LTR document searcher and docs to reproduce #835
Conversation
stephaniewhoo
commented
Oct 26, 2021
stephaniewhoo
commented
- add document scripts
- add document searcher
- add document repro doc
| @@ -0,0 +1,217 @@ | |||
| # | |||
There was a problem hiding this comment.
There's a lot of code duplicate here with: https://github.com/castorini/pyserini/blob/master/pyserini/ltr/search_msmarco_passage/__main__.py
|
|
||
| # We're going to explicitly use a local installation of Pyserini (as opposed to a pip-installed one). | ||
| # Comment these lines out to use a pip-installed one instead. | ||
| sys.path.insert(0, './') |
There was a problem hiding this comment.
this really suggests that this script should be in scripts/?
There was a problem hiding this comment.
For that matter, shouldn't pyserini/ltr/search_msmarco_passage/__main__.py be in scripts/ also?
There was a problem hiding this comment.
we reorganize the LTR pipeline according to #611, in order to have something like python -m pyserini.ltr.search_msmarco_passage \
If the main goes to scripts/, it will become sth python xxx.py. Is this what you mean?
They have many duplicate codes since the document exp uses passage-like segments in LTR process. Or I change codes in pyserini/ltr/search_msmarco_passage/ to make it work for both passage/segments in doc?
There was a problem hiding this comment.
I think it should be python xxx.py.
The reason is that python -m ... should work with just a pip install... but since you are depending on a checkout, python -m ... wouldn't work anyways, right?
|
close as new PR #844 |
