Files
supabase/apps/studio/evals/output.ts
T
Matt Rossman 517171b246 feat(assistant): online evals support and CI workflows (#43194)
Lays groundwork for online evals on Assistant chat logs.

https://www.braintrust.dev/docs/observe/score-online

### Changes

- New workflows:
- `braintrust-scorers-deploy.yml` keeps prod scorers in sync on push to
`master`
- `braintrust-preview-scorers-deploy.yml` deploys preview scorers to the
staging project for PRs labeled `preview-scorers`, posting a comment
with scorer links
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000097222))
- `braintrust-preview-scorers-cleanup.yml` deletes preview scorers when
the PR is closed
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000749847))
- Adds `evals/scorer-online.ts` entry point invoked with `pnpm
scorers:deploy`, registering scorers for online evals in the Braintrust
"Assistant" project
- Refactors scorer code to separate online-compatible scorers
(`scorer-online.ts`) from WASM-dependent ones (`scorer-wasm.ts`)
- "URL Validity" scorer now only checks Supabase domains to prevent
requests to untrusted origins
- Span `input` is now shaped `{ prompt: string }` instead of plain
`string` for compatibility with offline eval scorers
- Env vars `BRAINTRUST_STAGING_PROJECT_ID` and `BRAINTRUST_PROJECT_ID`
configured in GitHub repo settings
- `generateAssistantResponse` now uses `startSpan` + `withCurrent`
instead of `traced()` to manually manage the root span lifecycle — this
ensures `onFinish` logs output to the span _before_ `span.end()` is
called, which is when Braintrust triggers scoring automations

### Online Scorers

We share scoring logic across offline and online evals, but some of our
scorers aren't transferrable to an "online" setting due to runtime
challenges or ground truth requirements.

**Supported**
- Goal Completion
- Conciseness
- Completeness
- Docs Faithfulness
- URL Validity

**Unsupported**
- Correctness (requires ground truth output)
- Tool Usage (requires ground truth requiredTools)
- SQL Syntax (uses libpg-query WASM)
- SQL Identifier Quoting (uses libpg-query WASM)
 
### How to use these scorers

Going forward if you want to add/edit online eval scorers, add the
`preview-scorers` label to a PR. This deploys scorers to the [Assistant
(Staging
Scorers)](https://www.braintrust.dev/app/supabase.io/p/Assistant%20(Staging%20Scorers)?v=Overview)
project in Braintrust with branch-specific slugs, and comments on the PR
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000097222)).
From the Braintrust dashboard you can "Test" the scorer with traces from
any project.

<img width="1866" height="528" alt="CleanShot 2026-03-05 at 15 15 00@2x"
src="https://github.com/user-attachments/assets/4f15cebc-3f2d-4e8a-9ee2-fe8ef7bf4199"
/>

Once merged, scorers are deployed to the primary
[Assistant](https://www.braintrust.dev/app/supabase.io/p/Assistant)
project, and preview scorers are deleted from the staging project. Down
the road, scorers on the Assistant project will run automatically on a
sample of production traces.

Closes AI-437
2026-03-09 13:05:26 -04:00

67 lines
1.9 KiB
TypeScript

import { type ToolSet, type TypedToolCall, type TypedToolResult } from 'ai'
import { type AssistantEvalOutput } from './scorer'
type Step = {
text: string
toolCalls: TypedToolCall<ToolSet>[]
toolResults: TypedToolResult<ToolSet>[]
}
type ParsedToolCall = {
/** Query generated by `execute_sql` */
sqlQuery?: string
/** Docs text pulled in from `search_docs` */
docs?: string[]
}
function parseToolCall(
toolCall: TypedToolCall<ToolSet>,
toolResult: TypedToolResult<ToolSet>
): ParsedToolCall {
switch (toolCall.toolName) {
case 'execute_sql': {
const sqlQuery = toolCall.input?.sql
if (typeof sqlQuery !== 'string') return {}
return { sqlQuery }
}
case 'search_docs': {
const content = toolResult.output?.content
if (!content || !Array.isArray(content)) return {}
const docs = content.map((item) => item?.text).filter((text) => typeof text === 'string')
if (docs.length === 0) return {}
return { docs }
}
}
return {}
}
export function buildAssistantEvalOutput(
finishReason: AssistantEvalOutput['finishReason'],
steps: Step[]
): AssistantEvalOutput {
const simplifiedSteps = steps.map((step) => ({
text: step.text,
toolCalls: step.toolCalls.map((call) => ({
toolName: call.toolName,
input: call.input,
})),
}))
const toolNames: string[] = []
const sqlQueries: string[] = []
const docs: string[] = []
for (const step of steps) {
for (const [i, toolCall] of step.toolCalls.entries()) {
toolNames.push(toolCall.toolName)
const toolResult = step.toolResults.at(i)
if (!toolResult) continue
const parsed = parseToolCall(toolCall, toolResult)
if (parsed.sqlQuery) sqlQueries.push(parsed.sqlQuery)
if (parsed.docs) docs.push(...parsed.docs)
}
}
return { finishReason, steps: simplifiedSteps, toolNames, sqlQueries, docs }
}