Failing on valid WDL files

[anton-khodak]

I use pywdl to parse descriptions from the main repository, for instance, this one and this . Both descriptions have valid syntax (wdltool showed no errors on checking). However, when I run pywdl on either of them, I get the following errors:

$ wdl parse "/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/test/cases/ValidateBamsWf_170107.wdl"
Traceback (most recent call last):
 File "/home/anton/programming/venvs/pywdl/bin/wdl", line 4, in <module>
   __import__('pkg_resources').run_script('wdl==1.0.22', 'wdl')
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/setuptools-18.1-py3.5.egg/pkg_resources/__init__.py", line 735, in run_script
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/setuptools-18.1-py3.5.egg/pkg_resources/__init__.py", line 1659, in run_script
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/wdl-1.0.22-py3.5.egg/EGG-INFO/scripts/wdl", line 6, in <module>
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/wdl-1.0.22-py3.5.egg/wdl/main.py", line 76, in cli
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/wdl-1.0.22-py3.5.egg/wdl/parser.py", line 1222, in parse
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/wdl-1.0.22-py3.5.egg/wdl/parser.py", line 4077, in lex
 File "/home/anton/programming/venvs/pywdl/lib/python3.5/site-packages/wdl-1.0.22-py3.5.egg/wdl/parser.py", line 4073, in lex
AttributeError: 'str' object has no attribute 'unrecognized_token'

if I do this programmatically:

>>> with open('/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/test/cases/ValidateBamsWf_170107.wdl') as f:
    wdl.load(f)    
    
Traceback (most recent call last):
  File "<input>", line 2, in <module>
  File "/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/__init__.py", line 6, in load
    return parse_namespace(fp.read(), resource if resource is not None else fp.name)
  File "/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/binding.py", line 284, in parse_namespace
    tokens = wdl.parser.lex(string, resource, errors)
  File "/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/parser.py", line 4077, in lex
    return TokenStream(HermesLexer().lex(source, resource, errors, debug))
  File "/media/anton/ECFA959BFA95631E/Programming/pywdl/wdl/parser.py", line 4073, in lex
    raise ctx.errors.unrecognized_token(string_copy, ctx.line, ctx.col)
wdl.parser.SyntaxError: Unrecognized token on line 77, column 10:
    Array[File] validation_reports = ValidateBAM.validation_report
         ^

[anton-khodak]

UPD. I regenerated parser.py with hermes and tried again, with no success

[anton-khodak]

Please, tell me, is anything expected to be tackled regarding this issue in the nearest time? It is blocking me. If not, I will need to write just the same parser, but without bugs.

[vdauwera]

This was a proof of concept implementation, but we no longer use it and the language specification has evolved, which is probably why it seems broken. You're welcome to fork it and/or write your own as we do not plan to do anything more with it as far as I'm aware.

[geoffjentry]

@anton-khodak - @vdauwera is correct, our dev team is no longer actively maintaining this. We should make this more clear. That might change in the future, but for now if someone was interested in picking it up we think that'd be awesome.