Don't retrieve anything less than Z10000
Open, MediumPublic
Actions

Assigned To

None

Authored By

	cmassaro
	Feb 10 2025, 6:11 PM

Description

Currently, the orchestrator retrieves Z600 and Z610 for some reason. This makes no sense because these are builtin objects. We should dynamically determine which objects are builtins rather than hard-coding lists.

Desired behavior/Acceptance criteria (returned value, expected error, performance expectations, etc.)

write a function to determine whether a ZID is builtin
use that function in the orchestrator
stop retrieving Z6XX-range stuff
remove test objects that mirror Z6XX-range stuff (e.g., test_data/english_indefinite_noun_phrase/wikilambda/Z4[10].json)

Completion checklist

Before closing this task, review one by one the checklist available here: https://www.mediawiki.org/wiki/Abstract_Wikipedia_team/Definition_of_Done#Back-end_Task/Bug_completion_checklist

Event Timeline

cmassaro created this task.Feb 10 2025, 6:11 PM

Jdforrester-WMF triaged this task as Medium priority.Feb 12 2025, 6:06 PM

Jdforrester-WMF added a project: Performance Issue.

Jdforrester-WMF moved this task from To Triage to Engineering Backlog on the Abstract Wikipedia team board.

https://gitlab.wikimedia.org/repos/abstract-wiki/wikifunctions/function-orchestrator/-/commit/5cbc2c3885fa2f10bf0f92e0a7857ab70e1af7e1

I've tried to remedy this stuff, but it would result in some performance regressions. I don't think now is the time to mess with this further, but I leave the commit and branch there for posterity.

cmassaro updated the task description. (Show Details)Feb 25 2025, 12:21 PM

apine updated https://gitlab.wikimedia.org/repos/abstract-wiki/wikifunctions/function-orchestrator/-/merge_requests/332

Create helper function to log fetches, including time elapsed.

Don't retrieve anything less than Z10000Open, MediumPublicActions

Description

Description

Completion checklist

Event Timeline

Don't retrieve anything less than Z10000
Open, MediumPublic
Actions