Implement label inference pipeline #825
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
GabrielKS
commented
Jun 28, 2021
GabrielKS
commented
- Adds a stage to the intake pipeline to perform label inference
- Includes several placeholder label inference algorithms to test the system
- Unit tests coming soon!
+ Includes a placeholder prediction algorithm returning dummy data. + Adds inferred labels to confirmed trips. Testing done: Ensured that placeholder label predictions were being added to confirmed trip objects. No client-side work or testing done yet.
+ Adds two more placeholder label inference "algorithms" to facilitate front-end testing
shankari
requested changes
Jun 29, 2021
+ The label inference pipeline.py now handles multiple algorithms less awkwardly and has a place for combination of multiple predictions into an ensemble. + The pipeline now stores each algorithm's prediction and a final (perhaps ensemble) prediction in different places Tested by confirming that client-side behavior is the same as before.
+ The label inference pipeline now reads confirmed trips instead of cleaned trips + It then writes inference data directly to confirmed trips instead of relying on matcher.py to find the data in the database Tested by confirming that client-side behavior is the same as before.
shankari
reviewed
Jul 2, 2021
|
The other changes look good. |
The inference pipeline stage now takes cleaned trips and outputs "inferred trips," which the CREATE_CONFIRMED_OBJECTS stage then accepts. Tested by confirming that the client-side behavior is the same.
|
Et voilà, no |
|
This looks fine to me. Are you going to include unit tests now, or in a later PR? |
|
I'll write unit tests now. |
+ TestLabelInferencePipeline.py now tests the plumbing of labels.pipeline.py. It does not test the actual inference algorithms (which have not been implemented for real yet). + The pipeline code has been reworked slightly to fix bugs illuminated by the unit tests and to make placeholder_predictor_2 deterministic Tested by running the unit tests and by confirming that client-side behavior is the same
shankari
requested changes
Jul 8, 2021
shankari
added a commit
to shankari/e-mission-docker
that referenced
this pull request
Sep 16, 2021
This makes it compatible with https://github.com/e-mission/e-mission-server/releases/tag/v3.0.0 in particular, by merging e-mission/e-mission-server#829 and e-mission/e-mission-server#828 e-mission/e-mission-server#825
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.