modified history prompts

app/chains.py

@@ -5,6 +5,7 @@ import app.schemas as schemas
 from app.prompts import (
     raw_prompt,
     raw_prompt_formatted,
+    history_prompt_formatted,
     format_context,
     tokenizer
 )
@@ -23,10 +24,10 @@ llm = HuggingFaceEndpoint(
 simple_chain = (raw_prompt | llm).with_types(input_type=schemas.UserQuestion)
 
 # TODO: create formatted_chain by piping raw_prompt_formatted and the LLM endpoint.
-formatted_chain = raw_prompt_formatted | llm
+formatted_chain = (raw_prompt_formatted | llm).with_types(input_type=schemas.UserQuestion)
 
 # TODO: use history_prompt_formatted and HistoryInput to create the history_chain
-history_chain = None
+history_chain = (history_prompt_formatted | llm).with_types(input_type=schemas.HistoryInput)
 
 # TODO: Let's construct the standalone_chain by piping standalone_prompt_formatted with the LLM
 standalone_chain = None

app/crud.py

@@ -1,6 +1,8 @@
 from sqlalchemy.orm import Session
 import app.models as models
 import app.schemas as schemas
+from datetime import datetime
+
 
 
 def get_or_create_user(db: Session, username: str):
@@ -18,7 +20,14 @@ def add_message(db: Session, message: schemas.MessageBase, username: str):
     # - create a models.Message instance
     # - pass the retrieved user to the message instance
     # - save the message instance to the database
-    raise NotImplemented
+    user = get_or_create_user(db, username)
+    message = models.Message(message=message.message, type=message.type, user=user, timestamp=datetime.now(), user_id=user.id)
+    db.add(message)
+    db.commit()
+    db.refresh(message)
+    return message
 
 def get_user_chat_history(db: Session, username: str):
-    raise NotImplemented
+    
+    user = get_or_create_user(db, username)
+    return user.messages

app/main.py

@@ -6,13 +6,13 @@ from langserve.serialization import WellKnownLCSerializer
 from typing import List
 from sqlalchemy.orm import Session
 
-from app.chains import simple_chain, formatted_chain
+from app.chains import simple_chain, formatted_chain, history_chain
 import app.crud as crud
 import app.models as models
 import app.schemas as schemas
 from app.database import SessionLocal, engine
 from app.callbacks import LogResponseCallback
-
+from app.prompts import format_chat_history
 
 models.Base.metadata.create_all(bind=engine)
 
@@ -57,7 +57,14 @@ async def history_stream(request: Request, db: Session = Depends(get_db)):
     # - We add as part of the user history the current question by using add_message.
     # - We create an instance of HistoryInput by using format_chat_history.
     # - We use the history input within the history chain.
-    raise NotImplemented
+    data = await request.json()
+    history_input = schemas.HistoryInput(**data['input'])
+    user_name = history_input.username
+    crud.add_message(db, schemas.MessageBase(message=history_input.question, type="user"), user_name)
+    chat_history = crud.get_user_chat_history(db, user_name)
+    chat_history_str = format_chat_history(chat_history)
+    history_input.chat_history = chat_history_str
+    return EventSourceResponse(generate_stream(history_input, history_chain))
 
 
 @app.post("/rag/stream")

app/prompts.py

@@ -20,7 +20,10 @@ def format_prompt(prompt) -> PromptTemplate:
 def format_chat_history(messages: List[models.Message]):
     # TODO:  implement format_chat_history to format 
     # the list of Message into a text of chat history.
-    raise NotImplemented
+    chat_history = ""
+    for message in messages:
+        chat_history += f"{message.type}: {message.message}\n"
+    return chat_history
 
 def format_context(docs: List[str]):
     # TODO:  the output of the DataIndexer.search is a list of text, 
@@ -33,7 +36,11 @@ raw_prompt = "{question}"
 
 # TODO: Create the history_prompt prompt that will capture the question and the conversation history. 
 # The history_prompt needs a {chat_history} placeholder and a {question} placeholder.
-history_prompt: str = None
+history_prompt: str = """Given the following conversation provide a helpful answer to the follow up question.
+Chat History:
+{chat_history}
+Follow Up question: {question}
+helpful answer:"""
 
 # TODO: Create the standalone_prompt prompt that will capture the question and the chat history
 # to generate a standalone question. It needs a {chat_history} placeholder and a {question} placeholder,
@@ -48,11 +55,11 @@ raw_prompt_formatted = format_prompt(raw_prompt)
 raw_prompt = PromptTemplate.from_template(raw_prompt)
 
 # TODO: use format_prompt to create history_prompt_formatted
-history_prompt_formatted: PromptTemplate = None
+history_prompt_formatted: PromptTemplate = format_prompt(history_prompt)
 # TODO: use format_prompt to create standalone_prompt_formatted
-standalone_prompt_formatted: PromptTemplate = None
+standalone_prompt_formatted: PromptTemplate = format_prompt(standalone_prompt)
 # TODO: use format_prompt to create rag_prompt_formatted
-rag_prompt_formatted: PromptTemplate = None
+rag_prompt_formatted: PromptTemplate = format_prompt(rag_prompt)
 
 
 

app/schemas.py

@@ -6,12 +6,14 @@ class UserQuestion(BaseModel):
 
 # TODO: create a HistoryInput data model with a chat_history and question attributes.
 class HistoryInput(BaseModel):
-    pass
+    chat_history: str
+    question: str
 
 # TODO: let's create a UserRequest data model with a question and username attribute. 
 # This will be used to parse the input request.
 class UserRequest(BaseModel):
     username: str
+    question: str
 
 # TODO: implement MessageBase as a schema mapping from the database model to the 
 # FastAPI data model. Basically MessageBase should have the same attributes as models.Message