Fix CLIP2 model issue in app.py

app.py

@@ -2,6 +2,7 @@
 import os
 import uuid
 import io
+import base64
 from PIL import Image
 import gradio as gr
 import numpy as np
@@ -41,7 +42,6 @@ qclient = QdrantClient(url=QDRANT_URL, api_key=QDRANT_API_KEY)
 COLLECTION = "lost_found_items"
 VECTOR_SIZE = 512
 
-# Create collection if missing
 if not qclient.collection_exists(COLLECTION):
     qclient.create_collection(
         collection_name=COLLECTION,
@@ -52,24 +52,24 @@ if not qclient.collection_exists(COLLECTION):
 #  Helpers
 # -------------------------
 def embed_text(text: str):
-    vec = clip_model.encode([text], convert_to_numpy=True)[0]  # wrap in list
-    return vec
+    return clip_model.encode(text, convert_to_numpy=True)
 
 def embed_image_pil(pil_img: Image.Image):
-    vec = clip_model.encode([pil_img], convert_to_numpy=True)[0]  # wrap in list
-    return vec
+    return clip_model.encode(pil_img, convert_to_numpy=True)
 
-def gen_tags_from_image_file(local_path: str) -> str:
-    if not genai_client:
+def gen_tags_from_image_file(file_obj) -> str:
+    """file_obj can be path or BytesIO"""
+    if genai_client is None:
         return ""
-    file_obj = genai_client.files.upload(file=local_path)
+    uploaded_file = genai_client.files.upload(file=file_obj)
     prompt_text = (
         "Give 4 short tags (comma-separated) describing this item in the image. "
-        "Tags should be short single words or two-word phrases. Respond only with tags."
+        "Tags should be short single words or two-word phrases (e.g. 'black backpack', 'water bottle'). "
+        "Respond only with tags, no extra explanation."
     )
     response = genai_client.models.generate_content(
         model="gemini-2.5-flash",
-        contents=[prompt_text, file_obj],
+        contents=[prompt_text, uploaded_file],
     )
     return response.text.strip()
 
@@ -81,21 +81,29 @@ def add_item(mode: str, uploaded_image, text_description: str):
     payload = {"mode": mode, "text": text_description}
 
     if uploaded_image is not None:
-        tmp_path = f"/tmp/{item_id}.png"
-        uploaded_image.save(tmp_path)
+        # Save to BytesIO
+        img_bytes_io = io.BytesIO()
+        uploaded_image.save(img_bytes_io, format="PNG")
+        img_bytes_io.seek(0)
+
+        # Embed image
         vec = embed_image_pil(uploaded_image).tolist()
         payload["has_image"] = True
+
+        # Generate tags
         try:
-            tags = gen_tags_from_image_file(tmp_path)
+            tags = gen_tags_from_image_file(img_bytes_io)
         except Exception:
             tags = ""
         payload["tags"] = tags
-        with open(tmp_path, "rb") as f:
-            b64 = f.read()
-        payload["image_b64"] = True
+
+        # Store image as base64
+        img_bytes_io.seek(0)
+        payload["image_b64"] = base64.b64encode(img_bytes_io.read()).decode("utf-8")
     else:
         vec = embed_text(text_description).tolist()
         payload["has_image"] = False
+
         if genai_client:
             try:
                 resp = genai_client.models.generate_content(
@@ -108,6 +116,7 @@ def add_item(mode: str, uploaded_image, text_description: str):
         else:
             payload["tags"] = ""
 
+    # Upsert into Qdrant
     point = PointStruct(id=item_id, vector=vec, payload=payload)
     qclient.upsert(collection_name=COLLECTION, points=[point], wait=True)
 
@@ -119,53 +128,47 @@ def add_item(mode: str, uploaded_image, text_description: str):
 def search_items(query_image, query_text, limit: int = 5):
     if query_image is not None:
         qvec = embed_image_pil(query_image).tolist()
-    elif query_text and query_text.strip():
+    elif query_text:
         qvec = embed_text(query_text).tolist()
     else:
-        return "Please provide a query image or some query text."
+        return "Please provide a query image or text."
 
     hits = qclient.search(collection_name=COLLECTION, query_vector=qvec, limit=limit)
 
+    if not hits:
+        return "No results."
+
     results = []
     for h in hits:
         payload = h.payload or {}
         score = getattr(h, "score", None)
+        img_html = ""
+        if payload.get("has_image") and payload.get("image_b64"):
+            img_html = f'<img src="data:image/png;base64,{payload["image_b64"]}" width="200">'
         results.append(
-            {
-                "id": h.id,
-                "score": float(score) if score else None,
-                "mode": payload.get("mode", ""),
-                "text": payload.get("text", ""),
-                "tags": payload.get("tags", ""),
-                "has_image": payload.get("has_image", False),
-            }
+            f"{img_html}<br>ID:{h.id}<br>Score:{float(score) if score else 0:.4f}<br>"
+            f"Mode:{payload.get('mode','')}<br>Tags:{payload.get('tags','')}<br>Text:{payload.get('text','')}"
         )
 
-    if not results:
-        return "No results."
-    out_lines = [
-        f"id:{r['id']} score:{r['score']:.4f} mode:{r['mode']} tags:{r['tags']} text:{r['text']}"
-        for r in results
-    ]
-    return "\n\n".join(out_lines)
+    return "<br><br>".join(results)
 
 # -------------------------
 #  Gradio UI
 # -------------------------
 with gr.Blocks(title="Lost & Found — Simple Helper") as demo:
-    gr.Markdown("## Lost & Found Helper (image/text search)")
+    gr.Markdown("## Lost & Found Helper — Upload items and search by image or text.")
     with gr.Row():
         with gr.Column():
             mode = gr.Radio(choices=["lost", "found"], value="lost", label="Add as")
             upload_img = gr.Image(type="pil", label="Item photo (optional)")
             text_desc = gr.Textbox(lines=2, placeholder="Short description", label="Description (optional)")
             add_btn = gr.Button("Add item")
-            add_out = gr.Textbox(label="Add result", interactive=False)
+            add_out = gr.HTML(label="Add result")  # Changed to HTML to render images
         with gr.Column():
             query_img = gr.Image(type="pil", label="Search by image (optional)")
             query_text = gr.Textbox(lines=2, label="Search by text (optional)")
             search_btn = gr.Button("Search")
-            search_out = gr.Textbox(label="Search results", interactive=False)
+            search_out = gr.HTML(label="Search results")  # HTML to render images
 
     add_btn.click(add_item, inputs=[mode, upload_img, text_desc], outputs=[add_out])
     search_btn.click(search_items, inputs=[query_img, query_text], outputs=[search_out])