search

nycmeshnet / meshdb

A convenient, stable, and sane database for tracking Members and Nodes for use with robots and humans
https://db.nycmesh.net
MIT License
12 stars 14 forks source link

Add ability to replay join records #668

Closed WillNilges closed 5 days ago

WillNilges commented 3 weeks ago

This isn't quite done, but the general idea is there.

It adds a command you can use to replay join records. You'll have to exec into a pod to do it right now, but I can probs expose this via a web interface at some point.

It has a basic mock test, but I need to make it more robust before I trust it.

WillNilges commented 3 weeks ago

image

WillNilges commented 3 weeks ago

https://github.com/nycmeshnet/meshforms/pull/93

WillNilges commented 3 weeks ago

Had some thoughts on the train. I should include install number as the indicator or success.

codecov[bot] commented 3 weeks ago

Codecov Report

Attention: Patch coverage is 90.00000% with 11 lines in your changes missing coverage. Please review.

Project coverage is 94.06%. Comparing base (5caf101) to head (dc3995d). Report is 1 commits behind head on main.

Files with missing lines Patch % Lines
...meshapi/management/commands/replay_join_records.py 84.90% 8 Missing :warning:
src/meshapi/util/join_records.py 94.54% 3 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## main #668 +/- ## ========================================== - Coverage 94.19% 94.06% -0.14% ========================================== Files 80 82 +2 Lines 3293 3403 +110 ========================================== + Hits 3102 3201 +99 - Misses 191 202 +11 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

WillNilges commented 2 weeks ago

I would like to specify more rigid types and ensure parity between the join record format in JS and Python. They both read and write from the bucket so it's imperative that they're the same.

Also, I haven't addressed the problem that we're going to be getting a lot of these so the list will grow quite large.

WillNilges commented 2 weeks ago

Theoretically the bucket can grow quite large. Maybe one thing I could do is try to improve the way I fetch the failed submissions.

Then again, computers are fast, and the sub-optimal code might not matter until much, much later.

WillNilges commented 2 weeks ago

why is codecov mad I literally fixed it it literally has the wrong commit.

WillNilges commented 2 weeks ago

Added an issue for the post-launch stuff we need to do: https://github.com/nycmeshnet/meshforms/issues/102

WillNilges commented 2 weeks ago

i got rid of the mock, it was causing more problems than it was solving so now this is an integration test. Sue me and die.

WillNilges commented 1 week ago

I implemented start time.

Not going to do trust me bro yet in the interest of time and energy.

https://github.com/nycmeshnet/meshdb/issues/699