File size: 9,550 Bytes
b2646b2
 
8d79d7e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
beefdef
 
 
 
 
 
 
8d79d7e
 
 
c7ea54b
beefdef
 
 
 
 
 
8b057ae
beefdef
 
 
8b057ae
 
beefdef
 
8d79d7e
 
beefdef
 
 
 
 
 
b4b9a43
beefdef
 
 
b4b9a43
 
beefdef
 
8d79d7e
 
beefdef
 
 
 
 
 
 
 
333bd63
beefdef
 
8d79d7e
 
 
 
 
 
 
 
 
333bd63
 
 
8d79d7e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
import csv
import json
import streamlit as st
from dotenv import load_dotenv
from PyPDF2 import PdfReader
from langchain.text_splitter import CharacterTextSplitter, RecursiveCharacterTextSplitter
from langchain.embeddings import OpenAIEmbeddings, HuggingFaceInstructEmbeddings
from langchain.vectorstores import FAISS, Chroma
from langchain.embeddings import HuggingFaceEmbeddings  # General embeddings from HuggingFace models.
from langchain.chat_models import ChatOpenAI
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationalRetrievalChain
from htmlTemplates import css, bot_template, user_template
from langchain.llms import HuggingFaceHub, LlamaCpp, CTransformers  # For loading transformer models.
from langchain.document_loaders import PyPDFLoader, TextLoader, JSONLoader, CSVLoader
import tempfile # ์ž„์‹œ ํŒŒ์ผ์„ ์ƒ์„ฑํ•˜๊ธฐ ์œ„ํ•œ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์ž…๋‹ˆ๋‹ค.
import os


# PDF ๋ฌธ์„œ๋กœ๋ถ€ํ„ฐ ํ…์ŠคํŠธ๋ฅผ ์ถ”์ถœํ•˜๋Š” ํ•จ์ˆ˜์ž…๋‹ˆ๋‹ค.
def get_pdf_text(pdf_docs):
    temp_dir = tempfile.TemporaryDirectory()                        # ์ž„์‹œ ๋””๋ ‰ํ† ๋ฆฌ ์ƒ์„ฑ
    temp_filepath = os.path.join(temp_dir.name, pdf_docs.name)      # ์ž„์‹œ ํŒŒ์ผ ๊ฒฝ๋กœ ์ƒ์„ฑ
    with open(temp_filepath, "wb") as f:                # ์ž„์‹œ ํŒŒ์ผ ๋ฐ”์ด๋„ˆ๋ฆฌ ์“ฐ๊ธฐ ๋ชจ๋“œ๋กœ ์—ด๊ธฐ
        f.write(pdf_docs.getvalue())                    # PDF ๋ฌธ์„œ ๋‚ด์šฉ ์ž„์‹œ ํŒŒ์ผ์— ์“ฐ๊ธฐ
    pdf_loader = PyPDFLoader(temp_filepath)             # PyPDFLoader๋กœ PDF ๋กœ๋“œ
    pdf_doc = pdf_loader.load()                         # ํ…์ŠคํŠธ ์ถ”์ถœ
    return pdf_doc                                      # ์ถ”์ถœํ•œ ํ…์ŠคํŠธ ๋ฐ˜ํ™˜

# ๊ณผ์ œ
# ์•„๋ž˜ ํ…์ŠคํŠธ ์ถ”์ถœ ํ•จ์ˆ˜๋ฅผ ์ž‘์„ฑ
def get_text_file(docs):
    temp_dir2 = tempfile.TemporaryDirectory()                 # ์ž„์‹œ ๋””๋ ‰ํ† ๋ฆฌ ์ƒ์„ฑ
    temp_filepath2 = os.path.join(temp_dir2.name, docs.name)  # ์ž„์‹œ ํŒŒ์ผ ๊ฒฝ๋กœ ์ƒ์„ฑ
    with open(temp_filepath2, "wb") as f:                     # ์ž„์‹œ ํŒŒ์ผ ๋ฐ”์ด๋„ˆ๋ฆฌ ์“ฐ๊ธฐ ๋ชจ๋“œ๋กœ ์—ด๊ธฐ
        f.write(docs.getvalue())                              # text ๋ฌธ์„œ์˜ ๋‚ด์šฉ ์ž„์‹œ ํŒŒ์ผ์— ์“ฐ๊ธฐ
    txt_loader = TextLoader(                                  # TextLoader๋กœ text ํŒŒ์ผ ๋กœ๋“œ
        file_path=temp_filepath2,                             # text ๋ฌธ์„œ์˜ ๋‚ด์šฉ์ด ์“ฐ์ธ ํŒŒ์ผ ๊ฒฝ๋กœ
        txt_args={
            "delimiter": " ",                                 # ๋‚ด์šฉ์€ ๋„์–ด์“ฐ๊ธฐ๋กœ ๊ตฌ๋ถ„
            # ์ž‘๋™์„ ์•ˆ ํ•ด์„œ ์ž„์˜๋กœ ๋‚ด์šฉ ๋„ฃ๊ธฐ
            #"content":'"What is the most important thing in Team project? I think it is communication. No matter how good an individual ability is I think it is difficult to achieve good results without communicating with each other a lot."'
        }
    )
    txt_data = txt_loader.load()                              # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ์ €์žฅ
    return txt_data                                           # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ๋ฐ˜ํ™˜

def get_csv_file(docs):
    temp_dir3 = tempfile.TemporaryDirectory()                   # ์ž„์‹œ ๋””๋ ‰ํ† ๋ฆฌ ์ƒ์„ฑ
    temp_filepath3 = os.path.join(temp_dir3.name, docs.name)    # ์ž„์‹œ ํŒŒ์ผ ๊ฒฝ๋กœ ์ƒ์„ฑ
    with open(temp_filepath3, "wb") as f:                       # ์ž„์‹œ ํŒŒ์ผ ๋ฐ”์ด๋„ˆ๋ฆฌ ์“ฐ๊ธฐ ๋ชจ๋“œ๋กœ ์—ด๊ธฐ
        f.write(docs.getvalue())                                # csv ๋ฌธ์„œ์˜ ๋‚ด์šฉ ์ž„์‹œ ํŒŒ์ผ์— ์“ฐ๊ธฐ
    csv_loader = CSVLoader(                                     # CSVLoader๋กœ csv ํŒŒ์ผ ๋กœ๋“œ
        file_path=temp_filepath3,                               # CSV ๋ฌธ์„œ์˜ ๋‚ด์šฉ์ด ์“ฐ์ธ ํŒŒ์ผ ๊ฒฝ๋กœ
        csv_args={
            "delimiter": ",",                                   # ๋‚ด์šฉ์€ ์‰ผํ‘œ๋กœ ๊ตฌ๋ถ„
            "quotechar": '"',                                   # ๋ฌธ์ž์—ด์€ "" ์•ˆ์— ์“ฐ์ž„
            "fieldnames": ["name", "school", "address", "phone"],   # ํ•„๋“œ ์ด๋ฆ„ ๋‚˜์—ด
        },
    )
    csv_data = csv_loader.load()                                # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ์ €์žฅ
    return csv_data                                             # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ๋ฐ˜ํ™˜

def get_json_file(docs):
    temp_dir4 = tempfile.TemporaryDirectory()                   # ์ž„์‹œ ๋””๋ ‰ํ† ๋ฆฌ ์ƒ์„ฑ
    temp_filepath4 = os.path.join(temp_dir4.name, docs.name)    # ์ž„์‹œ ํŒŒ์ผ ๊ฒฝ๋กœ ์ƒ์„ฑ
    with open(temp_filepath4, "wb") as f:                       # ์ž„์‹œ ํŒŒ์ผ ๋ฐ”์ด๋„ˆ๋ฆฌ ์“ฐ๊ธฐ ๋ชจ๋“œ๋กœ ์—ด๊ธฐ
        f.write(docs.getvalue())                                # json ๋ฌธ์„œ์˜ ๋‚ด์šฉ ์ž„์‹œ ํŒŒ์ผ์— ์“ฐ๊ธฐ
    json_loader = JSONLoader(                                   # JSONLoader๋กœ json ํŒŒ์ผ ๋กœ๋“œ
        file_path=temp_filepath4,                               # json ๋ฌธ์„œ์˜ ๋‚ด์šฉ์ด ์“ฐ์ธ ํŒŒ์ผ ๊ฒฝ๋กœ
        jq_schema='.messages[].content',                        # json ๋ฌธ์„œ์—์„œ ์ถ”์ถœํ•  ๋‚ด์šฉ ์„ค์ •(์ฑ„ํŒ… ๋ฉ”์‹œ์ง€)
        text_content=False                                      # ์ถ”์ถœํ•œ ๋ฐ์ดํ„ฐ๋Š” ํ…์ŠคํŠธ ํ˜•์‹์œผ๋กœ
    )
    json_data = json_loader.load()                              # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ์ €์žฅ
    return json_data                                            # ์ถ”์ถœ๋œ ํ…์ŠคํŠธ ๋ฐ˜ํ™˜

    
# ๋ฌธ์„œ๋“ค์„ ์ฒ˜๋ฆฌํ•˜์—ฌ ํ…์ŠคํŠธ ์ฒญํฌ๋กœ ๋‚˜๋ˆ„๋Š” ํ•จ์ˆ˜์ž…๋‹ˆ๋‹ค.
def get_text_chunks(documents):
    text_splitter = RecursiveCharacterTextSplitter(
        chunk_size=1000, # ์ฒญํฌ์˜ ํฌ๊ธฐ๋ฅผ ์ง€์ •ํ•ฉ๋‹ˆ๋‹ค.
        chunk_overlap=200, # ์ฒญํฌ ์‚ฌ์ด์˜ ์ค‘๋ณต์„ ์ง€์ •ํ•ฉ๋‹ˆ๋‹ค.
        length_function=len # ํ…์ŠคํŠธ์˜ ๊ธธ์ด๋ฅผ ์ธก์ •ํ•˜๋Š” ํ•จ์ˆ˜๋ฅผ ์ง€์ •ํ•ฉ๋‹ˆ๋‹ค.
    )

    documents = text_splitter.split_documents(documents) # ๋ฌธ์„œ๋“ค์„ ์ฒญํฌ๋กœ ๋‚˜๋ˆ•๋‹ˆ๋‹ค
    return documents # ๋‚˜๋ˆˆ ์ฒญํฌ๋ฅผ ๋ฐ˜ํ™˜ํ•ฉ๋‹ˆ๋‹ค.


# ํ…์ŠคํŠธ ์ฒญํฌ๋“ค๋กœ๋ถ€ํ„ฐ ๋ฒกํ„ฐ ์Šคํ† ์–ด๋ฅผ ์ƒ์„ฑํ•˜๋Š” ํ•จ์ˆ˜์ž…๋‹ˆ๋‹ค.
def get_vectorstore(text_chunks):
    # OpenAI ์ž„๋ฒ ๋”ฉ ๋ชจ๋ธ์„ ๋กœ๋“œํ•ฉ๋‹ˆ๋‹ค. (Embedding models - Ada v2)

    embeddings = OpenAIEmbeddings()
    vectorstore = FAISS.from_documents(text_chunks, embeddings) # FAISS ๋ฒกํ„ฐ ์Šคํ† ์–ด๋ฅผ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค.

    return vectorstore # ์ƒ์„ฑ๋œ ๋ฒกํ„ฐ ์Šคํ† ์–ด๋ฅผ ๋ฐ˜ํ™˜ํ•ฉ๋‹ˆ๋‹ค.


def get_conversation_chain(vectorstore):
    gpt_model_name = 'gpt-3.5-turbo'
    llm = ChatOpenAI(model_name = gpt_model_name) #gpt-3.5 ๋ชจ๋ธ ๋กœ๋“œ
    
    # ๋Œ€ํ™” ๊ธฐ๋ก์„ ์ €์žฅํ•˜๊ธฐ ์œ„ํ•œ ๋ฉ”๋ชจ๋ฆฌ๋ฅผ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค.
    memory = ConversationBufferMemory(
        memory_key='chat_history', return_messages=True)
    # ๋Œ€ํ™” ๊ฒ€์ƒ‰ ์ฒด์ธ์„ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค.
    conversation_chain = ConversationalRetrievalChain.from_llm(
        llm=llm,
        retriever=vectorstore.as_retriever(),
        memory=memory
    )
    return conversation_chain

# ์‚ฌ์šฉ์ž ์ž…๋ ฅ์„ ์ฒ˜๋ฆฌํ•˜๋Š” ํ•จ์ˆ˜์ž…๋‹ˆ๋‹ค.
def handle_userinput(user_question):
    # ๋Œ€ํ™” ์ฒด์ธ์„ ์‚ฌ์šฉํ•˜์—ฌ ์‚ฌ์šฉ์ž ์งˆ๋ฌธ์— ๋Œ€ํ•œ ์‘๋‹ต์„ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค.
    response = st.session_state.conversation({'question': user_question})
    # ๋Œ€ํ™” ๊ธฐ๋ก์„ ์ €์žฅํ•ฉ๋‹ˆ๋‹ค.
    st.session_state.chat_history = response['chat_history']

    for i, message in enumerate(st.session_state.chat_history):
        if i % 2 == 0:
            st.write(user_template.replace(
                "{{MSG}}", message.content), unsafe_allow_html=True)
        else:
            st.write(bot_template.replace(
                "{{MSG}}", message.content), unsafe_allow_html=True)


def main():
    load_dotenv()
    st.set_page_config(page_title="Chat with multiple Files",
                       page_icon=":books:")
    st.write(css, unsafe_allow_html=True)

    if "conversation" not in st.session_state:
        st.session_state.conversation = None
    if "chat_history" not in st.session_state:
        st.session_state.chat_history = None

    st.header("Chat with multiple Files :")
    user_question = st.text_input("Ask a question about your documents:")
    if user_question:
        handle_userinput(user_question)

    with st.sidebar:
        openai_key = st.text_input("Paste your OpenAI API key (sk-...)")
        if openai_key:
            os.environ["OPENAI_API_KEY"] = openai_key

        st.subheader("Your documents")
        docs = st.file_uploader(
            "Upload your PDFs here and click on 'Process'", accept_multiple_files=True)
        if st.button("Process"):
            with st.spinner("Processing"):
                # get pdf text
                doc_list = []

                for file in docs:
                    print('file - type : ', file.type)
                    if file.type == 'text/plain':
                        # file is .txt
                        doc_list.extend(get_text_file(file))
                    elif file.type in ['application/octet-stream', 'application/pdf']:
                        # file is .pdf
                        doc_list.extend(get_pdf_text(file))
                    elif file.type == 'text/csv':
                        # file is .csv
                        doc_list.extend(get_csv_file(file))
                    elif file.type == 'application/json':
                        # file is .json
                        doc_list.extend(get_json_file(file))

                # get the text chunks
                text_chunks = get_text_chunks(doc_list)

                # create vector store
                vectorstore = get_vectorstore(text_chunks)

                # create conversation chain
                st.session_state.conversation = get_conversation_chain(
                    vectorstore)


if __name__ == '__main__':
    main()