Update app.py

app.py

@@ -2,13 +2,19 @@ import json
 import numpy as np
 import tensorflow as tf
 import gradio as gr
+from fastapi import FastAPI, HTTPException
+from typing import Dict, Any
 
-# ==== CONFIG ====
-MODEL_PATH = "best_model.h5"
-STATS_PATH = "means_std.json"
+# =========================
+# Config
+# =========================
+MODEL_PATH = "best_model.h5"          # your uploaded model
+STATS_PATH = "means_std.json"         # {"feature": {"mean": x, "std": y}, ...}
 CLASSES = ["Top", "Mid-Top", "Mid", "Mid-Low", "Low"]
-# ================
 
+# =========================
+# Load artifacts
+# =========================
 print("Loading model and stats...")
 model = tf.keras.models.load_model(MODEL_PATH, compile=False)
 
@@ -18,57 +24,124 @@ with open(STATS_PATH, "r") as f:
 FEATURES = list(stats.keys())
 print("Feature order:", FEATURES)
 
-def zscore(val, mean, sd):
+# =========================
+# Helpers
+# =========================
+def _zscore(val: Any, mean: float, sd: float) -> float:
     try:
-        val = float(val)
+        v = float(val)
     except Exception:
         return 0.0
-    if sd == 0 or sd is None:
+    if sd is None or sd == 0:
         return 0.0
-    return (val - mean) / sd
+    return (v - mean) / sd
 
-def coral_probs_from_logits(logits_np):
+def _coral_probs_from_logits(logits_np: np.ndarray) -> np.ndarray:
+    """
+    logits_np: (N, K-1) linear outputs.
+    Returns probabilities (N, K) with p_k = σ(z_{k-1}) - σ(z_k), and boundaries 1/0.
+    """
     logits = tf.convert_to_tensor(logits_np, dtype=tf.float32)
-    sig = tf.math.sigmoid(logits)
-    left = tf.concat([tf.ones_like(sig[:, :1]), sig], axis=1)
+    sig = tf.math.sigmoid(logits)                    # (N, K-1)
+    left  = tf.concat([tf.ones_like(sig[:, :1]), sig], axis=1)
     right = tf.concat([sig, tf.zeros_like(sig[:, :1])], axis=1)
     probs = tf.clip_by_value(left - right, 1e-12, 1.0)
     return probs.numpy()
 
-def predict_from_json(input_json):
-    # Expect a dictionary with feature:value pairs
-    if not isinstance(input_json, dict):
-        return {"error": "Expected JSON object mapping feature -> value."}
+def _predict_core(ratios: Dict[str, Any]) -> Dict[str, Any]:
+    """
+    ratios: dict mapping feature -> raw numeric value.
+    Returns: dict with predicted_state, probabilities, z_scores, missing, input_ok.
+    """
+    # Validate presence (we still accept missing and fill 0.0 after z-score)
+    missing = [f for f in FEATURES if f not in ratios]
 
-    # Prepare data
-    z_list = []
+    # Build z-score vector in exact FEATURE order
+    z_list, z_scores = [], {}
     for f in FEATURES:
-        v = input_json.get(f, 0)
-        z = zscore(v, stats[f]["mean"], stats[f]["std"])
+        z = _zscore(ratios.get(f, 0.0), stats[f]["mean"], stats[f]["std"])
         z_list.append(z)
-    X = np.array([z_list], dtype=np.float32)
+        z_scores[f] = z
 
+    X = np.array([z_list], dtype=np.float32)  # (1, D)
     raw = model.predict(X, verbose=0)
+
+    # Softmax (K) vs CORAL (K-1)
+    if raw.ndim != 2:
+        raise ValueError(f"Unexpected model output shape: {raw.shape}")
     if raw.shape[1] == len(CLASSES) - 1:
-        probs = coral_probs_from_logits(raw)[0]
+        probs = _coral_probs_from_logits(raw)[0]  # (K,)
+    elif raw.shape[1] == len(CLASSES):
+        probs = raw[0]                             # (K,)
+    else:
+        raise ValueError(f"Model output width {raw.shape[1]} incompatible with classes {len(CLASSES)}")
+
+    # Safety: normalize if not a perfect prob. vector
+    probs = np.maximum(probs, 0.0)
+    s = probs.sum()
+    if s <= 0:
+        # fallback uniform if something pathological happens
+        probs = np.ones(len(CLASSES), dtype=np.float32) / float(len(CLASSES))
     else:
-        probs = raw[0]
+        probs = probs / s
 
     pred_idx = int(np.argmax(probs))
-    output = {
+    return {
+        "input_ok": len(missing) == 0,
+        "missing": missing,
+        "z_scores": z_scores,
         "probabilities": {CLASSES[i]: float(probs[i]) for i in range(len(CLASSES))},
         "predicted_state": CLASSES[pred_idx],
     }
-    return output
 
-# ==== Gradio UI ====
+# =========================
+# Gradio adapter (UI)
+# =========================
+def _gradio_adapter(payload):
+    """
+    Accepts either:
+      - a dict {feature: value, ...}
+      - a list with one dict [ {feature: value, ...} ]
+    """
+    if isinstance(payload, list) and len(payload) == 1 and isinstance(payload[0], dict):
+        payload = payload[0]
+    if not isinstance(payload, dict):
+        return {"error": "Expected JSON object mapping feature -> value."}
+    return _predict_core(payload)
+
 demo = gr.Interface(
-    fn=predict_from_json,
-    inputs=gr.JSON(label="Feature dictionary (JSON)"),
+    fn=_gradio_adapter,
+    inputs=gr.JSON(label="ratios JSON (dict of feature -> value)"),
     outputs="json",
-    title="Static Fingerprint Predictor",
-    description="POST JSON to /run/predict with your feature values."
+    title="Static Fingerprint Model API",
+    description="Programmatic use: POST a raw dict to /predict. UI here is for quick manual checks.",
+    allow_flagging="never"
 )
 
+# =========================
+# FastAPI app (sync endpoint)
+# =========================
+api = FastAPI()
+
+@api.get("/health")
+def health():
+    return {"status": "ok", "features": FEATURES, "classes": CLASSES}
+
+@api.post("/predict")
+def predict_endpoint(payload: Any):
+    # Allow list-of-one and dict
+    if isinstance(payload, list) and len(payload) == 1 and isinstance(payload[0], dict):
+        payload = payload[0]
+    if not isinstance(payload, dict):
+        raise HTTPException(status_code=400, detail="Expected JSON object mapping feature -> value.")
+    try:
+        return _predict_core(payload)
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+# Mount Gradio UI at "/" and expose FastAPI routes alongside it
+app = gr.mount_gradio_app(api, demo, path="/")
+
 if __name__ == "__main__":
-    demo.launch()
+    # local dev run (HF Spaces will ignore this and use its own server)
+    demo.launch()