app.py

from langchain_community.document_loaders import (
    PyPDFLoader,
    TextLoader,
    DirectoryLoader,
    CSVLoader,
    UnstructuredExcelLoader,
    Docx2txtLoader,
)
from langchain.text_splitter import RecursiveCharacterTextSplitter, CharacterTextSplitter
import tiktoken
import chroma
import gradio as gr

import os, tempfile, glob, random
from pathlib import Path
from IPython.display import Markdown
from PIL import Image
from getpass import getpass
import numpy as np
from itertools import combinations

# LLM: openai and google_genai
import openai
from langchain_openai import OpenAI, OpenAIEmbeddings, ChatOpenAI
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_google_genai import GoogleGenerativeAIEmbeddings

# LLM: HuggingFace
from langchain_community.embeddings import HuggingFaceInferenceAPIEmbeddings
from langchain_community.llms import HuggingFaceHub

# langchain prompts, memory, chains...
from langchain.prompts import PromptTemplate, ChatPromptTemplate
from langchain.chains import ConversationalRetrievalChain
from langchain_community.chat_message_histories import StreamlitChatMessageHistory
from operator import itemgetter
from langchain_core.runnables import RunnableLambda, RunnableParallel, RunnablePassthrough
from langchain.schema import Document, format_document
from langchain_core.messages import AIMessage, HumanMessage, get_buffer_string

from langchain.retrievers.document_compressors import DocumentCompressorPipeline
from langchain.text_splitter import CharacterTextSplitter
from langchain_community.document_transformers import EmbeddingsRedundantFilter,LongContextReorder
from langchain.retrievers.document_compressors import EmbeddingsFilter
from langchain.retrievers import ContextualCompressionRetriever

from langchain.retrievers import ContextualCompressionRetriever
from langchain.retrievers.document_compressors import CohereRerank
from langchain_community.llms import Cohere

from langchain.memory import ConversationSummaryBufferMemory,ConversationBufferMemory


from langchain.schema import Document




def langchain_document_loader(TMP_DIR):
    """
    Load documents from the temporary directory (TMP_DIR). 
    Files can be in txt, pdf, CSV or docx format.
    """

    documents = []

    txt_loader = DirectoryLoader(
        TMP_DIR.as_posix(), glob="**/*.txt", loader_cls=TextLoader, show_progress=True
    )
    documents.extend(txt_loader.load())

    pdf_loader = DirectoryLoader(
        TMP_DIR.as_posix(), glob="**/*.pdf", loader_cls=PyPDFLoader, show_progress=True
    )
    documents.extend(pdf_loader.load())

    csv_loader = DirectoryLoader(
        TMP_DIR.as_posix(), glob="**/*.csv", loader_cls=CSVLoader, show_progress=True,
        loader_kwargs={"encoding":"utf8"}
    )
    documents.extend(csv_loader.load())

    doc_loader = DirectoryLoader(
        TMP_DIR.as_posix(),
        glob="**/*.docx",
        loader_cls=Docx2txtLoader,
        show_progress=True,
    )
    documents.extend(doc_loader.load())
    return documents



text_splitter = RecursiveCharacterTextSplitter(
    separators = ["\n\n", "\n", " ", ""],    
    chunk_size = 1600,
    chunk_overlap= 200
)

# Text splitting
chunks = text_splitter.split_documents(documents=documents)




def tiktoken_tokens(documents,model="gpt-3.5-turbo"):
    """Use tiktoken (tokeniser for OpenAI models) to return a list of token lengths per document."""    
    encoding = tiktoken.encoding_for_model(model) # returns the encoding used by the model.
    
    tokens_length = [len(encoding.encode(documents[i].page_content)) for i in range(len(documents))]

    return tokens_length


chunks_length = tiktoken_tokens(chunks,model="gpt-3.5-turbo")

print(f"Number of tokens - Average : {int(np.mean(chunks_length))}")
print(f"Number of tokens - 25% percentile : {int(np.quantile(chunks_length,0.25))}")
print(f"Number of tokens - 50% percentile : {int(np.quantile(chunks_length,0.5))}")
print(f"Number of tokens - 75% percentile : {int(np.quantile(chunks_length,0.75))}")



def select_embeddings_model(LLM_service="HuggingFace"):
    """Connect to the embeddings API endpoint by specifying 
    the name of the embedding model.
    if LLM_service == "OpenAI":
        embeddings = OpenAIEmbeddings(
            model='text-embedding-ada-002',
            api_key=openai_api_key)

    if LLM_service == "Google":
        embeddings = GoogleGenerativeAIEmbeddings(
            model="models/embedding-001",
            google_api_key=google_api_key
        )"""
    if LLM_service == "HuggingFace":
        embeddings = HuggingFaceInferenceAPIEmbeddings(    
            api_key=HF_key, 
            model_name="thenlper/gte-large"
        )
         
    return embeddings

#embeddings_OpenAI = select_embeddings_model(LLM_service="OpenAI")
#embeddings_google = select_embeddings_model(LLM_service="Google")
embeddings_HuggingFace = select_embeddings_model(LLM_service="HuggingFace")




def create_vectorstore(embeddings,documents,vectorstore_name):
    """Create a Chroma vector database."""
    persist_directory = (LOCAL_VECTOR_STORE_DIR.as_posix() + "/" + vectorstore_name)
    vector_store = Chroma.from_documents(
        documents=documents,
        embedding=embeddings,
        persist_directory=persist_directory
    )
    return vector_store



create_vectorstores = True # change to True to create vectorstores

if create_vectorstores:
    """
    vector_store_OpenAI,_ = create_vectorstore(
        embeddings=embeddings_OpenAI,
        documents = chunks,
        vectorstore_name="Vit_All_OpenAI_Embeddings",
    )
    print("vector_store_OpenAI:",vector_store_OpenAI._collection.count(),"chunks.")
    
    vector_store_google,new_vectorstore_name = create_vectorstore(
        embeddings=embeddings_google,
        documents = chunks,
        vectorstore_name="Vit_All_Google_Embeddings"
    )
    print("vector_store_google:",vector_store_google._collection.count(),"chunks.")
    """
    
    vector_store_HF = create_vectorstore(
        embeddings=embeddings_HuggingFace,
        documents = chunks,
        vectorstore_name="Vit_All_HF_Embeddings"
    )
    print("vector_store_HF:",vector_store_HF._collection.count(),"chunks.")
    print("")

"""
vector_store_OpenAI = Chroma(
    persist_directory = LOCAL_VECTOR_STORE_DIR.as_posix() + "/Vit_All_OpenAI_Embeddings",
    embedding_function=embeddings_OpenAI)
print("vector_store_OpenAI:",vector_store_OpenAI._collection.count(),"chunks.")

vector_store_google = Chroma(
    persist_directory = LOCAL_VECTOR_STORE_DIR.as_posix() + "/Vit_All_Google_Embeddings",
    embedding_function=embeddings_google)
print("vector_store_google:",vector_store_google._collection.count(),"chunks.")
"""

vector_store_HF = Chroma(
    persist_directory = LOCAL_VECTOR_STORE_DIR.as_posix() + "/Vit_All_HF_Embeddings",
    embedding_function=embeddings_HuggingFace)
print("vector_store_HF:",vector_store_HF._collection.count(),"chunks.")


def Vectorstore_backed_retriever(
vectorstore,search_type="similarity",k=4,score_threshold=None
):
    """create a vectorsore-backed retriever
    Parameters: 
        search_type: Defines the type of search that the Retriever should perform.
            Can be "similarity" (default), "mmr", or "similarity_score_threshold"
        k: number of documents to return (Default: 4) 
        score_threshold: Minimum relevance threshold for similarity_score_threshold (default=None)
    """
    search_kwargs={}
    if k is not None:
        search_kwargs['k'] = k
    if score_threshold is not None:
        search_kwargs['score_threshold'] = score_threshold

    retriever = vectorstore.as_retriever(
        search_type=search_type,
        search_kwargs=search_kwargs
    )
    return retriever

# similarity search
#base_retriever_OpenAI = Vectorstore_backed_retriever(vector_store_OpenAI,"similarity",k=10)
#base_retriever_google = Vectorstore_backed_retriever(vector_store_google,"similarity",k=10)
base_retriever_HF = Vectorstore_backed_retriever(vector_store_HF,"similarity",k=10)



def create_compression_retriever(embeddings, base_retriever, chunk_size=500, k=16, similarity_threshold=None):
    """Build a ContextualCompressionRetriever.
    We wrap the the base_retriever (a vectorstore-backed retriever) into a ContextualCompressionRetriever.
    The compressor here is a Document Compressor Pipeline, which splits documents
    into smaller chunks, removes redundant documents, filters out the most relevant documents,
    and reorder the documents so that the most relevant are at the top and bottom of the list.
    
    Parameters:
        embeddings: OpenAIEmbeddings, GoogleGenerativeAIEmbeddings or HuggingFaceInferenceAPIEmbeddings.
        base_retriever: a vectorstore-backed retriever.
        chunk_size (int): Documents will be splitted into smaller chunks using a CharacterTextSplitter with a default chunk_size of 500. 
        k (int): top k relevant chunks to the query are filtered using the EmbeddingsFilter. default =16.
        similarity_threshold : minimum relevance threshold used by the EmbeddingsFilter. default =None.
    """
    
    # 1. splitting documents into smaller chunks
    splitter = CharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=0, separator=". ")
    
    # 2. removing redundant documents
    redundant_filter = EmbeddingsRedundantFilter(embeddings=embeddings)

    # 3. filtering based on relevance to the query    
    relevant_filter = EmbeddingsFilter(embeddings=embeddings, k=k, similarity_threshold=similarity_threshold) # similarity_threshold and top K

    # 4. Reorder the documents 
    
    # Less relevant document will be at the middle of the list and more relevant elements at the beginning or end of the list.
    # Reference: https://python.langchain.com/docs/modules/data_connection/retrievers/long_context_reorder
    reordering = LongContextReorder()

    # 5. Create compressor pipeline and retriever
    
    pipeline_compressor = DocumentCompressorPipeline(
        transformers=[splitter, redundant_filter, relevant_filter, reordering]  
    )
    compression_retriever = ContextualCompressionRetriever(
        base_compressor=pipeline_compressor, 
        base_retriever=base_retriever
    )

    return compression_retriever

def CohereRerank_retriever(
    base_retriever, 
    cohere_api_key,cohere_model="rerank-multilingual-v2.0", top_n=8
):
    """Build a ContextualCompressionRetriever using Cohere Rerank endpoint to reorder the results based on relevance.
    Parameters:
       base_retriever: a Vectorstore-backed retriever
       cohere_api_key: the Cohere API key
       cohere_model: The Cohere model can be either 'rerank-english-v2.0' or 'rerank-multilingual-v2.0', with the latter being the default.
       top_n: top n results returned by Cohere rerank, default = 8.
    """
    
    compressor = CohereRerank(
        cohere_api_key=cohere_api_key, 
        model=cohere_model, 
        top_n=top_n
    )

    retriever_Cohere = ContextualCompressionRetriever(
        base_compressor=compressor,
        base_retriever=base_retriever
    )
    return retriever_Cohere



def instantiate_LLM(LLM_provider,api_key,temperature=0.5,top_p=0.95,model_name=None):
    """Instantiate LLM in Langchain.
    Parameters:
        LLM_provider (str): the LLM provider; in ["OpenAI","Google","HuggingFace"]
        model_name (str): in ["gpt-3.5-turbo", "gpt-3.5-turbo-0125", "gpt-4-turbo-preview", 
            "gemini-pro", "mistralai/Mistral-7B-Instruct-v0.2"].            
        api_key (str): google_api_key or openai_api_key or huggingfacehub_api_token 
        temperature (float): Range: 0.0 - 1.0; default = 0.5
        top_p (float): : Range: 0.0 - 1.0; default = 1.
    """
    if LLM_provider == "OpenAI":
        llm = ChatOpenAI(
            api_key=api_key,
            model=model_name, # in ["gpt-3.5-turbo", "gpt-3.5-turbo-0125", "gpt-4-turbo-preview"]
            temperature=temperature,
            model_kwargs={
                "top_p": top_p
            }
        )
    if LLM_provider == "Google":
        llm = ChatGoogleGenerativeAI(
            google_api_key=api_key,
            model=gemini-pro, # "gemini-pro"
            temperature=temperature,
            top_p=top_p,
            convert_system_message_to_human=True
        )
    if LLM_provider == "HuggingFace":
        llm = HuggingFaceHub(
            repo_id="mistralai/Mistral-7B-Instruct-v0.2", # "mistralai/Mistral-7B-Instruct-v0.2"
            huggingfacehub_api_token=api_key,
            model_kwargs={
                "temperature":temperature,
                "top_p": top_p,
                "do_sample": True,
                "max_new_tokens":1024
            },
        )
    return llm


def get_environment_variable(key):
    if key in os.environ:
        value = os.environ.get(key)
        print(f"\n[INFO]: {key} retrieved successfully.")
    else :
        print(f"\n[ERROR]: {key} is not found in your environment variables.") 
        value = getpass(f"Insert your {key}")
    return value

openai_api_key = os.environ['openai_key']
google_api_key = os.environ['gemini_key']
HF_key = os.environ['HF_token']
cohere_api_key = os.environ['cohere_api']




def create_memory(model_name='gpt-3.5-turbo',memory_max_token=None):
    """Creates a ConversationSummaryBufferMemory for gpt-3.5-turbo.
    Creates a ConversationBufferMemory for the other models."""
    
    if model_name=="gpt-3.5-turbo":
        if memory_max_token is None:
            memory_max_token = 1024 # max_tokens for 'gpt-3.5-turbo' = 4096
        memory = ConversationSummaryBufferMemory(
            max_token_limit=memory_max_token,
            llm=ChatOpenAI(model_name="gpt-3.5-turbo",openai_api_key=openai_api_key,temperature=0.1),
            return_messages=True,
            memory_key='chat_history',
            output_key="answer",
            input_key="question"
        )
    else:
        memory = ConversationBufferMemory(
            return_messages=True,
            memory_key='chat_history',
            output_key="answer",
            input_key="question",
        )  
    return memory

memory.save_context(inputs={"question":"..."},outputs={"answer":"...."})

standalone_question_template = """Given the following conversation and a follow up question, 
rephrase the follow up question to be a standalone question, in its original language.\n\n
Chat History:\n{chat_history}\n
Follow Up Input: {question}\n
Standalone question:"""

standalone_question_prompt = PromptTemplate(
    input_variables=['chat_history', 'question'], 
    template=standalone_question_template
)


def answer_template(language="english"):
    """Pass the standalone question along with the chat history and context 
    to the `LLM` wihch will answer"""
    
    template = f"""Answer the question at the end, using only the following context (delimited by <context></context>).
Your answer must be in the language at the end. 

<context>
{{chat_history}}

{{context}} 
</context>

Question: {{question}}

Language: {language}.
"""
    return template


chain = ConversationalRetrievalChain.from_llm(
    condense_question_prompt=standalone_question_prompt,
    combine_docs_chain_kwargs={'prompt': answer_prompt},
    condense_question_llm=instantiate_LLM(
        LLM_provider="Google",api_key=HF_key,temperature=0.1,
        model_name="gemini-pro"),
    memory=create_memory("gemini-pro"),
    retriever = retriever, 
    llm=instantiate_LLM(
        LLM_provider="Google",api_key=HF_key,temperature=0.5,
        model_name="gemini-pro"),
    chain_type= "stuff",
    verbose= False,
    return_source_documents=True   
)



# 1. load memory using RunnableLambda. Retrieves the chat_history attribute using itemgetter.
# `RunnablePassthrough.assign` adds the chat_history to the assign function

loaded_memory = RunnablePassthrough.assign(
    chat_history=RunnableLambda(memory.load_memory_variables) | itemgetter("chat_history"),
)

# 2. Pass the follow-up question along with the chat history to the LLM, and parse the answer (standalone_question).

condense_question_prompt = PromptTemplate(
    input_variables=['chat_history', 'question'], 
    template=standalone_question_template
)

condense_question_llm = instantiate_LLM(
    LLM_provider="Google",api_key=google_api_key,temperature=0.1,
    model_name="gemini-pro"
)

standalone_question_chain = {
    "standalone_question": {
        "question": lambda x: x["question"],
        "chat_history": lambda x: get_buffer_string(x["chat_history"]),
    }
    | condense_question_prompt
    | condense_question_llm
    | StrOutputParser(),
}

# 3. Combine load_memory and standalone_question_chain

chain_question = loaded_memory | standalone_question_chain


memory.clear()
memory.save_context(
    {"question": "What does DTC stand for?"},
    {"answer": "Diffuse to Choose."}
)
print("Chat history:\n",memory.load_memory_variables({}))

follow_up_question = "plaese give more details about it, including its use cases and implementation."
print("\nFollow-up question:\n",follow_up_question)

# invoke chain_question
response = chain_question.invoke({"question":follow_up_question})["standalone_question"]
print("\nStandalone_question:\n",response)



def _combine_documents(docs, document_prompt, document_separator="\n\n"):
    doc_strings = [format_document(doc, document_prompt) for doc in docs]
    return document_separator.join(doc_strings)

# 1. Retrieve relevant documents

retrieved_documents = {
    "docs": itemgetter("standalone_question") | retriever,
    "question": lambda x: x["standalone_question"],
}

# 2. Get variables ['chat_history', 'context', 'question'] that will be passed to `answer_prompt`

DEFAULT_DOCUMENT_PROMPT = PromptTemplate.from_template(template="{page_content}")
answer_prompt = ChatPromptTemplate.from_template(answer_template()) # 3 variables are expected ['chat_history', 'context', 'question']

answer_prompt_variables = {
    "context": lambda x: _combine_documents(docs=x["docs"],document_prompt=DEFAULT_DOCUMENT_PROMPT),
    "question": itemgetter("question"),
    "chat_history": itemgetter("chat_history") # get chat_history from `loaded_memory` variable
}

llm = instantiate_LLM(
    LLM_provider="Google",api_key=google_api_key,temperature=0.5,
    model_name="gemini-pro"
)

# 3. Load memory, format `answer_prompt` with variables (context, question and chat_history) and pass the `answer_prompt to LLM.
# return answer, docs and standalone_question

chain_answer = {
    "answer": loaded_memory | answer_prompt_variables | answer_prompt | llm,
    "docs": lambda x: [
        Document(page_content=doc.page_content,metadata=doc.metadata) # return only page_content and metadata 
        for doc in x["docs"]
    ], 
    "standalone_question": lambda x:x["question"] # return standalone_question
}


conversational_retriever_chain = chain_question | retrieved_documents | chain_answer
follow_up_question = "plaese give more details about it, including its use cases and implementation."

response = conversational_retriever_chain.invoke({"question":follow_up_question})
Markdown(response['answer'].content)


memory.save_context( 
 {"question": follow_up_question},
 {"answer": response['answer'].content} 
)









css = """
      #col-container {max-width: 80%; margin-left: auto; margin-right: auto;}
      #chatbox {min-height: 400px;}
      #header {text-align: center;}
      #prompt_template_preview {padding: 1em; border-width: 1px; border-style: solid; border-color: #e0e0e0; border-radius: 4px; min-height: 150px;}
      #total_tokens_str {text-align: right; font-size: 0.8em; color: #666;}
      #label {font-size: 0.8em; padding: 0.5em; margin: 0;}
      .message { font-size: 1.2em; }
      """

with gr.Blocks(css=css) as demo:
    
    state = gr.State(get_empty_state())


    with gr.Column(elem_id="col-container"):

           
        gr.Markdown("""## Ask questions of *needs assessment* experts,  
                    ## get responses from a *needs assessment experts* version of ChatGPT.  
                    Ask questions of all of them, or pick your expert below.
                    This is a free resource but it does cost us money to run. Unfortunately someone has been abusing this approach.
                    In response, we have had to temporarily turn it off until we can put improve the monitoring. Sorry for the inconvenience.""" ,
                    elem_id="header")
        
        
        with gr.Row():
            with gr.Column():
                chatbot = gr.Chatbot(elem_id="chatbox")
                input_message = gr.Textbox(show_label=False, placeholder="Enter your needs assessment question", visible=True).style(container=False)

                btn_submit = gr.Button("Submit")
                #total_tokens_str = gr.Markdown(elem_id="total_tokens_str")
                btn_clear_conversation = gr.Button("Start New Conversation")
            with gr.Column():
                prompt_template = gr.Dropdown(label="Choose an Expert:", choices=list(prompt_templates.keys()))
                prompt_template_preview = gr.Markdown(elem_id="prompt_template_preview")
                with gr.Accordion("Advanced parameters", open=False):
                    temperature = gr.Slider(minimum=0, maximum=2.0, value=0.7, step=0.1, label="Flexibility", info="Higher = More AI, Lower = More Expert")
                    max_tokens = gr.Slider(minimum=100, maximum=400, value=200, step=1, label="Length of Response.")
                    context_length = gr.Slider(minimum=1, maximum=5, value=2, step=1, label="Context Length", info="Number of previous questions you have asked.")

   
    btn_submit.click(submit_message, [ input_message, prompt_template, temperature, max_tokens, context_length, state], [input_message, chatbot,  state])
    input_message.submit(submit_message, [ input_message, prompt_template, temperature, max_tokens, context_length, state], [input_message, chatbot,  state])
    btn_clear_conversation.click(clear_conversation, [], [input_message, chatbot,  state])
    prompt_template.change(on_prompt_template_change_description, inputs=[prompt_template], outputs=[prompt_template_preview])

    
    demo.load(download_prompt_templates, inputs=None, outputs=[prompt_template], queur=False)


demo.queue(concurrency_count=10)
demo.launch(height='800px')