Add variant transfer prepper v2 #1013
Conversation
for more information, see https://pre-commit.ci
…to add-variant-transfer-prepper
for more information, see https://pre-commit.ci
… into add-variant-transfer-prepper-v2
for more information, see https://pre-commit.ci
| # NOTE: The return_df option doesn't work, the DataFrame.pivot() function fails. Fix is pending. | ||
| scores_list = scorer.score( | ||
| results, | ||
| use_dask=inputs.use_dask, | ||
| dask_client=dask_client, | ||
| return_df=False, | ||
| dask_failure_mode=inputs.dask_failure_mode, | ||
| ) | ||
| del results | ||
| import csv | ||
| with open(data_intermediates / "docking_scores_raw.csv", 'w', newline='') as myfile: | ||
| wr = csv.writer(myfile, quoting=csv.QUOTE_ALL) | ||
| for item in scores_list: | ||
| wr.writerow([item]) |
There was a problem hiding this comment.
@mariacm12 could you add info on the error you got? is it the target-related stuff?
There was a problem hiding this comment.
It's because the dataframe has duplicate entries I think. Although I couldn't see the duplicates myself when I generated the csv file the alternative way (from DataFrame.pivot() docs: Raises ValueError when there are any index, columns combinations with multiple values)
There was a problem hiding this comment.
Here (line 432) I had to use the hash of the reference complex, but the target name of the actual target. This is a quick fix because the hash and unique_name attributes depend on the ligand's InChIKey, which in this case the target doesn't have. Previously the target name of the reference was used so it was re-writting all of the docking results with the same name (the reference's).
The hash string still needs to be fixed though, I just don't know how to do it without breaking the other workflows. because the current fix still leads to problems when importing files from cache.
| results, | ||
| use_dask=inputs.use_dask, | ||
| dask_client=dask_client, | ||
| return_df=False, |
There was a problem hiding this comment.
If you return_df=True here will be a lot easier to wrangle, including CSV using df.save_csv below.
There was a problem hiding this comment.
@hmacdope see Maria's comments, return_df isn't working. ultimately we'd want to use that though
Description
copied over from #972
We want a way to transfer molecules from a crystal structure to an alphafold (or otherwise generated) structure and then re-dock them with OpenEye / etc.
Based on very old code from ben.
This is also a way around #927
Todos
Notable points that this PR has either accomplished or will accomplish.
make_du_from_new_ligQuestions
Status
Developers certificate of origin