Recent version of pyparsing (3.0.2) doesn't work well with pydot

Ref: https://github.com/pyparsing/pyparsing/issues/319

Minimal example:

import pydot
pydot.graph_from_dot_data("graph G { b1 [label=<Hello World>]; }")

graph G { b1 [label=<Hello World>]; }
      ^                                                                        
Expected '{', found 'G'  (at char 6), (line:1, col:7)

[MarcCote]

This issue is now resolved in pyparsing 3.0.3. That should probably be reflected in the requirements of pydot.

[ptmcg]

I’ve looked at the networkx issue, and the issue is due to the tightening up of whitespace-skipping that was pushed out in 3.0.2. This is intentional behavior. To correct in pydot, add the two lines shown below to dot_parser.py (I can submit as a PR if you prefer):

noncomma = "".join([c for c in printables if c != ","])
alphastring_ = OneOrMore(CharsNotIn(noncomma + " "))
# ========== add the following two lines ==============
# override pyparsing tightened whitespace-skipping logic
alphastring_.skipWhitespace = True

CharsNotIn has always suppressed whitespace, but in pyparsing 3.0, it is intended to preserve this whitespace-skipping suppression when enclosed in a surrounding expression. My understanding of alphastring_ is that it should collect non-comma characters after skipping over any initial whitespace. Adding this line passes the test suite. (The added comment is optional, but I thought it would be useful to document this atypical attribute setting.)

This fix also will address issues with pydot in networkx (pyparsing/pyparsing#338).