app.py

import streamlit as st
import pandas as pd
from openai import OpenAI
from PyPDF2 import PdfReader

# source: eagle0504/document-search-q-series
def read_and_textify_advanced(files, chunk_size):
    """
    Reads PDF files and extracts text from each page, breaking the text into specified segments.
    This function iterates over a list of uploaded PDF files, extracts text from each page,
    and compiles a list of texts and corresponding source information, segmented into smaller parts
    of approximately 'chunk_size' words each.
    Args:
    files (List[st.uploaded_file_manager.UploadedFile]): A list of uploaded PDF files.
    chunk_size (int): The number of words per text segment. Default is 50.
    Returns: A list of strings, where each string is a segment of text extracted from a PDF page.
    """

    text_list = []  # List to store extracted text segments

    # Iterate over each file
    for file in files:
        pdfReader = PdfReader(file)  # Create a PDF reader object
        # Iterate over each page in the PDF
        for i in range(len(pdfReader.pages)):
            pageObj = pdfReader.pages[i]  # Get the page object
            text = pageObj.extract_text()  # Extract text from the page
            if text:
                # Split text into chunks of approximately 'chunk_size' words
                words = text.split(".")
                for j in range(0, len(words), chunk_size):
                    # Get the chunk of text from j-chunk_size to j+chunk_size
                    # start = max(0, j - chunk_size)
                    # end = min(len(words), j + chunk_size + 1)
                    chunk = ".".join(words[j:j+chunk_size]) + '.'
                    chunk = chunk.strip()
                    text_list.append(chunk)
                    # Create a source identifier for each chunk and add it to the list
            else:
                # If no text extracted, still add a placeholder
                text_list.append("")
            pageObj.clear()  # Clear the page object (optional, for memory management)

    return text_list

def get_questions(context, instructions) -> str:
    """
    Given a text context, generates a list of questions using OpenAI's GPT-3 API.

    Args:
    - context: A string representing the context for which questions should be generated.

    Returns:
    - A string containing the question generated by the API.
    """

    try:
        response = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
              {"role": "system", "content": "You are a helpful assistant."},
              {"role": "user", "content": f"{instructions}\n\nText: {context}\n\nQuestions:\n"}
            ]
        )
        # Extract question text from the response
        question_text = response.choices[0].message.content
        return question_text
    except:
        # Return an empty string if there was an error
        return ""

def get_answers(row, instructions) -> str:
    """
    Given a dataframe row containing context and questions, generates an answer using OpenAI's GPT-3 API.

    Args:
    - row: A pandas dataframe row containing 'context' and 'questions' columns.

    Returns:
    - A string containing the answer generated by the API.
    """

    try:
        response = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
              {"role": "system", "content": "You are a helpful assistant."},
              {"role": "user", "content": f"{instructions}\n\nText: {row.context}\n\nQuestions:\n{row.questions}\n\nAnswers:\n"}
            ]
        )
        # Extract answer text from the response
        answer_text = response.choices[0].message.content
        return answer_text
    except Exception as e:
        # Print the error message and return an empty string if there was an error
        print (e)
        return ""

st.set_page_config(page_title="ChatbotGuide", layout="wide")
st.title("Chatbot Guide")

# Define the options in the dropdown menu
app_options = [
    "1) Create CSVs",
    "2) Merge CSVs",
    "3) Upload Datasets",
    "4) Create Chatbot"
]

# Sidebar dropdown for selecting the application
selected_app = st.sidebar.selectbox("Select Step (1-4)", app_options)

# Clear session state when switching apps
if 'last_selected_app' in st.session_state:
    if st.session_state.last_selected_app != selected_app:
        st.session_state.clear()

st.session_state.last_selected_app = selected_app

if 'submit' not in st.session_state:
    st.session_state.submit = False
if 'error' not in st.session_state:
    st.session_state.error = ""
if 'success' not in st.session_state:
    st.session_state.success = None

if selected_app == "1) Create CSVs":  
    if st.session_state.error != "":
        st.error(st.session_state.error)
    
    if st.session_state.success != None:
        st.success("Success! Download the Q/A pairs below / Click reset to upload more PDFs")
        st.download_button(
            label=f"Download CSV: length = {st.session_state.success[1]}",
            data=st.session_state.success[0],
            file_name='questions_answers.csv',
            mime='text/csv',
        )
        if st.button('Reset'):
            st.session_state.clear()
            st.rerun()
            
    else:
        uploaded_files = st.file_uploader("Upload PDFs Here", type="pdf", accept_multiple_files=True)
        
        question_protocol = st.text_input("Provide instructions for how questions should be generated", "Write a question based on the text")
        answer_protocol = st.text_input("Provide instructions for how answers should be generated", "Write an answer based on the text")
        
        sentence_chunks = st.number_input("Number sentences per Q/A pair", value=2, step=1, min_value=1, max_value=3)
        
        openai_api_key = st.text_input("Enter your OpenAI API key", type="password")
    
        submit = st.button("Submit")
        if submit:
            st.session_state.submit = True
        
        if st.session_state.submit:
            if uploaded_files:
                client = OpenAI(api_key=openai_api_key)

                with st.spinner("Loading, please be patient with us ... 🙏"):
                    # test api key
                    try:
                        response = client.chat.completions.create(
                            model="gpt-4o-mini",
                            messages=[
                              {"role": "system", "content": "You are a helpful assistant."},
                              {"role": "user", "content": "Say this is a test"}
                            ]
                        )
                    except:
                        st.session_state.clear()
                        st.session_state.error = "OpenAI API key is invalid"
                        st.rerun()
                    
                with st.spinner("Loading, please be patient with us ... 🙏"):
                    textify_output = read_and_textify_advanced(uploaded_files, sentence_chunks)
                    
                    df = pd.DataFrame(textify_output)
                    df.columns = ['context']
            
                    if question_protocol == "":
                        question_protocol = "Write questions based on the text"
                    df['questions'] = df.apply(lambda row: get_questions(row['context'], question_protocol), axis=1)
            
                    if answer_protocol == "":
                        answer_protocol = "Write answers based on the text"
                    df['answers'] = df.apply(lambda row: get_answers(row, answer_protocol), axis=1)
            
                    df = df.drop('context', axis=1)
                    length = len(df)
                    
                    csv = df.to_csv(index=False).encode('utf-8')

                    st.session_state.clear()
                    st.session_state.success = (csv, length)
                    st.rerun()
            else:
                st.session_state.clear()
                st.session_state.error = "Please upload at least 1 PDF"
                st.rerun()

if selected_app == "2) Merge CSVs":
    if st.session_state.error != "":
        st.error(st.session_state.error)
    
    if st.session_state.success != None:
        st.success("Success! Download the merged CSV with Q/A pairs below / Reset to merge more CSVs")
        st.download_button(
            label=f"Download CSV: length = {st.session_state.success[1]}",
            data=st.session_state.success[0],
            file_name='questions_answers.csv',
            mime='text/csv',
        )
        if st.button('Reset'):
            st.session_state.clear()
            st.rerun()
            
    else:
        uploaded_files = st.file_uploader("Upload CSV files to merge", accept_multiple_files=True, type="csv")
    
        submit = st.button("Submit")
        if submit:
            st.session_state.submit = True
        
        if st.session_state.submit:
            if len(uploaded_files) > 1:
                dfs = []
                for file in uploaded_files:
                    df = pd.read_csv(file)
                    
                    if "questions" in df.columns and "answers" in df.columns:
                        df = df[["questions", "answers"]]
                        dfs.append(df)
                
                    else:
                        st.session_state.clear()
                        st.session_state.error = "Please upload CSVs that have been generated from 1) Create CSV"
                        st.rerun()
                        
                df = pd.concat(dfs, ignore_index=True)
                length = len(df)
                
                csv = df.to_csv(index=False).encode('utf-8')
                
                st.session_state.clear()
                st.session_state.success = (csv, length)
                st.rerun()
                
            else:
                st.session_state.clear()
                st.session_state.error = "Please upload at least 2 CSVs to merge"
                st.rerun()

if selected_app == "3) Upload Datasets":
    st.markdown("Go to this [google colab link](https://colab.research.google.com/drive/1eCpk9HUoCKZb--tiNyQSHFW2ojoaA35m) to get started")

if selected_app == "4) Create Chatbot":
    if st.session_state.error != "":
        st.error(st.session_state.error)

    if st.session_state.success != None:
        st.success("Success! Copy/paste the requirements.txt and app.py files into your HuggingFace Space")
        
        st.write('requirements.txt')
        st.code(st.session_state.success[0], language='python')

        st.write('app.py')
        st.code(st.session_state.success[1], language='python')
    
        if st.button('Reset'):
            st.session_state.clear()
            st.rerun()
    
    else:
        organization_name = st.text_input("What is the name of your organization", "")
        num_domains = st.number_input("How many datasets do you have uploaded", value=1, step=1, min_value=1, max_value=10)

        st.divider()

        domain_info = []
        for i in range(num_domains):
            domain_link = st.text_input(f"Please enter link to dataset {i+1} with the format username/dataset_name", "Example: KeshavRa/About_YSA_Database")
            domain_name = st.text_input(f"What should domain {i+1} be called in the chatbot itself", "Example: About YSA")
            domain_purpose = st.text_area(f"What is the purpose of domain {i+1}, provide example questions (this will be visible to users of the chatbot)", '''Example: On this page, you can learn about what YSA does, how YSA was started, the advisory board, and the programs we offer.
            Example Questions
            --> What is the purpose of Youth Spirit Artworks?
            --> Who created YSA?
            --> What is the Advisory Board for Youth Spirit Artworks?
            --> What are the three empowerment-focused program areas of YSA?
            '''
            )
            domain_instructions = st.text_input(f"What baseline instructions/specifications should be sent to ChatGPT to answer questions in domain {i+1}", "Example: You are an assistant to help the user learn more about Youth Spirit Artworks")
            
            domain = {"link": domain_link, "name": domain_name, "purpose": domain_purpose, "instructions": domain_instructions}
            domain_info.append(domain)
            st.divider()
        
        submit = st.button("Submit")
        if submit:
            st.session_state.submit = True
        
        if st.session_state.submit:
            if organization_name == "":
                st.session_state.clear()
                st.session_state.error = "Please enter an organization name"
                st.rerun()
            
            requirements = '''
                openai
                scipy
                streamlit
                chromadb
                datasets
            '''
        
            app = f"""
                import os
                import streamlit as st
                from datasets import load_dataset
                import chromadb
                import string
            
                from openai import OpenAI
                
                import numpy as np
                import pandas as pd
                
                from scipy.spatial.distance import cosine
                
                from typing import Dict, List
                
                def merge_dataframes(dataframes):
                    # Concatenate the list of dataframes
                    combined_dataframe = pd.concat(dataframes, ignore_index=True)
            
                    # Ensure that the resulting dataframe only contains the columns "context", "questions", "answers"
                    combined_dataframe = combined_dataframe[['context', 'questions', 'answers']]
                
                    return combined_dataframe
            
                def call_chatgpt(prompt: str, directions: str) -> str:
                    '''
                    Uses the OpenAI API to generate an AI response to a prompt.
                    Args:
                        prompt: A string representing the prompt to send to the OpenAI API.
                    Returns:
                        A string representing the AI's generated response.
                    '''
                
                    # Use the OpenAI API to generate a response based on the input prompt.
                    client = OpenAI(api_key = os.environ["OPENAI_API_KEY"])
                
                    completion = client.chat.completions.create(
                    model="gpt-3.5-turbo-0125",
                    messages=[
                        {{"role": "system", "content": directions}},
                        {{"role": "user", "content": prompt}}
                    ]
                    )
                
                    # Extract the text from the first (and only) choice in the response output.
                    ans = completion.choices[0].message.content
                
                    # Return the generated AI response.
                    return ans
                
                def openai_text_embedding(prompt: str) -> str:
                    return openai.Embedding.create(input=prompt, model="text-embedding-ada-002")[
                        "data"
                    ][0]["embedding"]
                
                def calculate_sts_openai_score(sentence1: str, sentence2: str) -> float:
                    # Compute sentence embeddings
                    embedding1 = openai_text_embedding(sentence1)  # Flatten the embedding array
                    embedding2 = openai_text_embedding(sentence2)  # Flatten the embedding array
                
                    # Convert to array
                    embedding1 = np.asarray(embedding1)
                    embedding2 = np.asarray(embedding2)
            
                    # Calculate cosine similarity between the embeddings
                    similarity_score = 1 - cosine(embedding1, embedding2)
                
                    return similarity_score
            
                def add_dist_score_column(
                    dataframe: pd.DataFrame, sentence: str,
                ) -> pd.DataFrame:
                    dataframe["stsopenai"] = dataframe["questions"].apply(
                            lambda x: calculate_sts_openai_score(str(x), sentence)
                    )
                    
                    sorted_dataframe = dataframe.sort_values(by="stsopenai", ascending=False)
                
                
                    return sorted_dataframe.iloc[:5, :]
            
                def convert_to_list_of_dict(df: pd.DataFrame) -> List[Dict[str, str]]:
                    '''
                    Reads in a pandas DataFrame and produces a list of dictionaries with two keys each, 'question' and 'answer.'
                    Args:
                        df: A pandas DataFrame with columns named 'questions' and 'answers'.
                    Returns:
                        A list of dictionaries, with each dictionary containing a 'question' and 'answer' key-value pair.
                    '''
                
                    # Initialize an empty list to store the dictionaries
                    result = []
                
                    # Loop through each row of the DataFrame
                    for index, row in df.iterrows():
                        # Create a dictionary with the current question and answer
                        qa_dict_quest = {{"role": "user", "content": row["questions"]}}
                        qa_dict_ans = {{"role": "assistant", "content": row["answers"]}}
            
                        # Add the dictionary to the result list
                        result.append(qa_dict_quest)
                        result.append(qa_dict_ans)
            
                    # Return the list of dictionaries
                    return result
                
                st.sidebar.markdown('''This is a chatbot to help you learn more about {organization_name}''')
                
                domain = st.sidebar.selectbox("Select a topic", "")
                
                special_threshold = 0.3
                
                n_results = 3
                
                clear_button = st.sidebar.button("Clear Conversation", key="clear")
                
                if clear_button:
                    st.session_state.messages = []
                    st.session_state.curr_domain = ""
        
        
                
                ### 
                ###
                ### Load the dataset from a provided source.
                ###
                ###  
                  
                initial_input = "Tell me about {organization_name}"
            
                # Initialize a new client for ChromeDB.
                client = chromadb.Client()
                
                # Generate a random number between 1 billion and 10 billion.
                random_number: int = np.random.randint(low=1e9, high=1e10)
                
                # Generate a random string consisting of 10 uppercase letters and digits.
                random_string: str = "".join(
                    np.random.choice(list(string.ascii_uppercase + string.digits), size=10)
                )
                
                # Combine the random number and random string into one identifier.
                combined_string: str = f"{{random_number}}{{random_string}}"
            
                # Create a new collection in ChromeDB with the combined string as its name.
                collection = client.create_collection(combined_string)
                
                st.title("{organization_name} Chatbot")
                
                # Initialize chat history
                if "messages" not in st.session_state:
                    st.session_state.messages = []
            
                if "curr_domain" not in st.session_state:
                    st.session_state.curr_domain = ""
                
                ### 
                ###
                ### init_messages dict (one key per domain)
                ###
                ###
            
                ### 
                ###
                ### chatbot_instructions dict (one key per domain)
                ###
                ###
                
                # Embed and store the first N supports for this demo
                with st.spinner("Loading, please be patient with us ... 🙏"):
                    L = len(dataset["train"]["questions"])
                    
                    collection.add(
                        ids=[str(i) for i in range(0, L)],  # IDs are just strings
                        documents=dataset["train"]["questions"],  # Enter questions here
                        metadatas=[{{"type": "support"}} for _ in range(0, L)],
                    )
                    
                    if st.session_state.curr_domain != domain:
                        st.session_state.messages = []
                        
                        init_message = init_messages[domain]
                        st.session_state.messages.append({{"role": "assistant", "content": init_message}})
                        
                        st.session_state.curr_domain = domain
                
                # Display chat messages from history on app rerun
                for message in st.session_state.messages:
                    with st.chat_message(message["role"]):
                        st.markdown(message["content"])
                
                # React to user input
                if prompt := st.chat_input("Tell me about {organization_name}"):
                    # Display user message in chat message container
                    st.chat_message("user").markdown(prompt)
                    # Add user message to chat history
                    st.session_state.messages.append({{"role": "user", "content": prompt}})
                
                    question = prompt
                
                    results = collection.query(query_texts=question, n_results=n_results)
                
                    idx = results["ids"][0]
                    idx = [int(i) for i in idx]
                    ref = pd.DataFrame(
                        {{
                            "idx": idx,
                            "questions": [dataset["train"]["questions"][i] for i in idx],
                            "answers": [dataset["train"]["answers"][i] for i in idx],
                            "distances": results["distances"][0],
                        }}
                    )
                    # special_threshold = st.sidebar.slider('How old are you?', 0, 0.6, 0.1) # 0.3
                    # special_threshold = 0.3
                    filtered_ref = ref[ref["distances"] < special_threshold]
                    if filtered_ref.shape[0] > 0:
                        # st.success("There are highly relevant information in our database.")
                        ref_from_db_search = filtered_ref["answers"].str.cat(sep=" ")
                        final_ref = filtered_ref
                    else:
                        # st.warning(
                        #     "The database may not have relevant information to help your question so please be aware of hallucinations."
                        # )
                        ref_from_db_search = ref["answers"].str.cat(sep=" ")
                        final_ref = ref
                
                    engineered_prompt = f'''
                        Based on the context: {{ref_from_db_search}},
                        answer the user question: {{question}}.
                    '''
                
                    directions = chatbot_instructions[domain]
                
                    answer = call_chatgpt(engineered_prompt, directions)
                
                    response = answer
                    # Display assistant response in chat message container
                    with st.chat_message("assistant"):
                        st.markdown(response)
                        with st.expander("See reference:"):
                            st.table(final_ref) 
                    # Add assistant response to chat history
                    st.session_state.messages.append({{"role": "assistant", "content": response}})
            """
            
            st.session_state.clear()
            st.session_state.success = (requirements, app)
            st.rerun()