omicsnlp / auto-corpus Goto Github PK

I am having issues installing the dependencies in Auto-CORPUS.

When I run pip install . I get a conflicting dependencies error.

I am using Python 3.10.10 and Pip 24.0

@Tom-Shorter can we place the log file in the target_dir rather than base AC directory?
I think it's easier for users to see it there when inspecting their files?

Thanks for the useful library! I wonder if the authors could please clarify under which license the code is released?

The config_pmc.json file will fail for new HTML files from PMC due to the fact that the id tag is gone from the paragraph tag.

I suggest to change it to (where the first part includes Valentina's fix for some articles, and the second part is new to work on new PMC files):

"paragraphs": {
  "data": {},
  "defined-by": [
    {
      "tag": "p",
      "attrs": {"id": "_*[pP\\-|pP|Par]*\\d+"}
    },
	{
      "tag": "p",
      "attrs": {"class": "p"}
    }
  ]
},

I noticed a change compared to a previous version of the algorithm:

@Tom-Shorter, can we change the output to a list and if the abbreviation section and full text algorithm are the same, then skip, but otherwise list both abbreviations. Test with PMC4068805 (where RP is defined in abbreviation section as reverse phase, but in the text as reversed phase (this one is correct actually)).