Open hexylena opened 2 years ago
We've mostly been looking at scraping Event
s/CourseInstance
s to begin with, so the material scraping is pretty basic.
TeSS' Material model is also not fully aligned with the Bioschemas TrainingMaterial profile.
I think marking up with as much information as possible is never a bad thing, as long as the information is accurate and semantically/structurally correct. TeSS' scrapers are run regularly, so as the markup (or the scraping code and TeSS' data model) is changed, we can update existing entries in TeSS with new metadata.
Thanks @fbacall, I think I had the wrong impression of priorities, it makes sense now.
TeSS' scrapers are run regularly, so as the markup (or the scraping code and TeSS' data model) is changed, we can update existing entries in TeSS with new metadata.
can't wait to see our updated entries in TeSS! We (Galaxy) have been following the replacement scrapers progress for the last 2 years so we're really excited for this.
educationalLevel
per bioschemas TrainingMaterial 1.0 so that sounds like it'll work.I think ours is marked up pretty well for a11y items:
"accessMode": [
"textual",
"visual"
],
"accessModeSufficient": [
"textual",
"visual"
],
"accessibilityControl": [
"fullKeyboardControl",
"fullMouseControl"
],
"accessibilityFeature": [
"alternativeText",
"structuralNavigation",
"tableOfContents"
],
"accessibilitySummary": "The text aims to be as accessible as possible. Image descriptions will vary per tutorial, from images being completely inaccessible, to images with good descriptions for non-visual users.",
it could use a few more accessibilityFeature items, but in general that's correct. Probably accessMode
is one of the more important ones to surface, so learners know if the material is accessible via text, or video/audio content only. Same for accessibilityFeature
that the material is structured semantically, and uses alt text properly.
I'm looking at https://github.com/ElixirTeSS/tess-bioschemas-scraper/blob/master/src/TessApi/Material.ts#L29-L36 and noticing that we use a lot of extra fields. I'm wondering if any of these are in-scope for scraping and pushing to TeSS?
We write pretty extensive json-ld and we've been really curious to follow the scraper progress just to make sure we are compliant, but it looks like currently the scraper only extracts a limited vocabulary so maybe we've gone overboard? :) Example of our markup is available in: https://training.galaxyproject.org/training-material/topics/assembly/tutorials/debruijn-graph-assembly/tutorial.html