)
   1684                 except IndexError:
   1685                     raise ParseException(instring, len(instring), self.errmsg, self)

/usr/local/lib/python3.8/dist-packages/pyparsing.py in parseImpl(self, instring, loc, doActions)
   4779     def parseImpl(self, instring, loc, doActions=True):
   4780         try:
-> 4781             return super(ZeroOrMore, self).parseImpl(instring, loc, doActions)
   4782         except (ParseException, IndexError):
   4783             return loc, []

/usr/local/lib/python3.8/dist-packages/pyparsing.py in parseImpl(self, instring, loc, doActions)
   4695         if check_ender:
   4696             try_not_ender(instring, loc)
-> 4697         loc, tokens = self_expr_parse(instring, loc, doActions, callPreParse=False)
   4698         try:
   4699             hasIgnoreExprs = (not not self.ignoreExprs)

/usr/local/lib/python3.8/dist-packages/pyparsing.py in _parseNoCache(self, instring, loc, doActions, callPreParse)
   1681             if self.mayIndexError or preloc >= len(instring):
   1682                 try:
-> 1683                     loc, tokens = self.parseImpl(instring, preloc, doActions)
   1684                 except IndexError:
   1685                     raise ParseException(instring, len(instring), self.errmsg, self)

/usr/local/lib/python3.8/dist-packages/pyparsing.py in parseImpl(self, instring, loc, doActions)
   4252         for e in self.exprs:
   4253             try:
-> 4254                 ret = e._parse(instring, loc, doActions)
   4255                 return ret
   4256             except ParseException as err:

/usr/local/lib/python3.8/dist-packages/pyparsing.py in _parseNoCache(self, instring, loc, doActions, callPreParse)
   1714                 for fn in self.parseAction:
   1715                     try:
-> 1716                         tokens = fn(instring, tokensStart, retTokens)
   1717                     except IndexError as parse_action_exc:
   1718                         exc = ParseException("exception raised in parse action")

/usr/local/lib/python3.8/dist-packages/pyparsing.py in wrapper(*args)
   1314         while 1:
   1315             try:
-> 1316                 ret = func(*args[limit[0]:])
   1317                 foundArity[0] = True
   1318                 return ret

/usr/local/lib/python3.8/dist-packages/bibtexparser/bparser.py in <lambda>(s, l, t)
    199         self._expr.string_def.addParseAction(
    200             lambda s, l, t: self._add_string(t['StringName'].name,
--> 201                                              t['StringValue'])
    202             )
    203 

/usr/local/lib/python3.8/dist-packages/pyparsing.py in __getitem__(self, i)
    596         else:
    597             if i not in self.__accumNames:
--> 598                 return self.__tokdict[i][-1][0]
    599             else:
    600                 return ParseResults([v[0] for v in self.__tokdict[i]])

KeyError: 'StringValue'

[omangin]

Thanks @dominique-unruh for reporting this. I could reproduce the issue and open #276 with a candidate fix. Let us know if it solves the issue on your side.

[dominique-unruh]

Thanks for the quick reaction. Since you were able to reproduce the error, I assume your fix will fix it on my side, too. (I cannot easily test it at this point in the original scenario due to a workaround that I already put in place, but I am looking forward to removing it again when the next version is published on PyPi.)

[dkales]

Just happened to stumble across the same bug when trying to parse cryptobib, since they have some empty strings. After first patching it myself locally, and later finding this issue, I can confirm that #276 does indeed (as expected) fix the issue.

[MiWeiss]

This needs to be taken over by a new person. #276 is stale and broken.

[MiWeiss]

Fixed in v2.

PR for v1 is stale and thus closed. But if anyone ones to take this PR up again and get it to run, I am happy to re-open.