bobbai00 opened a new pull request, #4496:
URL: https://github.com/apache/texera/pull/4496
### What changes were proposed in this PR?
Stacks on #4495. Adds four tools to `TexeraAgent` so the LLM can edit and
run workflows: `addOperator`, `modifyOperator`, `deleteOperator` (validated
against the operator JSON-schema registry), and `executeOperator` (compiles the
current sub-DAG, runs it, returns sampled rows plus per-port shape metadata).
Adds `/api/execution/{wid}/{cuid}/run` on the JVM backend
(`SyncExecutionResource`) as the sync endpoint that backs `executeOperator`.
Forwards `ENV_AUTH_JWT_SECRET` to spawned computing-unit pods so they validate
the agent's JWT.
After each `addOperator`/`modifyOperator`, the agent re-executes the
affected operator and caches the result on the current step, so the model sees
fresh schemas and previews on its next step.
### Any related issues, documentation, discussions?
Stacked on #4495. Diff against `main` is cumulative until #4495 merges.
### How was this PR tested?
- `bun run typecheck` passes for `agent-service/`.
- Manual smoke: chat message that adds two operators and an edge streams
step events in order; canvas auto-persists within 500ms; `executeOperator`
returns sampled rows; concurrent runs against the same workflow id serialize
via the per-workflow `AsyncMutex`.
### Was this PR authored or co-authored using generative AI tooling?
Generated-by: Claude Code (Claude Opus 4.7)
--
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]
