ux(community): split Agent Supermarket into LLM and Open Agent shelves - #427
Open
Allan-Feng wants to merge 2 commits into
Open
ux(community): split Agent Supermarket into LLM and Open Agent shelves#427Allan-Feng wants to merge 2 commits into
Allan-Feng wants to merge 2 commits into
Conversation
Replace mixed strategy tiles with compact leaderboard-first cards that reuse the contest board's rank, return, and DJIA equity curves. Co-authored-by: Cursor <cursoragent@cursor.com>
Contributor
|
@Allan-Feng is attempting to deploy a commit to the allan-feng's projects Team on Vercel. A member of the Team first needs to authorize it. |
Correctness:
- formatMarketplaceReturnPct: null/'' is "no data", not 0% — Number(null)
is 0, so a failed leaderboard fetch painted every LLM card with a green
"0.0%" instead of the muted em-dash; sign now chosen after rounding so
-0.04% renders "0.0%", not "-0.0%".
- Mini chart: scale the DJIA path by its own point count — reusing the
agent-length x-scale drew the benchmark past the plot edge whenever the
two downsampled curves differed in length (align_equity_curve_asof can
legitimately return shorter curves).
- loadMarketplaceLeaderboard: only reuse the shared leaderboardPayload
when it is the CONTEST board (js/leaderboard.js writes the same global
for period=live), and reset the cache to null on a failed fetch so the
next Community visit retries instead of pinning blank stats; the
catalog docstring this PR displaced moves back onto loadMarketplace.
- marketplace.json: rename gemini-3-1-pro -> gemini-3-1-pro-preview so
the entry_id fallback join ('-'->'_') actually matches the board's
gemini_3_1_pro_preview; scope the DeepSeek "only model that beat the
baselines" claim to the fixed April 2026 contest window.
Tests:
- Pin the REAL join key: catalog names must equal the board entries'
`model` strings (the API never exposes `name`), plus a new guard that
every llms template_id maps onto its leaderboard entry_id.
- Pin the 7 LLM card prompts to DEFAULT_STARTER_INSTRUCTION.
- Bump app.js/styles.css cache-busters (v125/v131) and their four pins.
Cleanup (dead code this redesign orphaned):
- Remove the vendor-chip subsystem left wired to the deleted container:
marketplaceVendorFilter/setMarketplaceVendorFilter/
renderMarketplaceVendorChips, the dead listener, the unreachable
both-filters empty state, and formatModelProviderLabel; guard tests
now assert the machinery stays gone.
- Drop orphaned CSS (vendor chips, licence badge, clone-split/model
menu, card-body/meta/tag rows) and the no-op 1280px breakpoint.
- marketplace.py: the sort-key fallback branch was unreachable;
templateMarketplaceShelf uses the generic runtime fallback instead of
a per-runtime special case; drop never-read stats fields.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011ww26Uycvhrf4C5jShgQ6r
Collaborator
|
Review pass + fixes pushed as Two review findings deliberately left to the author:
|
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
2 participants
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
GET /api/v1/leaderboard?period=contest(2026-04-15 → 2026-05-15) for rank, cumulative return, and an agent-vs-DJIA mini chart. Open Agent cards stay unranked and show a short description plus the GitHub repo.Add to My Agentsunchanged. Competition Leaderboard APIs and page are untouched.Test plan
#N of 12, green/red Return, model-named curve vs DJIA, andDJIA 30 · Apr 15–May 15 · $10Kvirattt/ai-hedge-fundare visible; Add still clonespytest dashboard/backend/tests/test_frontend_model_facets.py dashboard/backend/tests/test_frontend_shelves.py dashboard/backend/tests/test_marketplace_catalog_models.py dashboard/backend/tests/test_agents_api.py -qMade with Cursor