New zenodo api bandaid #2942
New zenodo api bandaid #2942
Conversation
There was a problem hiding this comment.
Is this PR really supposed to be going into main rather than dev?
Another simplifying change I think we can make with the new API is getting rid of the API keys / tokens entirely, since it seems like read-only operations are fine without them (at least, curl works without them which I don't think was the case before!)
| for f in dpkg.json()["entries"]: | ||
| if f["key"] == "datapackage.json": | ||
| resp = self._fetch_from_url(f["links"]["content"]) |
There was a problem hiding this comment.
If our whole setup is contingent on being able to construct a path to a file given the filename and record ID, is there a reason to use the API to obtain the datapackage.json?
I guess if we want to verify the checksum on datapackage.json we need to get the checksum from Zenodo, since the file can't contain its own checksum.
There was a problem hiding this comment.
My reason for doing this is that "keeping the datapackage.json around lets me not think too hard about how to construct a DatapackageDescriptor."
There was a problem hiding this comment.
Sorry if I was unclear, I meant why not just pull down the datapackage.json by constructing its URL, rather than querying the API to get its URL -- having it around is clearly useful!
jdangerx
force-pushed
the
new-zenodo-api-bandaid
branch
from
932ec05
to
7db9f07
Compare
Codecov ReportAll modified lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## dev #2942 +/- ##
=====================================
Coverage 88.5% 88.5%
=====================================
Files 90 90
Lines 10797 10800 +3
=====================================
+ Hits 9564 9568 +4
+ Misses 1233 1232 -1
☔ View full report in Codecov by Sentry. |
Closes #2939 .
datapackage.json/ pretend that we were using the "best, most stable URLs" the whole time