backedn chat apis and uplaod apis integrated

app.py

@@ -78,7 +78,7 @@ async def get_api_key(api_key_header: str = Security(api_key_header)) -> str:
 class ChatRequest(BaseModel):
     resume_url: Optional[str] = None
     query: str=None
-    conversation_id: str
+    chat_id: int
     theme_id: Optional[int] = 1
     full_history_url: Optional[str] = None
     form_id:Optional[int] = None
@@ -95,7 +95,7 @@ class ChatResponse(BaseModel):
 
 class GeneratePDFRequest(BaseModel):
     resume_url: Optional[str] = None
-    conversation_id: str
+    chat_id: int
     theme_id: Optional[int] = 1
     full_history_url: Optional[str] = None
     form_id:Optional[int] = None
@@ -215,7 +215,7 @@ async def chat_endpoint(
             query = "Let's get started"
         response = ai_chat(
                 query=query,
-                conversation_id=request.conversation_id,
+                conversation_id=request.chat_id,
                 theme_id=request.theme_id,
                 resume=resume_docs,
                 full_history=full_history_docs,
@@ -304,16 +304,16 @@ async def generate_pdf_endpoint(
         # Here you would fetch the conversation data using the conversation_id
         # This is a placeholder - replace with your actual conversation data fetching logic
          # Get AI-generated theme content
+        # Get AI-generated theme content
         response = ai_chat(
             query="NOW GENERATE THE STARTPOP FRAMEWORK",
-            conversation_id=request.conversation_id,
+            conversation_id=request.chat_id,
             theme_id=request.theme_id,
             resume=resume_docs,
             full_history=full_history_docs,
             form_response=form_response_docs,
             generate_theme="YES"
         )
-
         print(f"AI Response for theme: {response}")
         
         # Ensure AI response is valid
@@ -321,19 +321,40 @@ async def generate_pdf_endpoint(
             raise HTTPException(status_code=500, detail="Invalid AI response format")
         
         # Generate PDF
-        response = json.loads(response)
+        response_data = json.loads(response)
-        pdf_content = create_pdf(response)
+        pdf_content = create_pdf(response_data)
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
-        filename = f"theme_{timestamp}.pdf"
+        file_path = f"theme_{timestamp}.pdf"
         
-        print(f"Returning PDF with filename: {filename}")
+        # Save the PDF locally temporarily
+        with open(file_path, "wb") as file:
+            file.write(pdf_content)
         
-        return Response(
+        # Upload the PDF to S3 using the API
-            content=pdf_content,
+        upload_url = f"{os.getenv('BACKEND_BASE_URL')}/v3/api/custom/theme/doc-upload?x-project={x_api_key}"
-            media_type="application/pdf",
+        with open(file_path, 'rb') as file:
-            headers={"Content-Disposition": f'attachment; filename="{filename}"'}
+            files = {'file': file}
-        )
+            upload_response = requests.post(upload_url, files=files)
         
+        # Check if the upload was successful
+        if upload_response.status_code != 200:
+            raise HTTPException(status_code=upload_response.status_code, detail="File upload to S3 failed: " + upload_response.text)
+        
+        upload_data = upload_response.json()  # Get the response in JSON format
+        
+        # Extract the uploaded file URL
+        theme_url = upload_data.get("url")  # Adjust this key based on the actual API response structure
+        if not theme_url:
+            raise HTTPException(status_code=500, detail="Failed to retrieve theme URL from upload response")
+        
+        # Clean up the temporary file
+        os.remove(file_path)
+        
+        # Return JSON response with theme URL and text
+        return {
+            "theme_url": theme_url,
+            "theme_text": response_data
+        }
     except Exception as e:
         print(f"Error generating theme: {str(e)}")
         raise HTTPException(status_code=500, detail=f"Error: {str(e)}")

src/llm.py

@@ -7,16 +7,28 @@ from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
 from langchain_core.messages import HumanMessage, AIMessage, BaseMessage
 from langgraph.checkpoint.memory import MemorySaver
 from langgraph.graph import START, MessagesState, StateGraph
-from utils.utils import format_questions_text
 from src.prompts import chat_prompt
 from langchain_openai import ChatOpenAI
 from src.models import Phase2Generation,Phase1Generation
 from config  import TEMPERATURE
+import os
+import requests
+import json
+from dotenv import load_dotenv
+from typing import List, Dict, Optional
+from dataclasses import dataclass
+from datetime import datetime
+from langchain_core.messages import HumanMessage, AIMessage
+
+# Load environment variables
+load_dotenv()
+
+
 @dataclass
 class Message:
     role: str  # 'human' or 'ai'
     content: str
-    timestamp: str
+    timestamp: str=None
 
 QUESTIONS_PATH = "./data/config_files/questions.json"
 with open(QUESTIONS_PATH, "r") as f:
@@ -46,63 +58,31 @@ def setup_prompt_template(theme: int, resume: str,full_history=None,form_respons
         ("system", chat_prompt(theme, resume,full_history,form_response,generate_theme)),
         MessagesPlaceholder(variable_name="messages")
     ])
+def fetch_conversation_history(conversation_id: str) -> List[Message]:
+    """
+    Fetch conversation history from the API using the conversation ID.
+    """
+    x_api_key = os.getenv("BACKEND_XAPI_KEY")
+    base_url = os.getenv("BACKEND_BASE_URL")
+    url = f"{base_url}/v3/api/custom/jordan/ai-chat/get-messages/{conversation_id}?x-project={x_api_key}"
     
-def parse_ai_response(content: str) -> Dict:
-    """Parse AI response content into expected format"""
     try:
-        response = json.loads(content)
+        response = requests.get(url)
-        return {
+        response.raise_for_status()  # Raise an error for bad responses
-            "message": response.get("message", ""),
+        data = response.json()["data"]  # First JSON parse
-            "end": response.get("end", "no") == "yes"
+        data = json.loads(json.loads(data))
-        }
+        # Parse the API response into Message objects
-    except json.JSONDecodeError:
+        messages = []
-        return {
+        for item in data:
-            "message": content,
+            role = item.get("role", "unknown")
-            "end": False
+            content = item.get("content", "")
-        }
+            timestamp = datetime.now().isoformat()  # Use current timestamp if not provided
+            messages.append(Message(role=role, content=content))
+        return messages
+    except requests.RequestException as e:
+        print(f"Error fetching conversation history: {e}")
+        return []
 
-def add_message(storage_path: Path, conversation_id: str, role: str, content: str) -> None:
-    """Add a message to the conversation history"""
-    message_data = {
-        "role": role,
-        "content": content,
-        "timestamp": datetime.now().isoformat()
-    }
-
-    conversations = load_conversations(storage_path)
-    if conversation_id not in conversations:
-        conversations[conversation_id] = {"messages": []}
-    conversations[conversation_id]["messages"].append(message_data)
-    save_conversations(storage_path, conversations)
-
-
-def get_conversation_history(conversation_id: str, storage_path: Path) -> List[Message]:
-    """Get the conversation history"""
-    conversations = load_conversations(storage_path)
-    if conversation_id not in conversations:
-        return None
-
-    return [
-        Message(
-            role=msg["role"],
-            content=msg["content"],
-            timestamp=msg["timestamp"]
-        )
-        for msg in conversations[conversation_id]["messages"]
-    ]
-
-def load_conversations(storage_path: Path) -> Dict:
-    """Load conversations from storage file"""
-    try:
-        with open(storage_path, 'r') as f:
-            return json.load(f)
-    except FileNotFoundError:
-        return {}
-
-def save_conversations(storage_path: Path, conversations: Dict) -> None:
-    """Save conversations to storage file"""
-    with open(storage_path, 'w') as f:
-        json.dump(conversations, f, indent=2)
 
 def convert_to_langchain_messages(messages: List[Message]) -> List[HumanMessage | AIMessage]:
     """Convert our Message objects to LangChain message objects"""
@@ -114,20 +94,19 @@ def convert_to_langchain_messages(messages: List[Message]) -> List[HumanMessage
             converted_messages.append(AIMessage(content=msg.content))
     return converted_messages
 
-
 def ai_chat(query: str, conversation_id: str, theme_id: int, resume: str, full_history=None, form_response=None, generate_theme="NO") -> str:
     """Main chat function that processes queries and manages conversation"""
-    storage_path = Path("conversations.json")
-    
     class State(TypedDict):
-        messages: Annotated[Sequence[BaseMessage], "The messages in the conversation"]
+        messages: List[HumanMessage | AIMessage]
         language: str
+
     # Initialize model and workflow
     model = ChatOpenAI(model=MODEL, temperature=TEMPERATURE)
     if generate_theme == "YES":
         model = model.with_structured_output(Phase2Generation)
     else:
         model = model.with_structured_output(Phase1Generation)
+
     workflow = StateGraph(state_schema=State)
     
     def call_model(state: State):
@@ -142,12 +121,11 @@ def ai_chat(query: str, conversation_id: str, theme_id: int, resume: str,full_hi
     workflow.add_edge(START, "model")
     workflow.add_node("model", call_model)
     
-
     memory = MemorySaver()
     app = workflow.compile(checkpointer=memory)
     
-    # Get conversation history
+    # Fetch conversation history from the API
-    history = get_conversation_history(conversation_id, storage_path)
+    history = fetch_conversation_history(conversation_id)
     
     config = {"configurable": {"thread_id": conversation_id}}
     language = "English"
@@ -155,65 +133,37 @@ def ai_chat(query: str, conversation_id: str, theme_id: int, resume: str,full_hi
     if not history:
         # New conversation
         input_messages = [HumanMessage(content=query)] if query else [HumanMessage(content="Let's get started")]
-        output = app.invoke(
-            {"messages": input_messages, "language": language},
-            config
-        )
     else:
         # Existing conversation
         history = convert_to_langchain_messages(history)
         input_messages = history + [HumanMessage(content=query)] if query else history
+    
     output = app.invoke(
         {"messages": input_messages, "language": language},
         config
     )
     
-    
-
     if generate_theme == "YES":
         structured_message = output["messages"][0]
         output = structured_message.json(by_alias=True)  # This returns a JSON string.
-        print(f"Output: {output}")
-        if query:
-            add_message(storage_path, conversation_id, "human", query)
-        add_message(storage_path, conversation_id, "ai", output)
     else:
-    
         structured_message = output["messages"][0]
         output = structured_message.json()  # This returns a JSON string.
         output = json.loads(output)
-        print(output)
+
         message = output.get("message")
         print(output)
-        if query:
+    
-            add_message(storage_path, conversation_id, "human", query)
-        add_message(storage_path, conversation_id, "ai", message)
-    print(output)
     return output
 
 
+
 # Example usage:
 if __name__ == "__main__":
-    # Sample resume
+    #conversation_id = "12345"  # Replace with the actual conversation ID
-    sample_resume = """
+    query = "Hello let us continue"
-    John Doe
+    theme_id = 1
-    EMT-B Certified
+    resume = "Emergency Response Specialist"
-    5 years experience as volunteer firefighter
+    conversation_id = 1
-    Bachelor's in Fire Science
+    response = ai_chat(query, conversation_id, theme_id, resume)
-    """
+    print(response)
-    
-    # Sample conversation
-    conversation_id = "12345"
-    theme_id = 1  # Customer Service theme
-    
-    # Start conversation
-    
-    
-    # Continue conversation
-    follow_up = ai_chat(
-        query="What was my last questions?",
-        conversation_id=conversation_id,
-        theme_id=theme_id,
-        resume=sample_resume
-    )
-    print("AI:", follow_up)

src/prompts.py

@@ -329,7 +329,7 @@ def chat_prompt(theme,resume,full_history=None, form_response=None,generate_them
                   NOTE: DO NOT KEEP THE CONVERSATION excessively long , CAREFULL ANALYZE USER RESUME AND THE PROVIDED EXAMPLES QUESTIONS AND ALL CONTEXT , ASK RELEVANT QUESTION BASED ON THE THEME AND THAT IS ALL 
 
 
-                  Follow these instructions exactly:
+                  FOLLOW THESE  INSTRUCTIONS  STTRICTLY:
 
                               You may receive chat history that includes a previously generated STARTPOP framework, and the user may provide feedback on it. Use this feedback to engage with the user and ask clarifying questions as needed.
 
@@ -346,6 +346,9 @@ def chat_prompt(theme,resume,full_history=None, form_response=None,generate_them
 
                               If the user says, "I just want to make the actions a little bit clearer," a good response is: "Okay noted, would you like to generate your theme now?" (with the appropriate "end" and "pop_theme_generation" values).
                               If the user responds "yes," a good response is: "Go ahead and click on the theme generation button, thanks."
+
+                              "NEVER RETURN THE STARTPOP FRAME WORK FORMAT PLEASE" , SEE EXAMPLE RESPONSES I GAVE
+                              WHENE THERE IS THERE IS NEED TO GENERATE THE STARTPOP THEME FRAME , JUST TELL THEM TO GO AHEAD AND CLICK ON button
 Strictly adhere to these guidelines
                   """
     return prompt

test.py

@@ -1,13 +1,28 @@
 import os
 import requests
 import json
+from typing import List
 from dotenv import load_dotenv
 load_dotenv()
-doc_id = 2
 
-x_api_key = os.getenv("BACKEND_XAPI_KEY")
-url = f"{os.getenv('BACKEND_BASE_URL')}/v3/api/custom/theme-document/answer/{doc_id}?x-project={x_api_key}"
 
-result = requests.get(url)
+from src.llm import ai_chat
-response_json = result.json()  # Return response in JSON format
+
-print(response_json)
+
+#conversation_id = "12345"  # Replace with the actual conversation ID
+query = "Hello let us continue"
+theme_id = 1
+resume = "Emergency Response Specialist"
+conversation_id = 1
+response = ai_chat(query, conversation_id, theme_id, resume)
+print(response)
+"""
+with open(file_path, 'rb') as file:
+    files = {'file': file}
+    response = requests.post(upload_url, files=files)
+    response.raise_for_status()  # Ensure we raise an error for bad responses
+    response_data = response.json()  # Get the response in JSON format
+    print(response_data)
+
+"""
+