chore(pipeline): reuse schema validation in pipeline #290
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
The trick is to leverage plpython to do the geocoding inside the database. Doing so, geocoding can now be modelled as a dbt model and orchestrate as such. The geocoding implementation has been moved to an actual package maintained next to the datawarehouse image. The plpython udf simply wraps the call. Is it worth it ? * it heavely simplifies the flow and set clearer concerns between airflow and dbt. Dbt does the transformation, airflow orchestrate it. * less error prone since we do not have to pull data from the db and map it to python ourselves. * geocoding now can simply be done per source. This would have been terribly cumbersome. * we can even leverage dbt to to the geocoding incrementally on inputs that have not been seen before. This will drastically reduce our carbon footprint... There are a few enhancements we would probably want : * obviously clean up * define a macro with the geocoding model that can be used for all sources * re-do geocodings for relatively old inputs
Currently validation differs between the pipeline and the api. The api relies on the schema validation, using pydantic. But the pipeline has its own sql-based validation. This can lead to inconsistencies between pipeline and api and duplicates code. This commit leverages plpython to reuse the schema validation in the pipeline. Validation is done at the source level. This commit also materializes validation errors in a dedicated model, rather than using dbt data tests, which are less convenient for products to use.
vperron
reviewed
Sep 12, 2024
| $(PIP_COMPILE) --extra=dev --output-file=requirements/dev-requirements.txt | ||
|
|
||
| test: | ||
| $(PIP_COMPILE) --extra=test --output-file=requirements/test-requirements.txt |
| FROM | ||
| structures, | ||
| LATERAL (SELECT * FROM VALIDATE('structure', TO_JSONB(structures))) AS errors | ||
| ), |
There was a problem hiding this comment.
Je vais etre tellement heureux de voir ces tables, c'est si pénible de débugger ce genre de trucs actuellement...
| OR {{ this }}.input != ARRAY[adresses.adresse, adresses.code_postal, adresses.commune] | ||
| {% endif %} | ||
| ) | ||
| ) AS geocodings ON adresses._di_surrogate_id = geocodings.adresse_id |
There was a problem hiding this comment.
comme dit quelque part, un unit test de ce code me rassurerait grandement
hlecuyer
approved these changes
Oct 3, 2024
vperron
reviewed
Oct 3, 2024
| data_tests: | ||
| - check_structure: | ||
| config: | ||
| severity: warn |
There was a problem hiding this comment.
faudra pas oublier de retirer les macros associées. Quid du check_adresses ?
| errors.input AS "input" | ||
| FROM | ||
| structures, | ||
| LATERAL (SELECT * FROM VALIDATE('structure', TO_JSONB(structures))) AS errors |
There was a problem hiding this comment.
question pour plus tard : quand on aura de la validation plus poussée, tu verras les choses comment ? On met tout dans le validate plpython qui appellerait par exemple, la validation pydantic + des validations HTTP par exemple ?
Ou plutot des etapes supplémentaires comme le geocoding (pour moi je penserais plutot à ça)
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.
depends on #270