Update app.py

app.py

@@ -5,277 +5,85 @@ import uuid
 from datetime import datetime
 import shutil
 from huggingface_hub import HfApi, create_repo, upload_file, upload_folder
-
-# Create directories for data storage
 os.makedirs("uploaded_images", exist_ok=True)
 os.makedirs("submissions", exist_ok=True)
-
-# Hugging Face Dataset configuration
 HF_TOKEN = os.environ.get("HF_TOKEN") 
 DATASET_NAME = "minemaster01/se-culture-dataset-results" 
 DATASET_CREATED = False
-
-# States by country dictionary
-states_by_country = {
-    "India": [
-        "Andhra Pradesh", "Arunachal Pradesh", "Assam", "Bihar", "Chhattisgarh", "Goa", "Gujarat", "Haryana",
-        "Himachal Pradesh", "Jharkhand", "Karnataka", "Kerala", "Madhya Pradesh", "Maharashtra", "Manipur",
-        "Meghalaya", "Mizoram", "Nagaland", "Odisha", "Punjab", "Rajasthan", "Sikkim", "Tamil Nadu", "Telangana",
-        "Tripura", "Uttar Pradesh", "Uttarakhand", "West Bengal", "Andaman and Nicobar Islands", "Chandigarh",
-        "Dadra and Nagar Haveli and Daman and Diu", "Delhi", "Jammu and Kashmir", "Ladakh", "Lakshadweep", "Puducherry"
-    ],
-    "Pakistan": [
-        "Balochistan", "Khyber Pakhtunkhwa", "Punjab", "Sindh", "Islamabad Capital Territory", "Gilgit-Baltistan"
-    ],
-    "Bangladesh": [
-        "Barisal", "Chittagong", "Dhaka", "Khulna", "Mymensingh", "Rajshahi", "Rangpur", "Sylhet"
-    ],
-    "Afghanistan": [
-        "Badakhshan", "Badghis", "Baghlan", "Balkh", "Bamyan", "Daykundi", "Farah", "Faryab", "Ghazni", "Ghor",
-        "Helmand", "Herat", "Jowzjan", "Kabul", "Kandahar", "Kapisa", "Khost", "Kunar", "Kunduz", "Laghman",
-        "Logar", "Nangarhar", "Nimruz", "Nuristan", "Paktia", "Paktika", "Panjshir", "Parwan", "Samangan",
-        "Sar-e Pol", "Takhar", "Uruzgan", "Wardak", "Zabul"
-    ],
-    "Bhutan": [
-        "Bumthang", "Chukha", "Dagana", "Gasa", "Haa", "Lhuentse", "Mongar", "Paro", "Pemagatshel", "Punakha",
-        "Samdrup Jongkhar", "Samtse", "Sarpang", "Thimphu", "Trashigang", "Trashiyangtse", "Trongsa", "Tsirang",
-        "Wangdue Phodrang", "Zhemgang"
-    ],
-    "Nepal": [
-        "Bagmati", "Gandaki", "Karnali", "Koshi", "Lumbini", "Madhesh", "Sudurpashchim"
-    ],
-    "Sri Lanka": [
-        "Central", "Eastern", "North Central", "Northern", "North Western", "Sabaragamuwa", "Southern", "Uva", "Western"
-    ]
-}
-south_asian_languages = [
-    "Hindi", "Bengali", "Urdu", "Punjabi", "Tamil", "Telugu", 
-    "Marathi", "Gujarati", "Kannada", "Malayalam", "Odia", 
-    "Sindhi", "Nepali", "Sinhala", "Pashto", "Dari", 
-    "Dzongkha", "Assamese", "Kashmiri", "Sanskrit", "Other"
-]
+states_by_country = {"India": ["Andhra Pradesh", "Arunachal Pradesh", "Assam", "Bihar", "Chhattisgarh", "Goa", "Gujarat", "Haryana", "Himachal Pradesh", "Jharkhand", "Karnataka", "Kerala", "Madhya Pradesh", "Maharashtra", "Manipur", "Meghalaya", "Mizoram", "Nagaland", "Odisha", "Punjab", "Rajasthan", "Sikkim", "Tamil Nadu", "Telangana", "Tripura", "Uttar Pradesh", "Uttarakhand", "West Bengal", "Andaman and Nicobar Islands", "Chandigarh", "Dadra and Nagar Haveli and Daman and Diu", "Delhi", "Jammu and Kashmir", "Ladakh", "Lakshadweep", "Puducherry"], "Pakistan": ["Balochistan", "Khyber Pakhtunkhwa", "Punjab", "Sindh", "Islamabad Capital Territory", "Other"], "Bangladesh": ["Barisal", "Chittagong", "Dhaka", "Khulna", "Mymensingh", "Rajshahi", "Rangpur", "Sylhet"], "Afghanistan": ["Badakhshan", "Badghis", "Baghlan", "Balkh", "Bamyan", "Daykundi", "Farah", "Faryab", "Ghazni", "Ghor", "Helmand", "Herat", "Jowzjan", "Kabul", "Kandahar", "Kapisa", "Khost", "Kunar", "Kunduz", "Laghman", "Logar", "Nangarhar", "Nimruz", "Nuristan", "Paktia", "Paktika", "Panjshir", "Parwan", "Samangan", "Sar-e Pol", "Takhar", "Uruzgan", "Wardak", "Zabul"], "Bhutan": ["Bumthang", "Chukha", "Dagana", "Gasa", "Haa", "Lhuentse", "Mongar", "Paro", "Pemagatshel", "Punakha", "Samdrup Jongkhar", "Samtse", "Sarpang", "Thimphu", "Trashigang", "Trashiyangtse", "Trongsa", "Tsirang", "Wangdue Phodrang", "Zhemgang"], "Nepal": ["Bagmati", "Gandaki", "Karnali", "Koshi", "Lumbini", "Madhesh", "Sudurpashchim"], "Sri Lanka": ["Central", "Eastern", "North Central", "Northern", "North Western", "Sabaragamuwa", "Southern", "Uva", "Western"]}
+south_asian_languages = ["Assamese", "Bengali", "Bhojpuri", "Bodo", "Dari", "Dzongkha", "Dogri", "Gujarati", "Hindi", "Kannada", "Kashmiri", "Konkani", "Maithili", "Malayalam", "Marathi", "Meitei", "Nepali", "Odia", "Pashto", "Punjabi", "Sanskrit", "Santali", "Sindhi", "Sinhala", "Tamil", "Telugu", "Tulu", "Urdu", "OTHER"]
 def setup_hf_dataset():
-    """Initialize the Hugging Face dataset if it doesn't exist"""
     global DATASET_CREATED
     if not DATASET_CREATED and HF_TOKEN:
         try:
             api = HfApi()
-            create_repo(
-                DATASET_NAME,
-                repo_type="dataset",
-                token=HF_TOKEN,
-                exist_ok=True
-            )
+            create_repo(DATASET_NAME, repo_type="dataset", token=HF_TOKEN, exist_ok=True)
             DATASET_CREATED = True
             print(f"Dataset {DATASET_NAME} is ready")
-        except Exception as e:
-            print(f"Error setting up dataset: {e}")
+        except Exception as e: print(f"Error setting up dataset: {e}")
     elif not HF_TOKEN:
         print("Warning: HF_TOKEN not set. Data will be stored locally only.")
-
 def update_state_dropdown(country):
-    """Update state dropdown based on selected country"""
     if country in states_by_country:
         return gr.Dropdown(choices=states_by_country[country], label=f"State/Province in {country}:", interactive=True)
     return gr.Dropdown(choices=[], label="State/Province:", interactive=True)
-
 def process_submission(input_img, language, country, state, city, se_asia_relevance, culture_knowledge, native_caption, english_caption,email):
-    # Generate unique ID for this submission
     submission_id = str(uuid.uuid4())
     timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-    
-    # Save the image if provided
     image_path = None
     if input_img is not None:
-        # Create filename with submission ID
         image_filename = f"{timestamp}.jpg"
         image_path = os.path.join("uploaded_images", image_filename)
-        
-        # Save the image
-        if isinstance(input_img, str):  # If it's a file path
-            shutil.copy(input_img, image_path)
-        else:  # If it's a PIL Image
-            input_img.save(image_path)
-    
-    # Create a data structure for the submission
-    submission_data = {
-        "id": submission_id,
-        "timestamp": timestamp,
-        "image_filename": os.path.basename(image_path) if image_path else None,
-        "cultural_relevance": language,
-        "country": country,
-        "state": state,
-        "city": city,
-        "se_asia_relevance": se_asia_relevance,
-        "cultural_knowledge_source": culture_knowledge,
-        "native_caption": native_caption,
-        "english_caption": english_caption,
-        "email": email
-    }
-    
-    # Save the data as JSON
+        if isinstance(input_img, str): shutil.copy(input_img, image_path)
+        else: input_img.save(image_path)
+    submission_data = {"id": submission_id, "timestamp": timestamp, "image_filename": os.path.basename(image_path) if image_path else None, "cultural_relevance": language, "country": country, "state": state, "city": city, "se_asia_relevance": se_asia_relevance, "cultural_knowledge_source": culture_knowledge, "native_caption": native_caption, "english_caption": english_caption, "email": email}
     json_filename = f"{timestamp}.json"
     json_path = os.path.join("submissions", json_filename)
-    with open(json_path, "w") as f:
-        json.dump(submission_data, f, indent=2)
-    
-    # Upload to Hugging Face Dataset if token is available
+    with open(json_path, "w") as f: json.dump(submission_data, f, indent=2)
     if HF_TOKEN and DATASET_CREATED:
         try:
             api = HfApi()
-            # Upload the JSON data
-            api.upload_file(
-                path_or_fileobj=json_path,
-                path_in_repo=f"submissions/{json_filename}",
-                repo_id=DATASET_NAME,
-                repo_type="dataset",
-                token=HF_TOKEN
-            )
-            # Upload the image if it exists
+            api.upload_file(path_or_fileobj=json_path, path_in_repo=f"submissions/{json_filename}", repo_id=DATASET_NAME, repo_type="dataset", token=HF_TOKEN)
             if image_path and os.path.exists(image_path):
-                api.upload_file(
-                    path_or_fileobj=image_path,
-                    path_in_repo=f"images/{os.path.basename(image_path)}",
-                    repo_id=DATASET_NAME,
-                    repo_type="dataset",
-                    token=HF_TOKEN
-                )
+                api.upload_file(path_or_fileobj=image_path, path_in_repo=f"images/{os.path.basename(image_path)}", repo_id=DATASET_NAME, repo_type="dataset", token=HF_TOKEN)
             print(f"Submission {submission_id} uploaded to Hugging Face Dataset")
-        except Exception as e:
-            print(f"Error uploading to dataset: {e}")
-    
-    # Return values to display in the interface
+        except Exception as e: print(f"Error uploading to dataset: {e}")
     location_info = f"Location: {city}, {state}, {country}" if state else f"Location: {city}, {country}"
     return input_img, f"Your text response: {language}", f"Selected location: {location_info}", f"SE Asia relevance: {se_asia_relevance}", f"Cultural knowledge source: {culture_knowledge}", f"Native caption: {native_caption}", f"English caption: {english_caption}"
-
 def clear_inputs():
     return None, "", None, None, "", None, None, "", "", ""
-
-# Initialize the dataset
 setup_hf_dataset()
-
-with gr.Blocks() as gradio_app:
+with gr.Blocks(theme='1024m/1024m-1') as gradio_app:
     gr.Markdown("# South Asian Image Data Collection")
     gr.Markdown("Upload an image and answer questions about its cultural significance.")
-    
-    # Split the interface into two columns
     with gr.Row():
-        # Left column for image upload and basic information
         with gr.Column(scale=1):
             input_img = gr.Image(label="Upload an image", sources=['upload', 'webcam'], type="pil")
-            language = gr.Dropdown(
-                choices=south_asian_languages,
-                label="Language:",
-                info="Select the native language relevant to the image",
-                interactive=True
-            )
-            
-            # Location information in the left column
-            country_dropdown = gr.Dropdown(
-                choices=["None","India", "Pakistan", "Bangladesh", "Afghanistan", "Bhutan", "Nepal", "Sri Lanka"],
-                label="Country where the image was taken:",
-                interactive=True
-            )
-            state_dropdown = gr.Dropdown(
-                choices=[],
-                label="State/Province:",
-                interactive=True
-            )
+            language = gr.Dropdown(choices=south_asian_languages, label="Language:", info="Select the native language relevant to the image", interactive=True)
+            country_dropdown = gr.Dropdown(choices=["None","India", "Pakistan", "Bangladesh", "Afghanistan", "Bhutan", "Nepal", "Sri Lanka"], label="Country where the image was taken:", interactive=True)
+            state_dropdown = gr.Dropdown(choices=[], label="State/Province:", interactive=True)
             city_textbox = gr.Textbox(label="City where the image was taken:", placeholder="Enter city name")
             email_input = gr.Textbox(label="Your Email:", placeholder="Enter your email address", info="Used as unique contributor ID")
-        
-        # Right column for additional information
         with gr.Column(scale=1):
-            se_asia_relevance = gr.Radio(
-                choices=[
-                    "Yes. Unique to South Asia",
-                    "Yes, people will likely think of South Asia when seeing the picture, but it may have low degree of similarity to other cultures.",
-                    "Maybe, this culture did not originate from South Asia, but it's quite dominant in South Asia",
-                    "Not really. It has some affiliation to South Asia, but actually does not represent South Asia or has stronger affiliation to cultures outside South Asia",
-                    "No. Totally unrelated to South Asia"
-                ],
-                label="Is the image culturally relevant in South Asia?"
-            )
-            
-            culture_knowledge = gr.Radio(
-                choices=[
-                    "I'm from this country/culture",
-                    "I checked online resources (e.g., Wikipedia, articles, blogs)"
-                ],
-                label="How do you know about this culture?",
-                info="Please do not consult LLMs (e.g., GPT-4o, Claude, Command-R, etc.)"
-            )
-            
+            se_asia_relevance = gr.Radio(choices=["Yes. Unique to South Asia", "Yes, people will likely think of South Asia when seeing the picture, but it may have low degree of similarity to other cultures.", "Maybe, this culture did not originate from South Asia, but it's quite dominant in South Asia", "Not really. It has some affiliation to South Asia, but actually does not represent South Asia or has stronger affiliation to cultures outside South Asia", "No. Totally unrelated to South Asia"], label="Is the image culturally relevant in South Asia?")
+            culture_knowledge = gr.Radio(choices=["I'm from this country/culture", "I checked online resources (e.g., Wikipedia, articles, blogs)"], label="How do you know about this culture?", info="Please do not consult LLMs (e.g., GPT-4o, Claude, Command-R, etc.)")
             native_caption = gr.Textbox(label="Caption in Native Language:", placeholder="Enter caption in the native language of the culture depicted")
             english_caption = gr.Textbox(label="English Caption:", placeholder="Enter caption in English")
-    
-    # Buttons row
     with gr.Row():
         clear_btn = gr.Button("Clear")
         submit_btn = gr.Button("Submit")
-    
-    # Output display section - also split into two columns
     with gr.Row():
         with gr.Column(scale=1):
             output_img = gr.Image(label="Submitted Image")
             output_text = gr.Text(label="Text Response")
             output_location = gr.Text(label="Location Information")
-        
         with gr.Column(scale=1):
             output_relevance = gr.Text(label="South Asia Cultural Relevance")
             output_knowledge = gr.Text(label="Cultural Knowledge Source")
             output_native = gr.Text(label="Native Language Caption")
             output_english = gr.Text(label="English Caption")
-    
-    # Set up event handlers
-    country_dropdown.change(
-        fn=update_state_dropdown,
-        inputs=country_dropdown,
-        outputs=state_dropdown
-    )
-    
-    submit_btn.click(
-        fn=process_submission,
-        inputs=[
-            input_img,
-            language,
-            country_dropdown,
-            state_dropdown,
-            city_textbox,
-            se_asia_relevance,
-            culture_knowledge,
-            native_caption,
-            english_caption,
-            email_input 
-        ],
-        outputs=[
-            output_img,
-            output_text,
-            output_location,
-            output_relevance,
-            output_knowledge,
-            output_native,
-            output_english
-        ]
-    )
-    
-    clear_btn.click(
-        fn=clear_inputs,
-        inputs=[],
-        outputs=[
-            input_img,
-            language,
-            country_dropdown,
-            state_dropdown,
-            city_textbox,
-            se_asia_relevance,
-            culture_knowledge,
-            native_caption,
-            english_caption,
-            email_input 
-        ]
-    )
-
-
+    country_dropdown.change(fn=update_state_dropdown, inputs=country_dropdown, outputs=state_dropdown)
+    submit_btn.click(fn=process_submission, inputs=[input_img, language, country_dropdown, state_dropdown, city_textbox, se_asia_relevance, culture_knowledge, native_caption, english_caption, email_input], outputs=[output_img, output_text, output_location, output_relevance, output_knowledge, output_native, output_english])
+    clear_btn.click(fn=clear_inputs, inputs=[], outputs=[input_img, language, country_dropdown, state_dropdown, city_textbox, se_asia_relevance, culture_knowledge, native_caption, english_caption, email_input])
 if __name__ == "__main__":
     gradio_app.launch()