ParserStateError

[muthupl]

Hi,

We are getting a parser state error.
Here is our test:

nlp = spacy.en.English()
nlp.matcher.add('splash', 'my entity', {},[ [{LEMMA: 'splash'}, {LEMMA: 'on'}]])
nlp('splash On')

Here is the trace:

Traceback (most recent call last):
File "lib/python3.5/site-packages/spacy/language.py", line 328, in call
proc(doc)
File "spacy/syntax/parser.pyx", line 146, in spacy.syntax.parser.Parser.call (spacy/syntax/parser.cpp:6114)
spacy.syntax.parser.ParserStateError: Error analysing doc -- no valid actions available. This should never happen, so please report the error on the issue tracker. Here's the thread to do so --- reopen it if it's closed:
#429
Please include the text that the parser failed on, which is:
'splash On'

Our spacy version: spacy==1.3.0
Python 3.5.1
Mac OSX EI Capitan

Any help or fix is much appreciated! Let us know if you need more details.

[honnibal]

Thanks! Will investigate.

[ejschoen]

Hi. Thanks for all your work. For what it's worth, I'm running into the same issue in version 1.5.0. I added a comment to issue #429 describing the code (match rule, with callback, doing a span merge) and the text I'm parsing, but I'm thinking that's not really the issue where it belongs.

Additionally, I'm able to do this to work around the ParserStateError:

d = nlp.tokenizer("Located 16 km southwest of the Ceres find, the Hebe-1 well tested at 5,956 BOPD.")
nlp.tagger(d)
nlp.parser(d)
nlp.entity(d)
nlp.matcher(d)

The resulting doc has "5,956 BOPD." merged into a single token, but the entity type is ORG, and not QUANTITY. However, the match rule for {like_num,"BOPD."} does try to define the match as a QUANTITY, as does the span.merge in the on_match callback.

One more update...

At least in 1.5.0, passing keyword args to span.merge didn't work for me.

        newspan = span.merge(ent_type='QUANTITY',tag='NNP')

However, this (I think deprecated form) did:

        newspan = span.merge('NNP',span.text,'QUANTITY')

I'm not a Python programmer, so this might be a noob issue on my part.

[honnibal]

I think this is fixed now. The code in this doc.merge() method is still quite poor, and needs to be rewritten. But there's now a passing test for this specific issue. Thanks again for reporting!

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.