File size: 4,843 Bytes
7d849d3
413cb20
9e4c9f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7d849d3
9830b8e
 
 
e4b3526
4b4013f
b16e227
81b851a
9830b8e
b16e227
7888633
 
 
 
 
 
 
 
874fc5b
63e083c
874fc5b
7888633
874fc5b
 
 
 
 
 
 
 
 
 
 
6a1c9b8
63e083c
e4b3526
4b4013f
413cb20
6cbfbad
e4b3526
4b4013f
6cbfbad
413cb20
e4b3526
674ea12
 
ef1eb58
 
 
 
 
 
 
 
 
4b4013f
005a493
7888633
413cb20
7888633
e4b3526
7888633
e4b3526
7888633
413cb20
7888633
8060e77
874fc5b
674ea12
874fc5b
8060e77
7888633
 
 
 
 
 
 
 
 
674ea12
 
 
 
 
7888633
ef1eb58
 
 
7888633
94f915f
7888633
94f915f
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
import streamlit as st
from tempfile import NamedTemporaryFile

import pprint
import google.generativeai as palm
import os
from dotenv import load_dotenv, find_dotenv
from langchain.embeddings import GooglePalmEmbeddings
from langchain.llms import GooglePalm

from langchain.document_loaders import UnstructuredURLLoader  #load urls into docoument-loader
from langchain.chains.question_answering import load_qa_chain
from langchain.indexes import VectorstoreIndexCreator #vectorize db index with chromadb
from langchain.text_splitter import CharacterTextSplitter #text splitter
from langchain.chains import RetrievalQA
from langchain.document_loaders import UnstructuredPDFLoader  #load pdf
from langchain.agents import create_pandas_dataframe_agent

import pandas as pd
import numpy as np
import pprint

radioButtonList = ["E-commerce CSV (https://www.kaggle.com/datasets/mervemenekse/ecommerce-dataset)",
"Upload my own CSV",
"Upload my own PDF",
"URL Chat with Google Alphabet's 2022 Q2 Earnings Report (https://tinyurl.com/f85wujsj)",
"Enter my own URL"]
genre = st.radio(
    "Choose dataset to finetune", radioButtonList, index=0
    )

# Initialize language model
load_dotenv(find_dotenv()) # read local .env file
api_key = st.secrets["PALM_API_KEY"] # put your API key here
os.environ["GOOGLE_API_KEY"] = st.secrets["PALM_API_KEY"]
palm.configure(api_key=api_key)
llm = GooglePalm()
llm.temperature = 0.1

pdfCSVURLText = ""
if genre==radioButtonList[0]:
    pdfCSVURLText = "CSV"
    dataDF = pd.read_csv('EcommerceDataset.csv', encoding= 'unicode_escape')
    # st.write('You selected comedy.')
    # else:
    # st.write(f'''Password streamlit app: {st.secrets["PSWD"]}''')
elif genre==radioButtonList[1]:
    pdfCSVURLText = "CSV"
elif genre==radioButtonList[2]:
    pdfCSVURLText = "PDF"
elif genre==radioButtonList[3]:
    pdfCSVURLText = "URL"
elif genre==radioButtonList[4]:
    pdfCSVURLText = "URL"

isCustomURL = genre==radioButtonList[4]
urlInput = st.text_input('Enter your own URL', '', placeholder="Type your URL here (e.g. https://tinyurl.com/f85wujsj)", disabled=not isCustomURL)

isCustomPDF = genre==radioButtonList[1] or genre==radioButtonList[2]
uploaded_file = st.file_uploader(f"Upload your own {pdfCSVURLText} here", type=pdfCSVURLText.lower(), disabled=not isCustomPDF)
uploadedFilename = ""
if uploaded_file is not None:
    with NamedTemporaryFile(dir='.', suffix=f'.{pdfCSVURLText.lower()}') as f:
        f.write(uploaded_file.getbuffer())
        uploadedFilename = f.name
        if genre==radioButtonList[1]: # Custom CSV Upload
            dataDF = pd.read_csv(uploadedFilename, encoding= 'unicode_escape')
        elif genre==radioButtonList[2]: # Custom PDF Upload
            pdf_loaders = [UnstructuredPDFLoader(uploadedFilename)]
            pdf_index = VectorstoreIndexCreator(
                    embedding=GooglePalmEmbeddings(),
                    text_splitter=CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)).from_loaders(pdf_loaders)
            pdf_chain = RetrievalQA.from_chain_type(llm=llm,
                                        chain_type="stuff",
                                        retriever=pdf_index.vectorstore.as_retriever(),
                                        input_key="question")

enableChatBox = False
if genre==radioButtonList[0]: # E-commerce CSV
    enableChatBox = True
elif genre==radioButtonList[1]: # Custom CSV Upload
    enableChatBox = uploadedFilename[-4:]==".csv"
elif genre==radioButtonList[2]: # Custom PDF Upload
    enableChatBox = uploadedFilename[-4:]==".pdf"
elif genre==radioButtonList[3]: # Google Alphabet URL Earnings Report
    enableChatBox = True
elif genre==radioButtonList[4]: # Custom URL
    enableChatBox = True

chatTextStr = st.text_input(f'Ask me anything about this {pdfCSVURLText}', '', placeholder="Type here (e.g. Question1: What was the most sold item? Question2: What was the most common payment?)", disabled=not enableChatBox)
chatWithPDFButton = "CLICK HERE TO START CHATTING"
if st.button(chatWithPDFButton, disabled=not enableChatBox and not chatTextStr): #  Button Cliked


    if genre==radioButtonList[0]: # E-commerce CSV
        # Initializing the agent
        agent = create_pandas_dataframe_agent(llm, dataDF, verbose=False)
        answer = agent.run(chatTextStr)
        st.write(answer)

    elif genre==radioButtonList[1]: # Custom CSV Upload
        # Initializing the agent
        agent = create_pandas_dataframe_agent(llm, dataDF, verbose=False)
        answer = agent.run(chatTextStr)
        st.write(answer)

    elif genre==radioButtonList[2]: # Custom PDF Upload
        pdf_answer = pdf_chain.run(chatTextStr)
        st.write(pdf_answer)

    elif genre==radioButtonList[3]: # Google Alphabet URL Earnings Report
        pass
    elif genre==radioButtonList[4]: # Custom URL
        pass