Page MenuHomePhabricator

Normalize Inputs to Orchestrator
Closed, ResolvedPublic

Event Timeline

DVrandecic triaged this task as Medium priority.Apr 21 2021, 4:16 PM

This will involve a couple of steps:

  • update function-orchestrator's .gitmodules to point to the most recent version of function-schemata; the update_local_submodule.sh script will help
    • this is needed to ensure that function-schemata's normalize() method is available
  • function calls start in src/orchestrate.js; the isFunctionCall method and caller need to be updated to account for the case where input is canonical or mixed format, so that is where normalization will go
    • I think the cleanest thing here is to add an argument to the orchestrate query (e.g. do_normalize) and update the API call and UI in the WikiLambda extension accordingly
    • otherwise, the canonical-form Z1 validator from function-schemata can tell whether the input is in canonical form
  • once function-orchestrator patch is submitted, it may be necessary to update the docker-registry container in extensions/WikiLambda/docker-compose-sample.yml
DVrandecic raised the priority of this task from Medium to High.May 5 2021, 4:38 PM

Change 685489 had a related patch set uploaded (by Lindsay Wardell; author: Lindsay Wardell):

[mediawiki/services/function-orchestrator@master] Normalize input to orchestrator

https://gerrit.wikimedia.org/r/685489

Change 685489 merged by jenkins-bot:

[mediawiki/services/function-orchestrator@master] Normalize input to orchestrator

https://gerrit.wikimedia.org/r/685489