Create app.py

app.py

@@ -0,0 +1,201 @@
+import gradio as gr
+import os
+import json
+import shutil
+from difflib import get_close_matches
+from functools import lru_cache
+
+from inference_sdk import InferenceHTTPClient
+
+# =====================
+# LOAD ENV
+# =====================
+ROBOFLOW_API_KEY = os.getenv("ROBOFLOW_API_KEY")
+
+# =====================
+# LOAD NUTRITION DB
+# =====================
+with open("nutrition_db.json", "r") as f:
+    NUTRITION_DB = json.load(f)
+
+print("✅ Loaded DB:", len(NUTRITION_DB))
+
+# =====================
+# ROBOFLOW CLIENT
+# =====================
+rf_client = InferenceHTTPClient(
+    api_url="https://serverless.roboflow.com",
+    api_key=ROBOFLOW_API_KEY
+)
+
+# =====================
+# NORMALIZATION
+# =====================
+def normalize_food_name(name):
+    name = name.lower()
+
+    mapping = {
+        "chapati": "wheat",
+        "roti": "wheat",
+        "naan": "wheat",
+        "paratha": "wheat",
+        "omelette": "egg",
+        "omellete": "egg",
+        "fried rice": "rice",
+        "plain rice": "rice"
+    }
+
+    return mapping.get(name, name)
+
+# =====================
+# FIND MATCH
+# =====================
+def find_food(name):
+    if name in NUTRITION_DB:
+        return name
+
+    matches = get_close_matches(name, NUTRITION_DB.keys(), n=1, cutoff=0.6)
+    return matches[0] if matches else None
+
+# =====================
+# FALLBACK
+# =====================
+def estimate_unknown_food(grams):
+    return {
+        "calories": round(1.5 * grams, 2),
+        "protein": round(0.05 * grams, 2),
+        "carbs": round(0.2 * grams, 2),
+        "fat": round(0.05 * grams, 2),
+    }
+
+# =====================
+# NUTRITION CACHE
+# =====================
+@lru_cache(maxsize=1000)
+def compute_nutrition_cached(food_key, grams):
+    base = NUTRITION_DB[food_key]
+    factor = grams / 100
+
+    return {
+        "calories": round(base["calories"] * factor, 2),
+        "protein": round(base["protein"] * factor, 2),
+        "carbs": round(base["carbs"] * factor, 2),
+        "fat": round(base["fat"] * factor, 2),
+    }
+
+# =====================
+# GET NUTRITION
+# =====================
+def get_nutrition(dish, grams):
+    dish = normalize_food_name(dish)
+    food_key = find_food(dish)
+
+    if not food_key:
+        return estimate_unknown_food(grams)
+
+    return compute_nutrition_cached(food_key, grams)
+
+# =====================
+# SIMPLE QUANTITY ESTIMATION
+# =====================
+def estimate_quantity(pred):
+    width = pred.get("width", 0)
+    height = pred.get("height", 0)
+
+    area = width * height
+    ratio = area / (640 * 640)
+
+    grams = 150 + (ratio * 300)
+    return round(grams, 1)
+
+# =====================
+# DETECTION
+# =====================
+def detect(image_path):
+    try:
+        result = rf_client.run_workflow(
+            workspace_name="rishab-5ghrt",
+            workflow_id="detect-count-and-visualize-4",
+            images={"image": image_path}
+        )
+
+        r = result[0]
+
+        if "predictions" in r:
+            preds = r["predictions"]
+
+            if isinstance(preds, dict):
+                return preds.get("predictions", [])
+
+            return preds
+
+        return []
+
+    except Exception as e:
+        return f"❌ Roboflow Error: {e}"
+
+# =====================
+# MAIN FUNCTION (GRADIO)
+# =====================
+def analyze_image(image):
+    if image is None:
+        return "Please upload an image"
+
+    path = "temp.jpg"
+    shutil.copy(image, path)
+
+    preds = detect(path)
+
+    if isinstance(preds, str):
+        return preds
+
+    if len(preds) == 0:
+        os.remove(path)
+        return "❌ No food detected"
+
+    output = ""
+    total = {"calories": 0, "protein": 0, "carbs": 0, "fat": 0}
+
+    for pred in preds:
+        dish = pred.get("class", "unknown")
+
+        grams = estimate_quantity(pred)
+        nutrition = get_nutrition(dish, grams)
+
+        output += f"🍽️ {dish}\n"
+        output += f"📏 {grams} g\n"
+        output += f"🔥 {nutrition['calories']} kcal\n"
+        output += f"💪 Protein: {nutrition['protein']} g\n"
+        output += f"🍞 Carbs: {nutrition['carbs']} g\n"
+        output += f"🧈 Fat: {nutrition['fat']} g\n"
+        output += "-"*30 + "\n"
+
+        for k in total:
+            total[k] += nutrition[k]
+
+    os.remove(path)
+
+    output += "\n🧾 TOTAL:\n"
+    output += f"🔥 Calories: {round(total['calories'],2)}\n"
+    output += f"💪 Protein: {round(total['protein'],2)} g\n"
+    output += f"🍞 Carbs: {round(total['carbs'],2)} g\n"
+    output += f"🧈 Fat: {round(total['fat'],2)} g\n"
+
+    return output
+
+# =====================
+# GRADIO UI
+# =====================
+demo = gr.Interface(
+    fn=analyze_image,
+    inputs=gr.Image(type="filepath"),
+    outputs="text",
+    title="🍽️ AI Nutritionist",
+    description="Upload a food image to get calories & macros"
+)
+
+# =====================
+# RUN
+# =====================
+if __name__ == "__main__":
+    demo.launch(server_name="0.0.0.0", server_port=7860)