Edit this page | Blame

Genewiki conversion

GeneNetwork1 contains a genewiki:

We want to migrate it to markdown documents that can be fetched from

So the steps are to (1) migrate the existing genewiki data in the database to named markdown documents in that repository and (2) create a rendered page that is found through

with an edit button, similar to

Tags

  • assigned: pjotrp, zsloan
  • type: enhancement
  • priority: medium
  • status: unclear
  • keywords: GN1, documentation

Tasks

  • [ ] Export Genewiki to markdown - one file per gene and store in git@github.com:genenetwork/gn-docs.git
  • [ ] Format output for GN using markdown parser (similar to other docs)
  • [ ] Provide edit link to github

Later we'll add automated links to wikidata and Uniprot etc.

Notes

Zach writes: How exactly do we want to store all of this? It appears to currently be stored across three SQL tables - GeneRIF, GeneRIFXRef, and GeneCategory. The first contains a row for each item a user adds (when displaying all items it queries by gene symbol), and the latter two are for storing the checkbox stuff (so there will presumably be a row in GeneRIFXRef for every checked box for each symbol, though this isn't totally clear to me because it's linked by GeneRIF.Id - which isn't unique - rather than GeneRIF.symbol which is what I would have assumed).

IIRC the issue I ran into (that isn't immediately apparent from looking at the web page) is that it's currently stored as a list of items. There isn't a single "free text" area - when a user edits they are either adding a new text item with its own row in the DB or editing one of the existing items, so I'm not sure how best to reasonably convert the current contents and editing method to markdown. Currently it doesn't even support any sort of user styling/formatting - users just enter basic text into a form. And if they were converted to markdown, how would we be storing the checkbox content?

It's probably possible to write a script that goes through those tables and generates a bunch of markdown files from them (one for each gene symbol, I think?), with the list of items just being converted into a single markdown file with those items formatted into a list. This would de-link GN1's GeneWiki from GN2's in the future, though (since the way things are stored would be fundamentally changed).

Pj: That is what we want. Create a markdown file for each gene symbol. Checklist can be part of that using markdown syntax.


(made with skribilo)