Add /api/data JSON endpoint (3 tracks incl. downstream) + CORS

app.py

@@ -2,8 +2,8 @@
 
 Computes the leaderboard live from the per-user substrate in the
 ``MyHeartCounts/OpenMHC-leaderboard-data`` HF dataset (see
-``leaderboard_compute.py``) and serves it as an HTML page. There is no JSON
-API — this Space *is* the leaderboard UI.
+``leaderboard_compute.py``) and serves it as an HTML page. The same data is
+also exposed as JSON at ``/api/data`` (CORS-enabled) for the public site.
 
 The dataset is public, so no token is required. Styling mirrors the public
 MyHeartCounts / OpenMHC site (light theme, red accent). The table is grouped by
@@ -13,10 +13,12 @@ sub-track (single-day / long-context) and can be filtered by method type.
 from __future__ import annotations
 
 import html
+import math
 from pathlib import Path
 
 from fastapi import FastAPI
-from fastapi.responses import FileResponse, HTMLResponse
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import FileResponse, HTMLResponse, JSONResponse
 
 from leaderboard_compute import (
     compute_downstream_rows,
@@ -26,6 +28,14 @@ from leaderboard_compute import (
 
 app = FastAPI(title="OpenMHC Leaderboard", docs_url=None, redoc_url=None)
 
+# The public site (myheartcounts.stanford.edu) reads /api/data cross-origin.
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["GET"],
+    allow_headers=["*"],
+)
+
 # Compute is mildly expensive (download + reduce); cache per track in-process.
 _CACHE: dict = {
     "downstream": {"rows": None, "error": None},
@@ -681,6 +691,66 @@ def health() -> dict:
     return {"status": "ok" if ok else "error", "tracks": tracks}
 
 
+# ---------------------------------------------------------------------------
+# JSON API — consumed cross-origin by the public MyHeartCounts site.
+# ---------------------------------------------------------------------------
+
+
+def _sanitize(row: dict) -> dict:
+    """Replace non-finite floats (NaN/inf) with None so the row is valid JSON.
+
+    Starlette's JSONResponse serialises with ``allow_nan=False``; an unsanitised
+    NaN would 500 the endpoint and break the client's ``.json()``.
+    """
+    return {
+        k: (None if isinstance(v, float) and not math.isfinite(v) else v)
+        for k, v in row.items()
+    }
+
+
+def _api_subtracks(cfg: dict, rows: list[dict]) -> list[tuple[str, str]]:
+    """Subtracks for the payload; mirrors the HTML's dynamic "Other" bucket.
+
+    The frontend silently drops rows whose ``subtrack`` matches no subtrack key,
+    so when any row falls outside the configured set we append ("other", "Other")
+    — exactly as ``_table`` does for the HTML page.
+    """
+    subs = list(cfg["subtracks"])
+    if subs:
+        known = {k for k, _ in subs}
+        if any(r.get("subtrack") not in known for r in rows):
+            subs = subs + [("other", "Other")]
+    return subs
+
+
+def _track_payload(cfg: dict, rows: list[dict] | None, error: str | None) -> dict:
+    # `fallback` is a downstream-only column in the public API contract.
+    columns = [
+        {"key": k, "label": label}
+        for k, label in cfg["columns"]
+        if not (k == "fallback" and cfg["key"] != "downstream")
+    ]
+    rows = rows or []
+    return {
+        "title": cfg["title"],
+        "tab": cfg["tab"],
+        "columns": columns,
+        "subtracks": [{"key": k, "label": label} for k, label in _api_subtracks(cfg, rows)],
+        "legend_html": cfg["note"],
+        "rows": [_sanitize(r) for r in rows],
+        "error": error,
+    }
+
+
+@app.get("/api/data")
+def api_data() -> JSONResponse:
+    payload = {
+        cfg["key"]: _track_payload(cfg, *_rows(cfg["key"], cfg["compute"]))
+        for cfg in TRACKS
+    }
+    return JSONResponse(payload)
+
+
 @app.get("/logo.png")
 def logo() -> FileResponse:
     return FileResponse(Path(__file__).parent / "logo.png", media_type="image/png")