fixed

app.py

@@ -1,13 +1,9 @@
-import os
-import uuid
-import tempfile
 import gradio as gr
+import uuid
 from qdrant_client import QdrantClient
-from qdrant_client.models import VectorParams, Distance, PointStruct
+from qdrant_client.models import PointStruct, VectorParams, Distance
 from sentence_transformers import SentenceTransformer
 from PIL import Image
-import torch
-import numpy as np
 
 # --------------------------
 # Qdrant Cloud Connection
@@ -19,7 +15,9 @@ COLLECTION_NAME = "lost_and_found"
 # CLIP model (text + image embeddings)
 MODEL_NAME = "sentence-transformers/clip-ViT-B-32"
 embedder = SentenceTransformer(MODEL_NAME)
-VECTOR_SIZE = embedder.get_sentence_embedding_dimension()
+
+# CLIP ViT-B/32 always gives 512-dimensional embeddings
+VECTOR_SIZE = 512
 
 # Qdrant Client (Cloud)
 qclient = QdrantClient(
@@ -36,106 +34,119 @@ qclient.recreate_collection(
 # --------------------------
 # Helper Functions
 # --------------------------
+
 def embed_text(text: str):
-    """Generate embedding for text"""
     return embedder.encode(text).tolist()
 
-def embed_image(image: Image.Image):
-    """Generate embedding for image"""
-    img_tensor = embedder.encode(image, convert_to_tensor=True)
-    return img_tensor.cpu().detach().numpy().tolist()
+def embed_image(img: Image.Image):
+    return embedder.encode(img).tolist()
 
-# --------------------------
-# Core Functions
-# --------------------------
-def add_item(description, image):
-    """Add a found item to Qdrant"""
-    if not description and image is None:
-        return "⚠️ Please provide description or image."
-
-    vectors = []
-    payload = {"description": description}
-
-    if description:
-        vectors = embed_text(description)
-
-    if image:
-        vectors = embed_image(image)
-        # Save uploaded image
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as tmp:
-            image.save(tmp.name)
-            payload["image_path"] = tmp.name
-
-    point = PointStruct(
-        id=str(uuid.uuid4()),
-        vector=vectors,
-        payload=payload
-    )
-    qclient.upsert(collection_name=COLLECTION_NAME, points=[point])
+def add_item(image, description, finder_name, finder_phone):
+    """Add a found item to Qdrant database"""
+    if image is None or description.strip() == "":
+        return "❌ Please provide both an image and a description."
 
-    return "✅ Item added successfully!"
+    # Encode image
+    embedding = embed_image(image)
 
-def search_items(query_text, query_image, max_results, min_score):
-    """Search lost items by text or image"""
-    vectors = None
+    # Store metadata
+    metadata = {
+        "description": description,
+        "finder_name": finder_name if finder_name.strip() else "NA",
+        "finder_phone": finder_phone if finder_phone.strip() else "NA",
+    }
 
-    if query_text:
-        vectors = embed_text(query_text)
+    # Insert into Qdrant
+    qclient.upsert(
+        collection_name=COLLECTION_NAME,
+        points=[
+            PointStruct(
+                id=str(uuid.uuid4()),
+                vector=embedding,
+                payload=metadata
+            )
+        ]
+    )
+    return "✅ Item successfully added!"
 
-    elif query_image:
-        vectors = embed_image(query_image)
+def search_items(query_text, query_image):
+    """Search by text or image"""
+    if not query_text and query_image is None:
+        return "❌ Please enter text or upload an image to search.", []
 
+    # Use text or image embedding
+    if query_image:
+        query_vector = embed_image(query_image)
     else:
-        return ["⚠️ Provide text or image to search."]
+        query_vector = embed_text(query_text)
 
+    # Search Qdrant
     results = qclient.search(
         collection_name=COLLECTION_NAME,
-        query_vector=vectors,
-        limit=max_results,
-        score_threshold=min_score,
+        query_vector=query_vector,
+        limit=5
     )
 
     if not results:
-        return ["No matches found."]
+        return "❌ No matches found.", []
 
-    outputs = []
+    # Format results
+    gallery = []
+    output_text = "✅ Found Matches:\n\n"
     for r in results:
         desc = r.payload.get("description", "No description")
-        img = r.payload.get("image_path", None)
-        score = round(r.score, 3)
-        if img:
-            outputs.append((img, f"{desc} (score: {score})"))
-        else:
-            outputs.append((None, f"{desc} (score: {score})"))
+        name = r.payload.get("finder_name", "NA")
+        phone = r.payload.get("finder_phone", "NA")
+        output_text += f"- **{desc}** (Finder: {name}, Phone: {phone})\n"
+        gallery.append(Image.new("RGB", (200, 200), color="gray"))  # Placeholder (no storage for images)
 
-    return outputs
+    return output_text, gallery
+
+def clear_database():
+    """Clear all stored items in Qdrant"""
+    qclient.delete_collection(COLLECTION_NAME)
+    qclient.recreate_collection(
+        collection_name=COLLECTION_NAME,
+        vectors_config=VectorParams(size=VECTOR_SIZE, distance=Distance.COSINE),
+    )
+    return "🗑️ Database cleared!"
 
 # --------------------------
 # Gradio UI
 # --------------------------
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🔍 Lost & Found System")
+
     with gr.Tab("➕ Add Found Item"):
         with gr.Row():
+            image_in = gr.Image(type="pil", label="Upload Found Item Image")
             desc_in = gr.Textbox(label="Item Description")
-            img_in = gr.Image(type="pil", label="Upload Image")
+        with gr.Row():
+            finder_name = gr.Textbox(label="Finder's Name")
+            finder_phone = gr.Textbox(label="Finder's Phone Number")
         add_btn = gr.Button("Add Item")
-        add_out = gr.Textbox(label="Status")
-        add_btn.click(fn=add_item, inputs=[desc_in, img_in], outputs=add_out)
-
-    with gr.Tab("🔍 Search Lost Item"):
-        query_text = gr.Textbox(label="Search by Text (optional)")
-        query_img = gr.Image(type="pil", label="Search by Image (optional)")
-        max_results = gr.Slider(1, 20, step=1, value=5, label="Max Results")
-        min_score = gr.Slider(0.0, 1.0, step=0.01, value=0.3, label="Min Similarity Score")
+        add_output = gr.Textbox(label="Status")
+
+    with gr.Tab("🔎 Search Items"):
+        with gr.Row():
+            search_text = gr.Textbox(label="Search by Text")
+            search_image = gr.Image(type="pil", label="Or Search by Image")
         search_btn = gr.Button("Search")
-        results_out = gr.Gallery(label="Search Results").style(grid=2, height="auto")
-        search_btn.click(fn=search_items, inputs=[query_text, query_img, max_results, min_score], outputs=results_out)
+        search_output = gr.Markdown(label="Results")
+        gallery = gr.Gallery(label="Matched Items", show_label=True, elem_id="gallery")
+
+    with gr.Tab("⚠️ Admin"):
+        clear_btn = gr.Button("Clear Entire Database")
+        clear_output = gr.Textbox(label="Status")
 
-    with gr.Tab("⚙️ Admin"):
-        gr.Markdown("Admin dashboard (future expansion).")
+    # Button actions
+    add_btn.click(add_item, inputs=[image_in, desc_in, finder_name, finder_phone], outputs=add_output)
+    search_btn.click(search_items, inputs=[search_text, search_image], outputs=[search_output, gallery])
+    clear_btn.click(clear_database, outputs=clear_output)
 
 # --------------------------
-# Run App
+# Launch App
 # --------------------------
 if __name__ == "__main__":
     demo.launch(server_name="0.0.0.0", server_port=7860)