[BUG] Error in setting up kegg database in new anvio (Marie)

[Harpreet525]

Short description of the problem

Error in setting up the kegg database in freshly installed latest version of anvio probably because of the naming error

anvi'o version

Anvio 8

anvi-self-test --version

anvi-self-test --suite mini worked fine.

System info

Linux Ubuntu , installation of anvio via pip and by default as indicated in anvio guide.

Detailed description of the issue

File "/bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/biochemistry/reactionnetwork.py", line 1148, in set_up os.mkdir(parent_dir) FileNotFoundError: [Errno 2] No such file or directory: '/bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/MISC/KEGG'

Here, the reactionnetwork.py is locating the data stored in the path /bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/MISC/KEGG.

However, it is unable to find the path because of the case sensitive letter "MISC" in the path. By default, the files are stored in data/misc/KEGG. But the reactionnetwork.py is locating it at data/MISC/KEGG.

Regards

[meren]

Thanks for the report, @Harpreet525.

This is interesting. Self note: If we can reproduce it, probably os.makedirs() instead of os.mkdir() will solve the problem. I'm not sure why the MISC/KEGG wouldn't be there, though.

[ivagljiva]

I probably missed this because I am always testing with the misc/KEGG directory already in place rather than from a fresh install without that directory :)

I fixed both the incorrect case for misc in the path KODatabase.default_dir, AND I switched the mkdir function to makedirs. Currently testing it with anvi-setup-kegg-data --kegg-data-dir TEST_KEGG -T 6 --mode modeling, but it will take about two hours to finish, so we won't know for a while if it worked. I will report back.

[meren]

Thank you very much for your quick action, @ivagljiva!

[Harpreet525]

i had replaced and changed the case of data/MISC/KEGG path with data/misc/KEGG in the reactionnetwork.py script and the kegg database setup completed successfully. Hoping this was the correct method !

Regards

[Harpreet525]

Total number of KOs/entry files ..............: 26,252 KEGG KO database version .....................: Release 108.0+/10-03, Oct 23 KEGG KO list .................................: /bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/misc/KEGG/KO_REACTION_NETWORK/ko_list.txt KEGG KO info .................................: /bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/misc/KEGG/KO_REACTION_NETWORK/ko_info.txt Table of select KEGG KO data .................: /bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/misc/KEGG/KO_REACTION_NETWORK/ko_data.tsv Archived KEGG KO entry files .................: /bettik/PROJECTS/pr-microbio/Singhhar/miniconda3/envs/anvio/lib/python3.10/site-packages/anvio/data/misc/KEGG/KO_REACTION_NETWORK/ko_entries.tar.gz

✓ anvi-setup-kegg-data took 1:45:28.341934

[ivagljiva]

That was exactly the right way to fix it @Harpreet525 :) Looks like it worked.

Now this issue should be fixed in anvio-dev as well!

[meren]

Should I update the tar.gz archive attached to the release, @ivagljiva? We didn't do anything major in the repo, and it would be nice for people to not have to switch to anvio-dev to address this issue. What do you think?

[ivagljiva]

I think that is a good idea @meren.

But please wait a little bit before you do it. I want my test to finish running just to make sure everything is A-OK first :)

(I'll ping you here when its done)

[semiller10]

Thank you very much for catching that, @Harpreet525, and for the quick fix, @ivagljiva. Updating the release sounds fine, @meren.

[ivagljiva]

@meren , this should be good to go :) Feel free to update the release (and I will close this issue). Thank you all very much!

[meren]

this should be good to go :) Feel free to update the release (and I will close this issue)

Just for posterity, this is now done, and I updated the tarball attached to the release reflecting the tested changes implemented in 59640ec0e513088754e9e53734481a4ac0459996 and 7ca5a8d30c6f1434f2af5e8b61490e10c66834c8.