Update app.py

README.md

@@ -1,12 +1,15 @@
----
-title: BibleStudy
-emoji: 💻
-colorFrom: yellow
-colorTo: pink
-sdk: docker
-pinned: false
-license: openrail++
-short_description: 'Bible Study on the book of Genesis '
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+## <h1 align="center" id="heading">An Agentic Bible Study Tool Built with LangChain and LangGraph</h1>
+
+Create an intelligent Bible study assistant that utilizes LLMs to enhance contextual understanding of scripture. This tool will enable users to pose questions, and the AI will provide answers grounded in the Bible, by accurately identifying and synthesizing information from relevant verses, chapters, and cross-references, promoting deeper comprehension and reducing misinterpretations.
+
+
+### Phase I
+    - Book of Genesis
+    - Examples of questions:
+        - How did GOD create the whole universe based on Genesis?
+        - Why LORD God make man leave garden?
+        - How did the Israelites, led by Jacob, end up in Egypt, and what role did Joseph play in their settlement there?
+
+
+## Ship 🚢
+Check out the prototype at https://huggingface.co/spaces/kcheng0816/BibleStudy

app.py

@@ -1,10 +1,10 @@
 import os
 from dotenv import load_dotenv
-from chainlit.types import AskFileResponse
 import chainlit as cl
 
 import pandas as pd
 from langchain_community.vectorstores import FAISS
+from langchain_openai.embeddings import OpenAIEmbeddings
 from langchain_core.documents import Document
 from langchain_community.document_loaders import DirectoryLoader
 from langchain_community.document_loaders import BSHTMLLoader
@@ -25,7 +25,10 @@ from langchain_core.documents import Document
 from langchain_core.messages import HumanMessage
 from langchain_core.tools import tool
 from langgraph.prebuilt import ToolNode
-
+from langchain_core.messages import AnyMessage
+from langgraph.graph.message import add_messages
+from typing import TypedDict, Annotated
+from langchain_core.documents import Document
 
 #Load API Keys
 load_dotenv()
@@ -35,7 +38,7 @@ path = "data/"
 loader = DirectoryLoader(path, glob="*.html")
 docs = loader.load()
 
-#Building RAG Graph with LangGraph
+#Text Splitter
 text_splitter = RecursiveCharacterTextSplitter(
     chunk_size = 750,
     chunk_overlap = 100
@@ -43,9 +46,11 @@ text_splitter = RecursiveCharacterTextSplitter(
 
 split_documents = text_splitter.split_documents(docs)
 len(split_documents)
+
 #fine tuned embedding model
 huggingface_embeddings = HuggingFaceEmbeddings(model_name="kcheng0816/finetuned_arctic_genesis")
 
+#vector datastore
 client = QdrantClient(":memory:")
 client.create_collection(
     collection_name="genesis_bible",
@@ -59,6 +64,7 @@ vector_store = QdrantVectorStore(
 )
 
 _ = vector_store.add_documents(documents=split_documents)
+
 #Retrieve
 retriever = vector_store.as_retriever(search_kwargs={"k": 5})
 
@@ -71,7 +77,7 @@ def retrieve_adjusted(state):
   retrieved_docs = compression_retriever.invoke(state["question"])
   return {"context" : retrieved_docs}
 
-
+#RAG prompt
 RAG_PROMPT = """\
 You are a helpful assistant who answers questions based on provided context. You must only use the provided context, and cannot use your own knowledge.
 
@@ -84,7 +90,7 @@ You are a helpful assistant who answers questions based on provided context. You
 rag_prompt = ChatPromptTemplate.from_template(RAG_PROMPT)
 
 
-#llm 
+#llm for RAG
 rate_limiter = InMemoryRateLimiter(
     requests_per_second=1,  # <-- make a request once every 1 seconds!!
     check_every_n_seconds=0.1,  # Wake up every 100 ms to check whether allowed to make a request,
@@ -99,9 +105,7 @@ def generate(state):
   response = llm.invoke(messages)
   return {"response" : response.content}
 
-
-
-
+#Build RAG graph
 class State(TypedDict):
   question: str
   context: List[Document]
@@ -112,8 +116,6 @@ graph_builder.add_edge(START, "retrieve_adjusted")
 graph = graph_builder.compile()
 
 
-
-
 @tool
 def ai_rag_tool(question: str) -> str:
     """Useful for when you need to answer questions about Bible """
@@ -127,24 +129,18 @@ tool_belt = [
     ai_rag_tool
 ]
 
-
+#llm for agent reasoning
 llm = init_chat_model("gpt-4o", temperature=0, rate_limiter=rate_limiter)
 llm_with_tools = llm.bind_tools(tool_belt)
 
 
-from langgraph.graph import END
-from langchain_core.messages import AnyMessage
-from langgraph.graph.message import add_messages
-from typing import TypedDict, Annotated
-from langchain_core.documents import Document
-
 
+#Build an agent graph
 class AgentState(TypedDict):
     messages: Annotated[list[AnyMessage], add_messages]
     context:List[Document]
 
 
-
 def call_mode(state):
     messages = state["messages"]
     response = llm_with_tools.invoke(messages)
@@ -164,7 +160,6 @@ def should_continue(state):
     return END
 
 
-#
 uncompiled_graph = StateGraph(AgentState)
 
 uncompiled_graph.add_node("agent", call_mode)
@@ -179,13 +174,11 @@ uncompiled_graph.add_conditional_edges(
 
 uncompiled_graph.add_edge("action", "agent")
 
-# Compile and display the graph for a visual overview
+# Compile the graph.
 compiled_graph = uncompiled_graph.compile()
 
 
-
-
-
+#user interface
 @cl.on_chat_start
 async def on_chat_start():
    cl.user_session.set("graph", compiled_graph)