chatGPT_own_PDFs

import tkinter
import customtkinter
from bs4 import BeautifulSoup

# Langchain loads:
from langchain.document_loaders import DirectoryLoader,PagedPDFSplitter
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import ElasticVectorSearch, Pinecone, Weaviate, FAISS, Qdrant
from langchain.chains.qa_with_sources import load_qa_with_sources_chain
from langchain.llms import OpenAI
from langchain.prompts import PromptTemplate

# import filedialog module
from tkinter import filedialog


customtkinter.set_appearance_mode("System")  # Modes: system (default), light, dark
customtkinter.set_default_color_theme("blue")  # Themes: blue (default), dark-blue, green

app = customtkinter.CTk()  # create CTk window like you do with the Tk window
app.geometry("1080x960")

OPENAI_API_KEY="YOUR API GOES HERE!"
embeddings = OpenAIEmbeddings(openai_api_key=OPENAI_API_KEY)

  
# Function for opening the
# file explorer window
def browseFiles():
    filename = filedialog.askdirectory()
    loader = DirectoryLoader(filename, glob="*.pdf")
    pdfs = loader.load_and_split()
    faiss_index = FAISS.from_documents(pdfs,embeddings) 
    faiss_index.save_local("faiss_index")


# Function for opening the
# file explorer window
def browseDocStore():
    faiss_file = filedialog.askdirectory(initialdir = "/", title = "Select a Vectorstore Index")
    faiss_index = FAISS.load_local(faiss_file , embeddings)


# Search Button
def button_function():
    # clear previous search
    textbox.delete("0.0", "end")
    textbox1.delete("0.0", "end")
    textbox2.delete("0.0", "end")
    textbox3.delete("0.0", "end")
    textbox4.delete("0.0", "end")

    textbox1b.delete("0.0", "end")
    textbox2b.delete("0.0", "end")
    textbox3b.delete("0.0", "end")
    textbox4b.delete("0.0", "end")

    # get question:
    query = entry.get() 
    faiss_index = FAISS.load_local("faiss_index" , embeddings)
    docs = faiss_index.similarity_search(query=query, k=5)

    # set up prompt:
    question_prompt_template = """Summarize the following text in 250 tokens or less:
{context}"""
    QUESTION_PROMPT = PromptTemplate(
        template=question_prompt_template, input_variables=["context"]
    )
    combine_prompt_template = """Given the following extracted parts of a number of long documents and a question, create a final answer. 
    If you don't know the answer, just say that you don't know. Don't try to make up an answer.

    QUESTION: {question}
    =========
    {summaries}
    =========
    FINAL ANSWER:"""
    COMBINE_PROMPT = PromptTemplate(
    template=combine_prompt_template, input_variables=["summaries", "question"] 
    )
    chain = load_qa_with_sources_chain(OpenAI(temperature=0,openai_api_key=OPENAI_API_KEY), chain_type="map_reduce", return_intermediate_steps=True, question_prompt=QUESTION_PROMPT, combine_prompt=COMBINE_PROMPT)
    
    # get answer and display the sources:
    answer = chain({"input_documents": docs, "question": query}, return_only_outputs=False)  
    textbox.insert("0.0", answer["output_text"])
    textbox1.insert("0.0", answer["intermediate_steps"][0])
    textbox1b.insert("0.0", answer["input_documents"][0])
    textbox2.insert("0.0", answer["intermediate_steps"][1])
    textbox2b.insert("0.0", answer["input_documents"][1])
    textbox3.insert("0.0", answer["intermediate_steps"][2])
    textbox3b.insert("0.0", answer["input_documents"][2])
    textbox4.insert("0.0", answer["intermediate_steps"][3])
    textbox4b.insert("0.0", answer["input_documents"][3])



# seach field
entry = customtkinter.CTkEntry(master=app, placeholder_text="query",width = 600)
entry.pack(padx=20, pady=10)
entry.place(relx=0.5, rely=0.03, anchor=tkinter.CENTER)
# output textbox
textbox = customtkinter.CTkTextbox(master=app,  width = 750, height=150)
textbox.pack(pady=20, padx=10)
textbox.place(relx=0.57, rely=0.15, anchor=tkinter.CENTER)


tabview = customtkinter.CTkTabview(master=app, width=750)
tabview.pack(pady=20, padx=10)
tabview.place(relx=0.5, rely=0.6, anchor=tkinter.CENTER)
tabview.add("Source and Summary 1")
tabview.add("Source and Summary 2")
tabview.add("Source and Summary 3")
tabview.add("Source and Summary 4")

    # Sources textbox
textbox1 = customtkinter.CTkTextbox(tabview.tab("Source and Summary 1"),  width = 950, height=150)
textbox1.pack(pady=20, padx=10)
textbox1b = customtkinter.CTkTextbox(tabview.tab("Source and Summary 1"),  width = 950, height=350)
textbox1b.pack(pady=20, padx=10)
textbox2 = customtkinter.CTkTextbox(tabview.tab("Source and Summary 2"),  width = 950, height=150)
textbox2.pack(pady=20, padx=10)
textbox2b = customtkinter.CTkTextbox(tabview.tab("Source and Summary 2"),  width = 950, height=350)
textbox2b.pack(pady=20, padx=10)
textbox3 = customtkinter.CTkTextbox(tabview.tab("Source and Summary 3"),  width = 950, height=150)
textbox3.pack(pady=20, padx=10)
textbox3b = customtkinter.CTkTextbox(tabview.tab("Source and Summary 3"),  width = 950, height=350)
textbox3b.pack(pady=20, padx=10)
textbox4 = customtkinter.CTkTextbox(tabview.tab("Source and Summary 4"),  width = 950, height=150)
textbox4.pack(pady=20, padx=10)
textbox4b = customtkinter.CTkTextbox(tabview.tab("Source and Summary 4"),  width = 950, height=350)
textbox4b.pack(pady=20, padx=10)




# Browse for folder of ducuments:
button = customtkinter.CTkButton(master=app, text="Index files", command=browseFiles)
button.place(relx=0.1, rely=0.1, anchor=tkinter.CENTER)

# Browse for an existing faiss store:
button = customtkinter.CTkButton(master=app, text="Load existing Index", command=browseDocStore)
button.place(relx=0.1, rely=0.05, anchor=tkinter.CENTER)


# defien seach button:
button2 = customtkinter.CTkButton(master=app, text="Search", command=button_function)
button2.place(relx=0.9, rely=0.03, anchor=tkinter.CENTER)



app.mainloop()

Sorry for your trouble! Now you click “index files” and select a folder of files to index. When they done you can search the file with ChatGPT!

Sorry to bother you. I am not able to get it to work. I still get the following error:
... # get question:
... query = entry.get()
NameError: name 'entry' is not defined

faiss_index = FAISS.load_local("faiss_index" , embeddings)
RuntimeError: Error in __cdecl faiss::FileIOReader::FileIOReader(const char *) at D:\a\faiss-wheels\faiss-wheels\faiss\faiss\impl\io.cpp:68: Error: 'f' failed: could not open faiss_index\index.faiss for reading: No such file or directory
docs = faiss_index.similarity_search(query=query, k=5)
NameError: name 'faiss_index' is not defined

Would it be possible to start with something much smaller. Open one PDF file and get it to provide the key points in a bullet for summary? THanks for the consideration.