New digests

[J08nY]

Fixes #420.

[codecov]

Codecov Report

Attention: Patch coverage is 92.70833% with 7 lines in your changes missing coverage. Please review.

Project coverage is 66.83%. Comparing base (c5b4f3d) to head (ee63131).
Report is 6 commits behind head on main.

Files	Patch %	Lines
src/sec_certs/sample/cc_scheme.py	91.08%	5 Missing ⚠️
src/sec_certs/sample/cc.py	91.67%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #424      +/-   ##
==========================================
- Coverage   68.36%   66.83%   -1.53%     
==========================================
  Files          62       62              
  Lines        7579     7553      -26     
==========================================
- Hits         5181     5047     -134     
- Misses       2398     2506     +108     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[J08nY]

@adamjanovsky What do you think about this solution? It moves to what we discussed in #420, using the filename parts of the st and report links. It also keeps the old digest attributes, perhaps a better naming there would be possible. I was thinking of having "alt_dgsts" or "old_dgsts" property that would have the older digests and would be a set. What do you think? This is necessary to be able to map the old to the new on the site (as well as for other of our users for some kind of compat).

[adamjanovsky]

Well, late to the review I guess :)

Anyways, I agreed with the proposal, didn't go through the implementation in the detail. One think to keep in mind is our usage of digests in Notebooks. I suppose that the backward-compatability redirect only applies to web, right?

What's going to happen when you load an old dataset with a new tool version?

[J08nY]

Anyways, I agreed with the proposal, didn't go through the implementation in the detail. One think to keep in mind is our usage of digests in Notebooks. I suppose that the backward-compatability redirect only applies to web, right?

Yes right now. However, you can get the old digests from the certs.

What's going to happen when you load an old dataset with a new tool version?

Interesting thought, haven't considered it much as I was looking at it from the POV of the web where a fresh run is running. However, I tested this a bit and what happens is that the new digests take precedence. So after loading it you get a dataset with new digests only (because dgst is a property computed from attributes) and if you store it again you will get the new digests also in the JSON. Basically, the old digests are ignored.

Also, regarding the quick merge. I wanted to get this in quickly because of the Monday update run. The web has been broken for a while and I wanted to fix it. This was a pre-requisite.

[adamjanovsky]

Interesting thought, haven't considered it much as I was looking at it from the POV of the web where a fresh run is running. However, I tested this a bit and what happens is that the new digests take precedence. So after loading it you get a dataset with new digests only (because dgst is a property computed from attributes) and if you store it again you will get the new digests also in the JSON. Basically, the old digests are ignored.

Cool, that's what I anticipated.

Also, regarding the quick merge. I wanted to get this in quickly because of the Monday update run. The web has been broken for a while and I wanted to fix it. This was a pre-requisite.

Understood 👍