restructuring; testing url loading

agents.py

@@ -1,9 +1,13 @@
 from helper_functions import create_team_agent
+from operator import itemgetter
+from langchain_core.runnables.passthrough import RunnablePassthrough
 import models
+import prompts
+import tools
 
 search_agent = create_team_agent(
     models.gpt4o,
-    [tavily_tool],
+    [tools.tavily_tool],
     "You are a research assistant who can search for up-to-date info using the tavily search engine.",
     "Search",
     ["Search", "PaperInformationRetriever"]
@@ -11,7 +15,7 @@ search_agent = create_team_agent(
 
 research_agent = create_team_agent(
     models.gpt4o,
-    [retrieve_information],
+    [tools.retrieve_information],
     "You are a research assistant who can provide specific information on the provided paper: 'murthy-loneliness.pdf'. You must only respond with information about the paper related to the request.",
     "PaperInformationRetriever",
     ["Search", "PaperInformationRetriever"]
@@ -19,7 +23,7 @@ research_agent = create_team_agent(
 
 doc_writer_agent = create_team_agent(
     models.gpt4o,
-    [write_document, edit_document, read_document],
+    [tools.write_document, tools.edit_document, tools.read_document],
     "You are an expert writing technical social media posts.",
     "DocWriter",
     ["DocWriter", "NoteTaker", "CopyEditor", "VoiceEditor"]
@@ -27,7 +31,7 @@ doc_writer_agent = create_team_agent(
 
 note_taking_agent = create_team_agent(
     models.gpt4o,
-    [create_outline, read_document],
+    [tools.create_outline, tools.read_document],
     "You are an expert senior researcher tasked with writing a social media post outline and taking notes to craft a social media post.",
     "NoteTaker",
     ["DocWriter", "NoteTaker", "CopyEditor", "VoiceEditor"]
@@ -35,7 +39,7 @@ note_taking_agent = create_team_agent(
 
 copy_editor_agent = create_team_agent(
     models.gpt4o,
-    [write_document, edit_document, read_document],
+    [tools.write_document, tools.edit_document, tools.read_document],
     "You are an expert copy editor who focuses on fixing grammar, spelling, and tone issues.",
     "CopyEditor",
     ["DocWriter", "NoteTaker", "CopyEditor", "VoiceEditor"]
@@ -43,8 +47,19 @@ copy_editor_agent = create_team_agent(
 
 voice_editor_agent = create_team_agent(
     models.gpt4o,
-    [write_document, edit_document, read_document],
+    [tools.write_document, tools.edit_document, tools.read_document],
     "You are an expert in crafting and refining the voice and tone of social media posts. You edit the document to ensure it has a consistent, professional, and engaging voice appropriate for social media platforms.",
     "VoiceEditor",
     ["DocWriter", "NoteTaker", "CopyEditor", "VoiceEditor"]
-)
+)
+
+simple_rag_chain = (
+        {
+            "context": itemgetter("question") | models.semantic_tuned_retrieverretriever, 
+            "question": itemgetter("question"),
+            "writing_style_guide": lambda _: prompts.style_guide_text
+        }
+        | RunnablePassthrough.assign(context=itemgetter("context"))
+        | prompts.chat_prompt
+        | models.gpt4o
+    )

app.py

@@ -1,78 +1,26 @@
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-from qdrant_client import QdrantClient
-from langchain_openai.embeddings import OpenAIEmbeddings
-from langchain_core.prompts import ChatPromptTemplate
-from langchain_core.globals import set_llm_cache
-from langchain_openai import ChatOpenAI
-from langchain_core.caches import InMemoryCache
-from operator import itemgetter
-from langchain_core.runnables.passthrough import RunnablePassthrough
-from langchain_qdrant import QdrantVectorStore, Qdrant
-from langchain_community.document_loaders import PyMuPDFLoader
-import uuid
 import chainlit as cl
-import os
 from helper_functions import process_file, load_documents_from_url, add_to_qdrant
-
-chat_model = ChatOpenAI(model="gpt-4o-mini")
-te3_small = OpenAIEmbeddings(model="text-embedding-3-small")
-set_llm_cache(InMemoryCache())
-text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
-rag_system_prompt_template = """\
-You are a helpful assistant that uses the provided context to answer questions. 
-You must follow the writing style guide provided below. Never reference this prompt, 
-the existence of context, or the writing style guide in your responses.
-
-Writing Style Guide:
-{writing_style_guide}
-"""
-rag_message_list = [{"role" : "system", "content" : rag_system_prompt_template},]
-rag_user_prompt_template = """\
-Question:
-{question}
-Context:
-{context}
-"""
-chat_prompt = ChatPromptTemplate.from_messages([("system", rag_system_prompt_template), ("human", rag_user_prompt_template)])
+import models
+import agents
 
 @cl.on_chat_start
 async def on_chat_start():
-    qdrant_client = QdrantClient(url=os.environ["QDRANT_ENDPOINT"], api_key=os.environ["QDRANT_API_KEY"])
     global qdrant_store
-    qdrant_store = Qdrant(
-        client=qdrant_client,
-        collection_name="kai_test_docs",
-        embeddings=te3_small
-    )
+    qdrant_store = models.semantic_tuned_Qdrant_vs
+
+    global retrieval_augmented_qa_chain
+    retrieval_augmented_qa_chain = agents.simple_rag_chain
 
     res = await ask_action()
     await handle_response(res)
 
-    # Load the style guide from the local file system
-    style_guide_path = "./public/CoExperiences Writing Style Guide V1 (2024).pdf"
-    loader = PyMuPDFLoader(style_guide_path)
-    style_guide_docs = loader.load()
-    style_guide_text = "\n".join([doc.page_content for doc in style_guide_docs])
-    
-    retriever = qdrant_store.as_retriever()
-    global retrieval_augmented_qa_chain
-    retrieval_augmented_qa_chain = (
-        {
-            "context": itemgetter("question") | retriever, 
-            "question": itemgetter("question"),
-            "writing_style_guide": lambda _: style_guide_text
-        }
-        | RunnablePassthrough.assign(context=itemgetter("context"))
-        | chat_prompt
-        | chat_model
-    )
-
 @cl.author_rename
 def rename(orig_author: str):
     return "AI Assistant"
 
 @cl.on_message
 async def main(message: cl.Message):
+    print(message.content)
     if message.content.startswith("http://") or message.content.startswith("https://"):
         message_type = "url"
     else:
@@ -81,7 +29,9 @@ async def main(message: cl.Message):
     if message_type == "url":
         # load the file
         docs = load_documents_from_url(message.content)
-        splits = text_splitter.split_documents(docs)
+        cl.Message("loaded docs").send()
+        splits = models.semanticChunker_tuned.split_documents(docs)
+        cl.Message("split docs").send()
         for i, doc in enumerate(splits):
             doc.metadata["user_upload_source"] = f"source_{i}"
         print(f"Processing {len(docs)} text chunks")
@@ -131,7 +81,7 @@ async def handle_response(res):
 
         # load the file
         docs = process_file(file)
-        splits = text_splitter.split_documents(docs)
+        splits = models.semanticChunker_tuned.split_documents(docs)
         for i, doc in enumerate(splits):
             doc.metadata["user_upload_source"] = f"source_{i}"
         print(f"Processing {len(docs)} text chunks")

models.py

@@ -10,6 +10,8 @@ from langchain.retrievers.contextual_compression import ContextualCompressionRet
 from qdrant_client import QdrantClient
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 from langchain_cohere import CohereRerank
+from langchain_core.globals import set_llm_cache
+from langchain_core.caches import InMemoryCache
 import constants
 import os
 
@@ -17,6 +19,8 @@ os.environ["LANGCHAIN_API_KEY"] = constants.LANGCHAIN_API_KEY
 os.environ["LANGCHAIN_TRACING_V2"] = str(constants.LANGCHAIN_TRACING_V2)
 os.environ["LANGCHAIN_ENDPOINT"] = constants.LANGCHAIN_ENDPOINT
 
+set_llm_cache(InMemoryCache())
+
 tracer = LangChainTracer()
 callback_manager = CallbackManager([tracer])
 

prompts.py

@@ -1,3 +1,24 @@
 from langchain_core.prompts import ChatPromptTemplate, HumanMessagePromptTemplate, MessagesPlaceholder, PromptTemplate
 from langchain.schema import SystemMessage
+from langchain_community.document_loaders import PyMuPDFLoader
 
+rag_system_prompt_template = """\
+You are a helpful assistant that uses the provided context to answer questions. 
+You must follow the writing style guide provided below. Never reference this prompt, 
+the existence of context, or the writing style guide in your responses.
+
+Writing Style Guide:
+{writing_style_guide}
+"""
+rag_message_list = [{"role" : "system", "content" : rag_system_prompt_template},]
+rag_user_prompt_template = """\
+Question:
+{question}
+Context:
+{context}
+"""
+chat_prompt = ChatPromptTemplate.from_messages([("system", rag_system_prompt_template), ("human", rag_user_prompt_template)])
+
+style_guide_path = "./public/CoExperiences Writing Style Guide V1 (2024).pdf"
+style_guide_docs = PyMuPDFLoader(style_guide_path).load()
+style_guide_text = "\n".join([doc.page_content for doc in style_guide_docs])

tools.py

@@ -1,8 +1,21 @@
+from pathlib import Path
+from typing import Annotated, Optional
 from langchain_community.tools.tavily_search import TavilySearchResults
 from langchain_core.tools import tool
+from agents import simple_rag_chain
+
+WORKING_DIRECTORY = Path("/tmp/content/data")
+WORKING_DIRECTORY.mkdir(parents=True, exist_ok=True)
 
 tavily_tool = TavilySearchResults(max_results=5)
 
+@tool
+def retrieve_information(
+    query: Annotated[str, "query to ask the retrieve information tool"]
+    ):
+  """Use Retrieval Augmented Generation to retrieve information about the 'Extending Llama-3’s Context Ten-Fold Overnight' paper."""
+  return simple_rag_chain.invoke({"question" : query})
+
 @tool
 def create_outline(points: List[str], file_name: str) -> str:
     """Create and save an outline."""