Update app.py

app.py

@@ -14,34 +14,25 @@ from google.genai import types
 # Aapki Gemini API Key yahan daal di gayi hai.
 GEMINI_API_KEY = "AIzaSyCrhWiAEQmCidtE2QZw3CTiLt7F8yv5M7A"
 
-## <<< CHANGE START: API Client ko globally initialize karein taaki baar-baar na bane.
-# API key configure karein
-genai.configure(api_key=GEMINI_API_KEY)
-# File service ke liye client banayein
-file_client = genai.FilesClient(api_key=GEMINI_API_KEY)
-## <<< CHANGE END
-
 def save_binary_file(file_name, data):
     with open(file_name, "wb") as f:
         f.write(data)
 
-def generate(text, file_name, model="gemini-1.5-flash-latest"): # <<< CHANGE: Model ko stable flash version se update kiya gaya
+def generate(text, file_name, model="gemini-2.0-flash-exp"):
+    # Client ko hardcoded API key se initialize karein.
+    client = genai.Client(api_key=GEMINI_API_KEY)
     
-    # Client ko global scope se istemal karein
-    print(f"Uploading file: {file_name}")
-    uploaded_file = file_client.upload_file(path=file_name)
-    print(f"Completed uploading file: {uploaded_file.name}")
-
-    # Model ko initialize karein
-    model_instance = genai.GenerativeModel(model_name=model)
+    files = [
+        client.files.upload(file=file_name),
+    ]
 
     contents = [
         types.Content(
             role="user",
             parts=[
                 types.Part.from_uri(
-                    file_uri=uploaded_file.uri,
-                    mime_type=uploaded_file.mime_type,
+                    file_uri=files[0].uri,
+                    mime_type=files[0].mime_type,
                 ),
                 types.Part.from_text(text=text),
             ],
@@ -61,90 +52,49 @@ def generate(text, file_name, model="gemini-1.5-flash-latest"): # <<< CHANGE: Mo
 
     with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as tmp:
         temp_path = tmp.name
-        print("Generating content...")
-        # Stream ki jagah seedha generate karein, image generation ke liye yeh aam taur par behtar hai
-        response = model_instance.generate_content(
+        for chunk in client.models.generate_content_stream(
+            model=model,
             contents=contents,
-            generation_config=generate_content_config,
-        )
-        print("Content generated.")
-        
-        # Response se image data extract karein
-        if response.candidates and response.candidates[0].content and response.candidates[0].content.parts:
-            inline_data = response.candidates[0].content.parts[0].inline_data
+            config=generate_content_config,
+        ):
+            if not chunk.candidates or not chunk.candidates[0].content or not chunk.candidates[0].content.parts:
+                continue
+            inline_data = chunk.candidates[0].content.parts[0].inline_data
             if inline_data:
                 save_binary_file(temp_path, inline_data.data)
                 print(
-                    f"File of mime type {inline_data.mime_type} saved to: {temp_path}"
+                    "File of mime type "
+                    f"{inline_data.mime_type} saved to: {temp_path} and prompt input :{text}"
                 )
             else:
-                 # Agar image nahi hai, to text response print karein
-                if response.text:
-                    print(f"Model Response (Text): {response.text}")
-                else:
-                    print("No image data or text found in the response.")
-        else:
-            print("Could not get a valid response from the model.")
+                print(chunk.text)
 
-
-    # Upload ki gayi file ko delete karein taaki storage na bhare
-    print(f"Deleting uploaded file: {uploaded_file.name}")
-    file_client.delete_file(name=uploaded_file.name)
-    
+    del files
     return temp_path
  
 
 def process_image_and_prompt(composite_pil, prompt):
-    ## <<< CHANGE START: Image ko API par bhejne se pehle resize karein
-    MAX_SIZE = (1024, 1024)  # Maximum dimensions
-    composite_pil.thumbnail(MAX_SIZE, Image.Resampling.LANCZOS)
-    print(f"Image resized to fit within {MAX_SIZE} dimensions.")
-    ## <<< CHANGE END
-
-    # Resized image ko temporary file mein save karein
+    # Save the composite image to a temporary file.
     with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as tmp:
         composite_path = tmp.name
-        # Agar image mein transparency (alpha channel) hai to use RGB mein convert karein
-        if composite_pil.mode == "RGBA":
-            composite_pil = composite_pil.convert("RGB")
-        composite_pil.save(composite_path, "PNG")
+        composite_pil.save(composite_path)
 
     file_name = composite_path  
     input_text = prompt 
-    
-    gemma_edited_image_path = generate(text=input_text, file_name=file_name)
-    
-    if os.path.exists(gemma_edited_image_path) and os.path.getsize(gemma_edited_image_path) > 0:
-        print("image_path ", gemma_edited_image_path)
-        result_img = Image.open(gemma_edited_image_path)
-        # Safai ke liye temporary files delete karein
-        os.remove(composite_path)
-        os.remove(gemma_edited_image_path)
-        return [result_img]
-    else:
-        # Agar koi image generate nahi hui to error handle karein
-        print("Failed to generate image or the generated file is empty.")
-        os.remove(composite_path) # Input temp file ko phir bhi delete karein
-        # Aap yahan user ko ek error message dikha sakte hain
-        # For now, we return nothing, which will clear the gallery
-        return []
+    model = "gemini-2.0-flash-exp" 
 
+    gemma_edited_image_path = generate(text=input_text, file_name=file_name, model=model)
+    print("image_path ", gemma_edited_image_path)
+    result_img = Image.open(gemma_edited_image_path)
+    if result_img.mode == "RGBA":
+        result_img = result_img.convert("RGB")
+    return [result_img]
 
 # Build a Blocks-based interface.
 with gr.Blocks() as demo:
-    gr.Markdown("## Gen AI Image Editing\nUpload an image and enter a prompt to generate outputs in the gallery. Do not Use NFSW Images")
-    
-    # Define examples to be shown within the Gradio interface
-    examples = [
-        ["data/1.webp", 'change text to "AMEER"'],
-        ["data/2.webp", "remove the spoon from  hand only"],
-        ["data/3.webp", 'change text to "Make it "'],
-        ["data/1.jpg", "add  joker style only on face"],
-        ["data/1777043.jpg", "add  joker style only on face"],
-        ["data/2807615.jpg","add lipstick on lip only "],
-        ["data/76860.jpg", "add lipstick on lip only "],
-        ["data/2807615.jpg", "make it happy looking face only"],
-    ]
+    # HTML Header ko yahan se hata diya gaya hai.
+    gr.Markdown("## Gen AI Image Editing")
+    gr.Markdown("Upload an image and enter a prompt to generate outputs in the gallery. Do not Use NFSW Images")
 
     with gr.Row():
         with gr.Column():
@@ -153,6 +103,7 @@ with gr.Blocks() as demo:
                 label="Upload Image",
                 image_mode="RGBA"
             )
+            # API Key Textbox ko yahan se hata diya gaya hai
             prompt_input = gr.Textbox(
                 lines=2,
                 placeholder="Enter prompt here...",
@@ -162,16 +113,33 @@ with gr.Blocks() as demo:
         with gr.Column():
             output_gallery = gr.Gallery(label="Generated Outputs")
 
+    # Define examples to be shown within the Gradio interface
+    examples = [
+        # Each example is a list corresponding to the inputs:
+        # [Input Image, Prompt]
+        ["data/1.webp", 'change text to "AMEER"'],
+        ["data/2.webp", "remove the spoon from  hand only"],
+        ["data/3.webp", 'change text to "Make it "'],
+        ["data/1.jpg", "add  joker style only on face"],
+         ["data/1777043.jpg", "add  joker style only on face"],
+         ["data/2807615.jpg","add lipstick on lip only "],
+         
+         ["data/76860.jpg", "add lipstick on lip only "],
+         ["data/2807615.jpg", "make it happy looking face only"],
+        
+    
+    ]
+
     # Set up the interaction.
     submit_btn.click(
         fn=process_image_and_prompt,
-        inputs=[image_input, prompt_input],
+        inputs=[image_input, prompt_input], # Inputs se API key hata di gayi hai
         outputs=output_gallery,
-        show_progress="full"  # <<< CHANGE: UI mein loading progress dikhayein
+ 
     )
     gr.Examples(
         examples=examples,
-        inputs=[image_input, prompt_input],
+        inputs=[image_input, prompt_input], # Inputs se API key hata di gayi hai
         label="Try these examples"
     )