pierrejeambrun opened a new pull request, #74334:
URL: https://github.com/apache/airflow/pull/74334
Stacked on #74272 -- this PR surfaces the `model_name`/`usage` XComs that PR
adds as a first-class "Model" tab on the task-instance page, instead of leaving
them only reachable through the raw XCom tab.
The tab reads the XComs directly from the core REST API (same-origin, cookie
auth) rather than adding a new backend endpoint, and is scoped via `applies_to`
to the operators that actually publish this data (`LLMOperator`,
`AgentOperator`, and their `@task.llm`/`@task.agent` TaskFlow forms) so it
doesn't show up as a dead end on unrelated tasks.
- Model-only state (LLMOperator / `@task.llm`, which doesn't publish
`usage`):
- placeholder-llm-only
- Full state (AgentOperator, model + token usage + cost):
- placeholder-agent-usage
- Cost tooltip (best-effort pricing explanation):
- placeholder-cost-tooltip
- Tab correctly absent on an unrelated task (applies_to scoping):
- placeholder-scoped-absent
Also fixes a latent bug in `update_providers_dependencies.py`: it sorted
each provider's plugin list by comparing the raw dicts directly, which only
worked because no provider had more than one plugin before this one.
---
##### Was generative AI tooling used to co-author this PR?
- [X] Yes — Claude Code (Opus 4.7)
Generated-by: Claude Code (Opus 4.7) following [the
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
