Update app.py

app.py

@@ -1,4 +1,294 @@
-# import gradio as gr
+import gradio as gr
+from phi.agent import Agent
+from phi.model.groq import Groq
+import os
+import logging
+from sentence_transformers import CrossEncoder
+from backend.semantic_search import table, retriever
+import numpy as np
+from time import perf_counter
+import requests
+from jinja2 import Environment, FileSystemLoader
+from pathlib import Path
+
+# Set up logging
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# API Key setup
+api_key = os.getenv("GROQ_API_KEY")
+if not api_key:
+    gr.Warning("GROQ_API_KEY not found. Set it in 'Repository secrets'.")
+    logger.error("GROQ_API_KEY not found.")
+    api_key = ""  # Fallback to empty string, but this will fail without a key
+else:
+    os.environ["GROQ_API_KEY"] = api_key
+
+# Bhashini API setup
+bhashini_api_key = os.getenv("API_KEY", "").strip()
+bhashini_user_id = os.getenv("USER_ID", "").strip()
+
+def bhashini_translate(text: str, from_code: str = "en", to_code: str = "hi") -> dict:
+    """Translates text from source language to target language using the Bhashini API."""
+    if not text.strip():
+        print('Input text is empty. Please provide valid text for translation.')
+        return {"status_code": 400, "message": "Input text is empty", "translated_content": None}
+    else:
+        print('Input text - ', text)
+    print(f'Starting translation process from {from_code} to {to_code}...')
+    gr.Warning(f'Translating to {to_code}...')
+    
+    url = 'https://meity-auth.ulcacontrib.org/ulca/apis/v0/model/getModelsPipeline'
+    headers = {
+        "Content-Type": "application/json",
+        "userID": bhashini_user_id,
+        "ulcaApiKey": bhashini_api_key
+    }
+    for key, value in headers.items():
+        if not isinstance(value, str) or '\n' in value or '\r' in value:
+            print(f"Invalid header value for {key}: {value}")
+            return {"status_code": 400, "message": f"Invalid header value for {key}", "translated_content": None}
+
+    payload = {
+        "pipelineTasks": [{"taskType": "translation", "config": {"language": {"sourceLanguage": from_code, "targetLanguage": to_code}}}],
+        "pipelineRequestConfig": {"pipelineId": "64392f96daac500b55c543cd"}
+    }
+    
+    print('Sending initial request to get the pipeline...')
+    response = requests.post(url, json=payload, headers=headers)
+    
+    if response.status_code != 200:
+        print(f'Error in initial request: {response.status_code}, Response: {response.text}')
+        return {"status_code": response.status_code, "message": "Error in translation request", "translated_content": None}
+
+    print('Initial request successful, processing response...')
+    response_data = response.json()
+    print('Full response data:', response_data)
+    if "pipelineInferenceAPIEndPoint" not in response_data or "callbackUrl" not in response_data["pipelineInferenceAPIEndPoint"]:
+        print('Unexpected response structure:', response_data)
+        return {"status_code": 400, "message": "Unexpected API response structure", "translated_content": None}
+
+    service_id = response_data["pipelineResponseConfig"][0]["config"][0]["serviceId"]
+    callback_url = response_data["pipelineInferenceAPIEndPoint"]["callbackUrl"]
+    
+    print(f'Service ID: {service_id}, Callback URL: {callback_url}')
+    
+    headers2 = {
+        "Content-Type": "application/json",
+        response_data["pipelineInferenceAPIEndPoint"]["inferenceApiKey"]["name"]: response_data["pipelineInferenceAPIEndPoint"]["inferenceApiKey"]["value"]
+    }
+    compute_payload = {
+        "pipelineTasks": [{"taskType": "translation", "config": {"language": {"sourceLanguage": from_code, "targetLanguage": to_code}, "serviceId": service_id}}],
+        "inputData": {"input": [{"source": text}], "audio": [{"audioContent": None}]}
+    }
+    
+    print(f'Sending translation request with text: "{text}"')
+    compute_response = requests.post(callback_url, json=compute_payload, headers=headers2)
+    
+    if compute_response.status_code != 200:
+        print(f'Error in translation request: {compute_response.status_code}, Response: {compute_response.text}')
+        return {"status_code": compute_response.status_code, "message": "Error in translation", "translated_content": None}
+    
+    print('Translation request successful, processing translation...')
+    compute_response_data = compute_response.json()
+    translated_content = compute_response_data["pipelineResponse"][0]["output"][0]["target"]
+    
+    print(f'Translation successful. Translated content: "{translated_content}"')
+    return {"status_code": 200, "message": "Translation successful", "translated_content": translated_content}
+
+# Initialize PhiData Agent
+agent = Agent(
+    name="Science Education Assistant",
+    role="You are a helpful science tutor for 10th-grade students",
+    instructions=[
+        "You are an expert science teacher specializing in 10th-grade curriculum.",
+        "Provide clear, accurate, and age-appropriate explanations.",
+        "Use simple language and examples that students can understand.",
+        "Focus on concepts from physics, chemistry, and biology.",
+        "Structure responses with headings and bullet points when helpful.",
+        "Encourage learning and curiosity."
+    ],
+    model=Groq(id="llama3-70b-8192", api_key=api_key),
+    markdown=True
+)
+
+# Set up Jinja2 environment
+proj_dir = Path(__file__).parent
+env = Environment(loader=FileSystemLoader(proj_dir / 'templates'))
+template = env.get_template('template.j2')  # For document context
+template_html = env.get_template('template_html.j2')  # For HTML output
+
+# Response Generation Function
+def retrieve_and_generate_response(query, cross_encoder_choice, history=None):
+    """Generate response using semantic search and LLM"""
+    top_rerank = 25
+    top_k_rank = 20
+    
+    if not query.strip():
+        return "Please provide a valid question.", []
+    
+    try:
+        start_time = perf_counter()
+        
+        # Encode query and search documents
+        query_vec = retriever.encode(query)
+        documents = table.search(query_vec, vector_column_name="vector").limit(top_rerank).to_list()
+        documents = [doc["text"] for doc in documents]
+        
+        # Re-rank documents using cross-encoder
+        cross_encoder_model = CrossEncoder('BAAI/bge-reranker-base') if cross_encoder_choice == '(ACCURATE) BGE reranker' else CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')
+        query_doc_pair = [[query, doc] for doc in documents]
+        cross_scores = cross_encoder_model.predict(query_doc_pair)
+        sim_scores_argsort = list(reversed(np.argsort(cross_scores)))
+        documents = [documents[idx] for idx in sim_scores_argsort[:top_k_rank]]
+        
+        # Create context from top documents
+        context = "\n\n".join(documents[:10]) if documents else ""
+        context = f"Context information from educational materials:\n{context}\n\n"
+        
+        # Add conversation history for context
+        history_context = ""
+        if history and len(history) > 0:
+            for user_msg, bot_msg in history[-2:]:  # Last 2 exchanges
+                if user_msg and bot_msg:
+                    history_context += f"Previous Q: {user_msg}\nPrevious A: {bot_msg}\n"
+        
+        # Create full prompt
+        full_prompt = f"{history_context}{context}Question: {query}\n\nPlease answer the question using the context provided above. If the context doesn't contain relevant information, use your general knowledge about 10th-grade science topics."
+        
+        # Generate response
+        response = agent.run(full_prompt)
+        response_text = response.content if hasattr(response, 'content') else str(response)
+        
+        logger.info(f"Response generation took {perf_counter() - start_time:.2f} seconds")
+        return response_text, documents  # Return documents for template
+    
+    except Exception as e:
+        logger.error(f"Error in response generation: {e}")
+        return f"Error generating response: {str(e)}", []
+
+def simple_chat_function(message, history, cross_encoder_choice):
+    """Chat function with semantic search and retriever integration"""
+    if not message.strip():
+        return "", history, ""
+    
+    # Generate response and get documents
+    response, documents = retrieve_and_generate_response(message, cross_encoder_choice, history)
+    
+    # Add to history
+    history.append([message, response])
+    
+    # Render template with documents and query
+    prompt_html = template_html.render(documents=documents, query=message)
+    
+    return "", history, prompt_html
+
+def translate_text(selected_language, history):
+    """Translate the last response in history to the selected language."""
+    iso_language_codes = {
+        "Hindi": "hi", "Gom": "gom", "Kannada": "kn", "Dogri": "doi", "Bodo": "brx", "Urdu": "ur",
+        "Tamil": "ta", "Kashmiri": "ks", "Assamese": "as", "Bengali": "bn", "Marathi": "mr",
+        "Sindhi": "sd", "Maithili": "mai", "Punjabi": "pa", "Malayalam": "ml", "Manipuri": "mni",
+        "Telugu": "te", "Sanskrit": "sa", "Nepali": "ne", "Santali": "sat", "Gujarati": "gu", "Odia": "or"
+    }
+    
+    to_code = iso_language_codes[selected_language]
+    response_text = history[-1][1] if history and history[-1][1] else ''
+    print('response_text for translation', response_text)
+    translation = bhashini_translate(response_text, to_code=to_code)
+    return translation.get('translated_content', 'Translation failed.')
+
+# Gradio Interface with layout template
+with gr.Blocks(title="Science Chatbot", theme='gradio/soft') as demo:
+    # Header section
+    with gr.Row():
+        with gr.Column(scale=10):
+            gr.HTML(value="""<div style="color: #FF4500;"><h1>Welcome! I am your friend!</h1>Ask me !I will help you<h1><span style="color: #008000">I AM A CHATBOT FOR 10TH SCIENCE WITH TRANSLATION IN 22 LANGUAGES</span></h1></div>""")
+            gr.HTML(value=f"""<p style="font-family: sans-serif; font-size: 16px;">A free chat bot developed by K.M.RAMYASRI,TGT,GHS.SUTHUKENY using Open source LLMs for 10 std students</p>""")
+            gr.HTML(value=f"""<p style="font-family: Arial, sans-serif; font-size: 14px;"> Suggestions may be sent to <a href="mailto:[email protected]" style="color: #00008B; font-style: italic;">[email protected]</a>.</p>""")
+        with gr.Column(scale=3):
+            try:
+                gr.Image(value='logo.png', height=200, width=200)
+            except:
+                gr.HTML("<div style='height: 200px; width: 200px; background-color: #f0f0f0; display: flex; align-items: center; justify-content: center;'>Logo</div>")
+
+    # Chat and input components
+    chatbot = gr.Chatbot(
+        [],
+        elem_id="chatbot",
+        avatar_images=('https://aui.atlassian.com/aui/8.8/docs/images/avatar-person.svg',
+                       'https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg'),
+        bubble_full_width=False,
+        show_copy_button=True,
+        show_share_button=True,
+    )
+
+    with gr.Row():
+        msg = gr.Textbox(
+            scale=3,
+            show_label=False,
+            placeholder="Enter text and press enter",
+            container=False,
+        )
+        submit_btn = gr.Button(value="Submit text", scale=1, variant="primary")
+
+    # Additional controls
+    cross_encoder = gr.Radio(
+        choices=['(FAST) MiniLM-L6v2', '(ACCURATE) BGE reranker'],
+        value='(ACCURATE) BGE reranker',
+        label="Embeddings Model",
+        info="Select the model for document ranking"
+    )
+    language_dropdown = gr.Dropdown(
+        choices=[
+            "Hindi", "Gom", "Kannada", "Dogri", "Bodo", "Urdu", "Tamil", "Kashmiri", "Assamese", "Bengali", "Marathi",
+            "Sindhi", "Maithili", "Punjabi", "Malayalam", "Manipuri", "Telugu", "Sanskrit", "Nepali", "Santali",
+            "Gujarati", "Odia"
+        ],
+        value="Hindi",
+        label="Select Language for Translation"
+    )
+    translated_textbox = gr.Textbox(label="Translated Response")
+    prompt_html = gr.HTML()  # Add HTML component for the template
+
+    # Event handlers
+    def update_chat_and_translate(message, history, cross_encoder_choice, selected_language):
+        if not message.strip():
+            return "", history, "", ""
+        
+        # Generate response and get documents
+        response, documents = retrieve_and_generate_response(message, cross_encoder_choice, history)
+        history.append([message, response])
+        
+        # Translate response
+        translated_text = translate_text(selected_language, history)
+        
+        # Render template with documents and query
+        prompt_html_content = template_html.render(documents=documents, query=message)
+        
+        return "", history, translated_text, prompt_html_content
+
+    msg.submit(update_chat_and_translate, [msg, chatbot, cross_encoder, language_dropdown], [msg, chatbot, translated_textbox, prompt_html])
+    submit_btn.click(update_chat_and_translate, [msg, chatbot, cross_encoder, language_dropdown], [msg, chatbot, translated_textbox, prompt_html])
+
+    clear = gr.Button("Clear Conversation")
+    clear.click(lambda: ([], "", "", ""), outputs=[chatbot, msg, translated_textbox, prompt_html])
+
+    # Example questions
+    gr.Examples(
+        examples=[
+            'What is the difference between metals and non-metals?',
+            'What is an ionic bond?',
+            'Explain asexual reproduction',
+            'What is photosynthesis?',
+            'Explain Newton\'s laws of motion'
+        ],
+        inputs=msg,
+        label="Try these example questions:"
+    )
+
+if __name__ == "__main__":
+    demo.launch(server_name="0.0.0.0", server_port=7860)# import gradio as gr
 # from phi.agent import Agent
 # from phi.model.groq import Groq
 # import os