Repository navigation
feat(app): agent chat page, result card and .adapt() button - #12114
Merged
Merged
Conversation
…() button Admin-only chat page at /debug/agent?spec=&library=&language= over the /debug/agent BFF: data panel (200 KB counter, parse, 20-row preview, warnings, one dropdown per spec role with the server's defaults), chat thread with the fixed refusal and errors with code and ref, a progress timeline and queue counter from the anyplot/1 status events, Stop (AbortController plus the cancel route) and the library pills. ResultCard per version: the rendered PNG from its blob URL, a light/dark switch through the theme toggle route, copy image (Clipboard API with a download fallback), download PNG, open full size, plot.py in CodeHighlighter with copy code, downloads of plot.py and data.csv, the change list, residual notes, a refine composer and a feedback slot. lib/sse.ts parses SSE over fetchWithAuth's body reader (pings, CRLF, split chunks); hooks/useAgentSession.ts holds the session state machine. The .adapt() overlay button shows only with VITE_ENABLE_AGENT_CHAT, the admin hint DebugPage now sets after /debug/status, and an eligible pair. Without the flag the chunk is not built. Analytics are enum-only (agent_open, agent_data_parsed, agent_plot_rendered, agent_guardrail_block, copy_code page agent_chat) and documented in plausible.md. app/scripts/agent-bff-mock.mjs serves the BFF routes and a scripted stream locally for browser verification. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GdMSqLR5ww4ji74EUSmk9M
…actions
Review fixes for the "Use with my data" chat page:
- The plot event carries `version`, the number the agents service stored
the result under (PlotResult.version, PLOT_FIELDS, the BFF allowlist);
the page addresses the artifact and theme toggle routes by it and counts
plot events only against a server without the field.
- The dataset answer lists the spec's roles {name, kinds, required,
variadic, description}; the data panel gives every role a dropdown and a
variadic family one per member plus one for the next, kept contiguous.
A refused binding set keeps the check's lines (at most 20) through the BFF.
- The page starts no turn, parse, binding change, library switch or theme
render while another is in flight; Stop keeps the stream until the
server's done (at most 30 s), so a late plot still arrives and the next
turn does not meet 409 run_active.
- The session is deleted on pagehide with keepalive; a request without an
answer is `unreachable`, reloads once for Cloudflare Access (the guard is
shared with DebugPage) and then offers a reload link.
- An ineligible pair offers the other libraries as links.
- The mock BFF listens on loopback only, serves line-multi with a series
family, the roles, the error lines and the version; README and design doc
name AGENT_USER_ID_KEY and ADMIN_TOKEN for a local run.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GdMSqLR5ww4ji74EUSmk9M
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GdMSqLR5ww4ji74EUSmk9M
Codecov Report❌ Patch coverage is 📢 Thoughts on this report? Let us know! |
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
The mock does not preserve version artifacts, language inputs can become inconsistent, and the SSE line guard is bypassable.
4 open findings
What changed in this PR
Adds the feature-flagged, admin-only agent chat UI, result cards, .adapt() entry point, and supporting agent/BFF contract updates.
Changes:
- Adds the complete agent-chat frontend and local mock.
- Adds server-provided plot versions, roles, and binding diagnostics.
- Documents, tests, and deploy-gates the feature.
| File | Description |
|---|---|
agents/README.md |
Documents local chat setup and mock usage. |
agents/anyplot/pipeline.py |
Assigns stored version numbers to plot results. |
agents/anyplot/schemas.py |
Adds validated plot-result versions. |
agents/main.py |
Returns role metadata with parsed datasets. |
agents/stream.py |
Includes versions in plot events. |
api/routers/agent.py |
Relays versions and sanitized binding errors. |
app/ARCHITECTURE.md |
Documents agent-chat frontend structure. |
app/Dockerfile |
Adds the chat build argument. |
app/cloudbuild.yaml |
Keeps the production chat build disabled. |
app/scripts/agent-bff-mock.mjs |
Implements the local BFF mock. |
app/src/global-config.ts |
Defines the build-time feature flag. |
app/src/hooks/useAgentEligibility.test.ts |
Tests eligibility gating. |
app/src/hooks/useAgentEligibility.ts |
Gates .adapt() eligibility requests. |
app/src/hooks/useAgentSession.test.ts |
Tests chat session behavior. |
app/src/hooks/useAgentSession.ts |
Implements the chat state machine. |
app/src/lib/agent.test.ts |
Tests the BFF client and protocol parsing. |
app/src/lib/agent.ts |
Adds the typed agent-chat client. |
app/src/lib/sse.test.ts |
Tests incremental SSE parsing. |
app/src/lib/sse.ts |
Adds fetch-based SSE parsing. |
app/src/pages/AgentChatPage.test.tsx |
Tests page gating and startup. |
app/src/pages/AgentChatPage.tsx |
Adds the agent-chat page. |
app/src/pages/DebugPage.test.tsx |
Tests admin-hint updates. |
app/src/pages/DebugPage.tsx |
Shares admin authentication state. |
app/src/pages/SpecPage.tsx |
Connects eligible plots to the chat. |
app/src/routes/index.tsx |
Registers the feature-gated route. |
app/src/routes/paths.test.ts |
Tests agent-chat URL generation. |
app/src/routes/paths.ts |
Adds the agent-chat path helper. |
app/src/sections/agent-chat/ChatThread.tsx |
Renders conversation items and results. |
app/src/sections/agent-chat/DataPanel.test.tsx |
Tests role-binding controls. |
app/src/sections/agent-chat/DataPanel.tsx |
Adds data parsing, preview, and bindings. |
app/src/sections/agent-chat/ProgressTimeline.tsx |
Displays queue and pipeline progress. |
app/src/sections/agent-chat/ReloadHint.tsx |
Adds recoverable-error reload actions. |
app/src/sections/agent-chat/ResultCard.test.tsx |
Tests result actions and refinement. |
app/src/sections/agent-chat/ResultCard.tsx |
Adds plot, artifact, theme, and code controls. |
app/src/sections/agent-chat/bindings.ts |
Builds binding rows for role families. |
app/src/sections/agent-chat/files.ts |
Adds clipboard and download helpers. |
app/src/sections/agent-chat/messages.ts |
Maps protocol errors to UI copy. |
app/src/sections/agent-chat/styles.ts |
Defines shared chat styles. |
app/src/sections/spec-detail/SpecDetailView.test.tsx |
Tests .adapt() visibility and interaction. |
app/src/sections/spec-detail/SpecDetailView.tsx |
Adds the .adapt() overlay action. |
app/src/utils/adminAuth.ts |
Centralizes admin token and hint storage. |
app/src/vite-env.d.ts |
Types the feature flag. |
changelog.d/agent-chat-ui.md |
Records the feature and contract changes. |
docs/concepts/agent-network.md |
Updates the implementation status and design. |
docs/reference/api.md |
Documents roles, versions, and binding errors. |
docs/reference/plausible.md |
Documents agent-chat analytics. |
docs/reference/repository.md |
Adds the new frontend structure. |
tests/unit/agents/runtime/test_service_flow.py |
Tests roles and version sequencing. |
tests/unit/agents/runtime/test_stream.py |
Tests role serialization and plot versions. |
tests/unit/agents/test_schemas.py |
Tests version validation rules. |
tests/unit/api/test_agent_router.py |
Tests BFF filtering and error relaying. |
🧠 Review effort: Balanced
💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
…lain action CTA, mock versions snapshot their data Copilot's four findings on #12114: - The SSE parser checked the line limit only on the unfinished remainder of the buffer, so an oversized line whose newline arrived in the same chunk was parsed. The limit now applies to every line before it is parsed; two tests cover the terminated oversized line and a line at the limit. - The chat page took `language` from the query independently of `library`, so a mismatched pair picked the wrong highlighter and a back link to a page that does not exist. The language now follows the library through `LIB_TO_LANG`, as the catalogue pairs them. - `.create_plot()` used the filled hero CTA, which the style guide reserves for the landing hero. It is now the regular method-call action with a rule border; `ctaButtonSx` is gone. - The mock's versions read the session's current dataset and bindings, so an earlier result card served the new data after a reparse. A version now keeps the dataset and bindings it was made from, and its code, CSV and later theme renders read from that snapshot. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GdMSqLR5ww4ji74EUSmk9M
# Conflicts: # docs/concepts/agent-network.md
# Conflicts: # agents/README.md # docs/concepts/agent-network.md # tests/unit/agents/runtime/test_stream.py
1 of 2 tasks
MarkusNeusinger
added a commit
that referenced
this pull request
Oct 10, 2026
## Summary - The tool-safety allowlist for the `plot_pipeline` result did not include `version` (added to `PlotResult` in #12114 for the chat UI), so every shipped plot reached the root as `invalid_result` and the root's closing reply told the user the run had failed while the plot was already on screen. Seen in the first local run through the `/debug/agent` gate. - The allowlist is now derived from `PlotResult.model_fields` plus the tool-error shape, so a schema field can never silently break the root again; a plugin test passes a shipped result with `version=1` through the callback and asserts the schema's fields stay allowed. ## Plan N/A (bug found while bringing the stack up locally; the eval harness never checks the root's text). ## Test plan - [x] `ruff check`, `ruff format --check`, `mypy agents`, `pytest tests/unit/agents/runtime/test_plugins.py tests/unit/agents/runtime/test_service_flow.py` (85 passed), changelog check - [ ] CI green; after merge the local run's root reply reads the shipped result ("your plot is ready" instead of "internal error") 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01GdMSqLR5ww4ji74EUSmk9M --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
/debug/agent?spec=&library=&language=(app/src/pages/AgentChatPage.tsx): data panel with a 200 KB counter, 20-row preview, a column dropdown for every spec role (series families get one dropdown per member), the chat thread, a progress timeline that follows theanyplot/1stream, a queue-position counter in the chat's corner,.stop()through the cancel route, and library pills. Built only withVITE_ENABLE_AGENT_CHAT=true(off inapp/cloudbuild.yaml, so production ships without the chunk) and opens only in local development or for a browser carrying the admin hint.app/src/sections/agent-chat/ResultCard.tsx) with the plot page's overlay actions (copy image, download PNG, open full size), a light/dark switch through the theme toggle route, the adaptedplot.pywith copy and download,data.csvdownload, the change list and residual notes, a refine composer and the quick-feedback slot; plus the.adapt()overlay button on the plot page, gated by build flag, admin hint and the BFF's eligibility route so public visitors never call the debug API.plotevent carriesversion(agents translator and BFF allowlist), and the dataset answer lists the spec'sroleswith a 422 that keeps the binding check's lines. Analytics events (enum properties only) are documented indocs/reference/plausible.md; a loopback mock of the BFF (app/scripts/agent-bff-mock.mjs) drives the whole flow without the API, the agents service or a model.Plan
Design:
docs/concepts/agent-network.md(sections "Frontend", "Result card", "Analytics"); this is the frontend PR of the phase-1 roadmap, built on the merged runtime core (#12111) and run queue (#12112).Test plan
cd app && yarn lint && yarn fm:check && yarn type-check && yarn test(81 files, 750 tests) andyarn buildwith the flag on and off; the default build contains no agent clientuv run --extra dev ruff check . && ruff format --check .,mypy api core agents,pytest tests/unit tests/integration(7816 passed), changelog check,uv lock --checkplot.py/data.csv, open full size, copy code), theme switch, repair round, refusal, error, stop then a new turn without a 409, ineligible pair; no horizontal scroll, 16 px gutter at 390 pxdeploy-appanddeploy-apiCloud Builds succeed; production stays unchanged (_VITE_ENABLE_AGENT_CHATfalse,AGENT_ENABLEDfalse)/debug/agent?spec=scatter-basic&library=matplotlibas admin and run one "Create plot" end to end