File size: 3,460 Bytes
9177215
 
7d1720e
9177215
7d1720e
9177215
 
 
 
 
 
 
7d1720e
9177215
7d1720e
 
 
 
 
9177215
 
 
 
 
 
 
 
 
f1387c9
 
9177215
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8598421
 
9177215
 
 
 
 
 
 
 
 
 
 
8598421
9177215
 
 
8598421
f1387c9
 
8598421
 
 
9177215
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8598421
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
import streamlit as st

from components.sidebar import sidebar

from ui import (
    wrap_doc_in_html,
    is_query_valid,
    is_file_valid,
    is_open_ai_key_valid,
    display_file_read_error,
)

from core.caching import bootstrap_caching

from core.parsing import read_file
from core.chunking import chunk_file
from core.embedding import embed_files
from core.qa import query_folder
from core.utils import get_llm


EMBEDDING = "openai"
VECTOR_STORE = "faiss"
MODEL_LIST = ["gpt-3.5-turbo", "gpt-4"]

# Uncomment to enable debug mode
# MODEL_LIST.insert(0, "debug")

st.set_page_config(page_title="☘️InspectSolv", page_icon="☘️", layout="wide")
st.header("☘️InspectSolv")

# Enable caching for expensive functions
bootstrap_caching()

sidebar()

openai_api_key = st.session_state.get("OPENAI_API_KEY")


if not openai_api_key:
    st.warning(
        "Enter your OpenAI API key in the sidebar. You can get a key at"
        " https://platform.openai.com/account/api-keys."
    )




uploaded_file = st.file_uploader(
    "Upload a pdf, docx, or txt file",
    type=["pdf", "docx", "txt"],
    help="Scanned documents are not supported yet!",
)

# model: str = st.selectbox("Model", options=MODEL_LIST)  # type: ignore

model = MODEL_LIST[0]


from importlib import metadata
with st.expander("Advanced Options"):
    return_all_chunks = st.checkbox("Show all chunks retrieved from vector search")
    show_full_doc = st.checkbox("Show parsed contents of the document")
    st.markdown('# config')
    import sys
    st.markdown(f'Python version: {sys.version}')
    for dist in metadata.distributions():
        st.markdown((f"{dist.name}=={dist.version}"))



if not uploaded_file:
    st.stop()

try:
    file = read_file(uploaded_file)
except Exception as e:
    display_file_read_error(e, file_name=uploaded_file.name)

chunked_file = chunk_file(file, chunk_size=300, chunk_overlap=0)

if not is_file_valid(file):
    st.stop()


if not is_open_ai_key_valid(openai_api_key, model):
    st.stop()


with st.spinner("Indexing document... This may take a while⏳"):
    folder_index = embed_files(
        files=[chunked_file],
        embedding=EMBEDDING if model != "debug" else "debug",
        vector_store=VECTOR_STORE if model != "debug" else "debug",
        openai_api_key=openai_api_key,
    )


with st.form(key="qa_form"):
    options = ['List all pre existing conditions which may affect home insurance', 'Show the problematic components!', 'Show repair needs!']
    query = st.selectbox('Select an option', options)
    submit = st.form_submit_button("Submit")


if show_full_doc:
    with st.expander("Document"):
        # Hack to get around st.markdown rendering LaTeX
        st.markdown(f"<p>{wrap_doc_in_html(file.docs)}</p>", unsafe_allow_html=True)


if submit:
    if not is_query_valid(query):
        st.stop()

    # Output Columns
    answer_col, sources_col = st.columns(2)

    llm = get_llm(model=model, openai_api_key=openai_api_key, temperature=0)
    result = query_folder(
        folder_index=folder_index,
        query=query,
        return_all=return_all_chunks,
        llm=llm,
    )

    with answer_col:
        st.markdown("#### Answer")
        st.markdown(result.answer)

    with sources_col:
        st.markdown("#### Sources")
        for source in result.sources:
            st.markdown(source.page_content)
            st.markdown(source.metadata["source"])
            st.markdown("---")