Closed xiki-tempula closed 1 year ago
Merging #317 (7347a9c) into master (4e590cc) will increase coverage by
0.01%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #317 +/- ##
==========================================
+ Coverage 98.74% 98.75% +0.01%
==========================================
Files 26 27 +1
Lines 1754 1772 +18
Branches 382 387 +5
==========================================
+ Hits 1732 1750 +18
Misses 2 2
Partials 20 20
Impacted Files | Coverage Δ | |
---|---|---|
src/alchemlyb/parsing/parquet.py | 100.00% <100.00%> (ø) |
|
src/alchemlyb/workflows/abfe.py | 99.67% <100.00%> (+<0.01%) |
:arrow_up: |
@orbeckst Do you mind have a review of this PR, please? Thank you.
@orbeckst Thanks for the review. I have change the environment.yml. I will raise an issue with meta.yml when we are doing a release.
So the T problem is a bit tricky in that parquet serialisation doesn't preserve the df.attrs. So the dataframe loaded here doesn't contain temperature and the temperature is assigned via extract function. I will add a not to state that.
So the T problem is a bit tricky in that parquet serialisation doesn't preserve the df.attrs. So the dataframe loaded here doesn't contain temperature and the temperature is assigned via extract function. I will add a not to state that.
Ok, I blithely assumed that parquet would save everything — then T
is not optional. Definitely document this requirement.
@orbeckst
Ok, I blithely assumed that parquet would save everything
I'm annoyed by this as well but it seems that parquet is still the best serialiser. This is the only format that preserves index besides to_pickle. Also it faithfully preserve all the data in its original datatype. The to_pickle, though it preserves everything, is slow to read and write and also would not be safe between different versions of pandas.
Definitely document this requirement.
I have added this as a note to both alchemlyb.parsing.parquet.extract_u_nk and alchemlyb.parsing.parquet.extract_dHdl.
Let me know when I need to review again.
@orbeckst I addressed the comments. Do you mind having another review? Thank you.
Please squash-merge when ready.
This PR add a parquet parser, which will load the serialised dataframe and permits the usage of
Fix https://github.com/alchemistry/alchemlyb/issues/316