srlearn / rnlp

Relational NLP: Convert text into relational facts.
https://rnlp.readthedocs.io/en/latest/
GNU General Public License v3.0
9 stars 5 forks source link

Output files #9

Closed hayesall closed 4 years ago

hayesall commented 6 years ago

Present implementation appends facts, blocks, sentences, etc. to files in the user's current directory.

Ideally there should be an output flag (-o) with which users may specify where these should be written to.

adamjsawicki commented 4 years ago

If -o is specified, should all writing be done to that file regardless of whether it's a fact, block, sentence, etc?

hayesall commented 4 years ago

Currently this outputs four files (see parse.py)

I think a -o flag would be helpful for pre-pending a path, for example: -o financial_documents/ would write financial_documents/facts.txt, financial_documents/blockIDs.txt, etc.

@Swixx do you think separate output flags for each file would be more useful?

adamjsawicki commented 4 years ago

No, I think that overcomplicates things without a good use case, but I'm willing to be convinced.

hayesall commented 4 years ago

100% agree. I think your solution in #21 is on the right track!

I can imagine some cases where we include/exclude specific output files, but this seems like a premature optimization.