Decoding error

[nguyenlab]

I am running penman with output from an AMR parser. It outputs ":null_edge (x20 / 876-9))))))"
Hence, penman cannot decode it because of the relation "null_edge".
I replace the underscore with a minus then it can be solved.

I am not sure whether "null_edge" violates the annotation of penman or the decoder can not recognize this pattern?

Many thanks

[goodmami]

Hi, thanks for the bug report. I'm having trouble reproducing. Can you fill in a few details:

• Penman version (release version or commit hash)
• Whether you used penman.decode(), penman.AMRCodec().decode(), or something else
• The full input

I'm finding that the problem is the hyphen in the concept rather than the underscore in the relation. Sometimes the error messages from the penman module don't accurately point to the source of the error.

Here is what I see, both in the development version and release v0.6.1. First, with AMRCodec:

>>> penman.AMRCodec().decode('(a / blah :null_edge (x20 / 876-9))')
Traceback (most recent call last):
[...]
penman.DecodeError: Expected ":" or "/" at position 31
>>> penman.AMRCodec().decode('(a / blah :null-edge (x20 / 876-9))')
Traceback (most recent call last):
[...]
penman.DecodeError: Expected ":" or "/" at position 31
>>> penman.AMRCodec().decode('(a / blah :null_edge (x20 / 8769))')
<Graph object (top=a) at 140002126592336>

Then with the default codec:

>>> penman.decode('(a / blah :null_edge (x20 / 876-9))')
Traceback (most recent call last):
[...]
penman.DecodeError: Expected ":" or "/" at position 31
>>> penman.decode('(a / blah :null-edge (x20 / 876-9))')
Traceback (most recent call last):
[...]
penman.DecodeError: Expected ":" or "/" at position 31
>>> penman.decode('(a / blah :null_edge (x20 / 8769))')
<Graph object (top=a) at 140002126592208>

I would expect the original graph to parse with the default codec. Investigating this allowed me to see a possible bug in the AMRCodec, though, that is overly restrictive on concept parsing.

[goodmami]

I believe this is fixed in the latest develop version. It fixes parsing of 876-9 as a node type (concept). I didn't do anything about null_edge because it didn't cause any problems for me.
Can you please check?

[riyajatar37003]

DecodeError: Expected: LPAREN

here is text where it thows error: DART ("Data Access for Responsible Testing")
how can solve this issue, is there any way i can bypass this and is there any impact on output if i bypass this

[goodmami]

Hi @riyajatar37003, please open a new issue with a minimal working example, including the exact string you're trying to parse.