Write data and package versions to data/output

[glass-ships]

Closes #394

• [x] Split data and packages into separate sections
• [ ] Find a way to get versions of other data sources
• [ ] May need beautifulsoup to scrape some web pages
• [ ] Maybe use kghubdownloader / python-jq?? to supply API endpoint and get versions?

[glass-ships]

Currently, output/versions.yaml looks like:

kg-version: 2024-01-23

packages:
  biolink: 3.6.0
  koza: 0.5.3
  monarch-ingest: 0.8.0

data:
  phenio: v2024-01-12
  alliance: 6.0.0

@kevinschaper should we a) rename this to metadata.yaml b) bring this in as is c) wait to bring it in and keep looking for ways to get versions of additional data sources

[glass-ships]

@matentzn Just wanted to bring this to your attention as a WIP solution to the issue of including data/tooling versions in some metadata file that gets uploaded alongside the knowledge graph / related files

[matentzn]

@glass-ships thank you very much for tagging me! This is fantastic. I have talked a bit to @kevinschaper about this, and I will propose a much more comprehensive dump pf information - but not very soon (after the 14th of February). The concept will be the same, but the written file will look different (JSON instead of yaml, and more metadata fields). For now, the only thing I would suggest is renaming the Koza parameter to --write-metadata, and go ahead with your PR, with the understanding that the file will look a bit different in the end.

[caufieldjh]

@matentzn I'm curious about the choice of JSON vs YAML here (but generally in favor of anything allowing better version tracking)

[matentzn]

@matentzn I'm curious about the choice of JSON vs YAML here (but generally in favor of anything allowing better version tracking)

@caufieldjh lets discuss this in a bespoke issue - I will make sure to loop you in also because of the b2ai work during the brainstorming process.