Bug: Sync subclass: ancestor inconsistencies

[joeflack4]

Overview

Nico was trying to fix some bugs in the pipeline (#403) when he noticed an issue regarding ancestor inconsistencies. Basically, we expect that ancestor() edges should give the same results as the indirect edges + direct edges. But we're getting an error where that is not the case.

Sub-tasks

• 1. Find root cause of issue & fix
  • May include Bug: Sync subclass: ancestor inconsistencies #408
• 2. Consider changing back: logging.error() -> raise ValueError()
• 3. Change logger to print info?

3. Change logger to print info?

Details

This is unrelated to this issue, but basically I was expecting logging.info() to print, but it is not. It would help me to debug if this would print. Should we print these? Should we change these to logging.debug() and maybe I change my code so that if I'm running what I typically run in the debugger, run_defaults(), it sets the logging to debug mode?

Shamefully I don't have enough experience with logging. I tried to fix with some suggestions here, but most of what I tried didn't work (nor did ChatGPT work). One suggestion did work, but it ended up printing everything twice.

[joeflack4]

@matentzn Frustratingly, I can't replicate your issue. I probably screwed up somewhere along the way. Maybe you can figure it out? Here's what I tried.

Details

What I did

Setup 1: My current mondo-ingest clone

1. Branch off of your reinstate-sync PR
2. Make new branch / PR bug-sync-ancestor
3. Add %.db bugfix I suggested. Otherwise it was erroring out for me as expected.
4. rm src/ontology/reports/*subclass*
5. Deleted and remade subclass prereqs tmp/mondo-ingest.db tmp/mondo.db tmp/mondo.sssom.tsv`

• Pretty sure I did this without also adding IMP=false COMP=false MIR=false. I know at least they changed somehow because I backed them up first and I just checked md5 and they are different.

7. Tried running in both (i) debugger, and (ii) through ODK with IMP=false COMP=false MIR=false

Setup 2: Fresh mondo-ingest clone

1. Fresh git clone
2. Check out my bug-sync-ancestor branch (with bugfix in step (3) above) from 'setup 1'.
3. rm src/ontology/reports/*subclass*
4. Ran just through ODK sh run.sh make sync

Questions

1. Can you tell me anything that would help me replicate your issue?
2. What source was running when that error happened, if you remember?

[matentzn]

I will rerun it again now and see if it was a de-synchronised issue on my end.. I noticed another issue which I caused (missing pip install). So if all goes well, we can just merge this and we are good again!

[joeflack4]

@matentzn Assigned this one to you now! If you re run and there's no issue feel free to close.

[matentzn]

Yes this was fixed a few PRs ago! Thanks for ping, closing.