Update app.py

app.py

@@ -1,147 +1,147 @@
-import gradio as gr
-from langchain_openai import ChatOpenAI
-from langchain.memory import ConversationBufferMemory
-from langchain.chains import ConversationalRetrievalChain
-from langchain_community.document_loaders import UnstructuredCSVLoader
-from langchain_openai import OpenAIEmbeddings
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain_community.vectorstores import DocArrayInMemorySearch
-from operator import itemgetter
-from queue import SimpleQueue
-from langchain_core.runnables import RunnableLambda, RunnablePassthrough
-from langchain.prompts import PromptTemplate
-
-import os
-from dotenv import load_dotenv
-load_dotenv()
-API_KEY = os.getenv("OPENAI_API_KEY")
-#print("Hi Krish")
-
-def setup_qa_chain(file_path, openai_model):
-    docs = []
-    for filepath in file_path:
-        loader = UnstructuredCSVLoader(filepath)
-        docs.extend(loader.load())
-    return docs
-
-def split_docs(docs):
-    # Split documents
-    text_splitter = RecursiveCharacterTextSplitter(
-        chunk_size=500,
-        chunk_overlap=100
-    )
-    return text_splitter.split_documents(docs)
-
-def embed(docs, API_KEY):
-    # Create embeddings and store in vectordb
-    embeddings = OpenAIEmbeddings(openai_api_key=API_KEY)
-    vectordb = DocArrayInMemorySearch.from_documents(docs, embeddings)
-    return vectordb
-
-def create_retriever(vectordb):
-    # Define retriever
-    retriever = vectordb.as_retriever(
-        search_type='mmr'
-    )
-    return retriever
-
-def create_mem():
-    # Setup memory for contextual conversation
-    memory = ConversationBufferMemory(
-        memory_key='chat_history',
-        return_messages=True,
-        output_key='answer'  # Explicitly set the output key to avoid ambiguity
-    )
-    return memory
-
-def create_chat_chain(file_path, openai_model):
-    docs = setup_qa_chain(file_path, openai_model)
-    split_documents = split_docs(docs)
-    embeddings = embed(split_documents, API_KEY=API_KEY)
-    retr = create_retriever(embeddings)
-    mem = create_mem()
-    llm_def = ChatOpenAI(model=openai_model, temperature=0, api_key=API_KEY, max_tokens=1100)
-    chain = ConversationalRetrievalChain.from_llm(
-        llm=llm_def,
-        retriever=retr,
-        memory=mem,
-        return_source_documents=True,
-        verbose=True
-    )
-    return chain
-
-if __name__ == "__main__":
-    openai_model = "gpt-4o"
-    # file_path = [
-    #     r"C:\Users\Admin\Documents\CodeProjects_KD\RAG_ChatBot\people-1000.csv"
-        
-    # ]
-    file_path = [
-        r"people-1000.csv"
-        
-    ]
-    #user_query = input(">>")
-    chain = create_chat_chain(file_path, openai_model)
-
-queue = SimpleQueue()
-LLM_MODEL = "gpt-4o"
-llm = ChatOpenAI(
-    openai_api_key=os.environ['OPENAI_API_KEY'],
-    model=LLM_MODEL,
-    temperature=0.5
-)
-
-template = "You are an agent responsible for summarizing, and giving accurate responses regarding the High Jump Warehouse System Guide. '{question}' "
-prompt = PromptTemplate.from_template(template)
-memory = create_mem()
-
-# Inspect the output of memory.load_memory_variables
-#print(memory.load_memory_variables("inputs"))
-
-def safe_itemgetter(key):
-    def getter(d):
-        if key not in d:
-            raise KeyError(f"Key '{key}' not found in the dictionary. Available keys: {list(d.keys())}")
-        return d[key]
-    return getter
-
-conversation_chain = (
-    RunnablePassthrough.assign(
-        history=RunnableLambda(memory.load_memory_variables) | safe_itemgetter("chat_history")
-    )
-    | prompt
-    | llm
-)
-
-global_cb = None
-with gr.Blocks() as demo:
-    with gr.Row():
-        gr.Markdown("<h1 style='text-align: center; width: 100%;'>RAG with Memory (Warehouse Assistant)</h1>")
-    with gr.Row():
-        with gr.Column(scale=1):
-            gr.Markdown("**Here's how to use this demo:**")
-            gr.Markdown("""This a conversational agent responsible for summarizing, and giving key highlights about the High Jump Warehouse System Guide .\n
-
-            Feel free to experiment with your own different queries and usage or accesibility requests!
-            """)
-        with gr.Column(scale=3):
-            ex = ["What is the job title for Lindsey Rice?", "What is the Job Title for Joshua Carey?", "What is the user id for Dakota Gomez?"]
-            chatbot = gr.Chatbot()
-            msg = gr.Textbox()
-            gr.Examples(ex, msg)
-            clear = gr.Button("Clear Chat History ")
-            #but = gr.Button(value="Get Callback Data ")
-
-            def process_input(user_message, history):
-                inputs = {"question": user_message}
-                outputs = chain(inputs)
-                print(outputs)  # Debugging: Print the structure of the outputs
-                history.append([user_message, outputs['answer']])  # Use the correct key
-                return "", history
-
-            msg.submit(process_input, [msg, chatbot], [msg, chatbot])
-
-            clear.click(lambda: chain.memory.clear(), None, chatbot)
-
-demo.queue()
+import gradio as gr
+from langchain_openai import ChatOpenAI
+from langchain.memory import ConversationBufferMemory
+from langchain.chains import ConversationalRetrievalChain
+from langchain_community.document_loaders import UnstructuredCSVLoader
+from langchain_openai import OpenAIEmbeddings
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_community.vectorstores import DocArrayInMemorySearch
+from operator import itemgetter
+from queue import SimpleQueue
+from langchain_core.runnables import RunnableLambda, RunnablePassthrough
+from langchain.prompts import PromptTemplate
+
+import os
+from dotenv import load_dotenv
+load_dotenv()
+API_KEY = os.getenv("OPENAI_API_KEY")
+#print("Hi Krish")
+
+def setup_qa_chain(file_path, openai_model):
+    docs = []
+    for filepath in file_path:
+        loader = UnstructuredCSVLoader(filepath)
+        docs.extend(loader.load())
+    return docs
+
+def split_docs(docs):
+    # Split documents
+    text_splitter = RecursiveCharacterTextSplitter(
+        chunk_size=500,
+        chunk_overlap=100
+    )
+    return text_splitter.split_documents(docs)
+
+def embed(docs, API_KEY):
+    # Create embeddings and store in vectordb
+    embeddings = OpenAIEmbeddings(openai_api_key=API_KEY)
+    vectordb = DocArrayInMemorySearch.from_documents(docs, embeddings)
+    return vectordb
+
+def create_retriever(vectordb):
+    # Define retriever
+    retriever = vectordb.as_retriever(
+        search_type='mmr'
+    )
+    return retriever
+
+def create_mem():
+    # Setup memory for contextual conversation
+    memory = ConversationBufferMemory(
+        memory_key='chat_history',
+        return_messages=True,
+        output_key='answer'  # Explicitly set the output key to avoid ambiguity
+    )
+    return memory
+
+def create_chat_chain(file_path, openai_model):
+    docs = setup_qa_chain(file_path, openai_model)
+    split_documents = split_docs(docs)
+    embeddings = embed(split_documents, API_KEY=API_KEY)
+    retr = create_retriever(embeddings)
+    mem = create_mem()
+    llm_def = ChatOpenAI(model=openai_model, temperature=0, api_key=API_KEY, max_tokens=1100)
+    chain = ConversationalRetrievalChain.from_llm(
+        llm=llm_def,
+        retriever=retr,
+        memory=mem,
+        return_source_documents=True,
+        verbose=True
+    )
+    return chain
+
+if __name__ == "__main__":
+    openai_model = "gpt-4o"
+    # file_path = [
+    #     r"C:\Users\Admin\Documents\CodeProjects_KD\RAG_ChatBot\people-1000.csv"
+        
+    # ]
+    file_path = [
+        r"downloads/people-1000.csv"
+        
+    ]
+    #user_query = input(">>")
+    chain = create_chat_chain(file_path, openai_model)
+
+queue = SimpleQueue()
+LLM_MODEL = "gpt-4o"
+llm = ChatOpenAI(
+    openai_api_key=os.environ['OPENAI_API_KEY'],
+    model=LLM_MODEL,
+    temperature=0.5
+)
+
+template = "You are an agent responsible for summarizing, and giving accurate responses regarding the High Jump Warehouse System Guide. '{question}' "
+prompt = PromptTemplate.from_template(template)
+memory = create_mem()
+
+# Inspect the output of memory.load_memory_variables
+#print(memory.load_memory_variables("inputs"))
+
+def safe_itemgetter(key):
+    def getter(d):
+        if key not in d:
+            raise KeyError(f"Key '{key}' not found in the dictionary. Available keys: {list(d.keys())}")
+        return d[key]
+    return getter
+
+conversation_chain = (
+    RunnablePassthrough.assign(
+        history=RunnableLambda(memory.load_memory_variables) | safe_itemgetter("chat_history")
+    )
+    | prompt
+    | llm
+)
+
+global_cb = None
+with gr.Blocks() as demo:
+    with gr.Row():
+        gr.Markdown("<h1 style='text-align: center; width: 100%;'>RAG with Memory (Warehouse Assistant)</h1>")
+    with gr.Row():
+        with gr.Column(scale=1):
+            gr.Markdown("**Here's how to use this demo:**")
+            gr.Markdown("""This a conversational agent responsible for summarizing, and giving key highlights about the High Jump Warehouse System Guide .\n
+
+            Feel free to experiment with your own different queries and usage or accesibility requests!
+            """)
+        with gr.Column(scale=3):
+            ex = ["What is the job title for Lindsey Rice?", "What is the Job Title for Joshua Carey?", "What is the user id for Dakota Gomez?"]
+            chatbot = gr.Chatbot()
+            msg = gr.Textbox()
+            gr.Examples(ex, msg)
+            clear = gr.Button("Clear Chat History ")
+            #but = gr.Button(value="Get Callback Data ")
+
+            def process_input(user_message, history):
+                inputs = {"question": user_message}
+                outputs = chain(inputs)
+                print(outputs)  # Debugging: Print the structure of the outputs
+                history.append([user_message, outputs['answer']])  # Use the correct key
+                return "", history
+
+            msg.submit(process_input, [msg, chatbot], [msg, chatbot])
+
+            clear.click(lambda: chain.memory.clear(), None, chatbot)
+
+demo.queue()
 demo.launch()