Spaces:
Sleeping
Sleeping
Commit
·
6b63bdc
1
Parent(s):
aac6922
add collumns
Browse files
app_V2.py
CHANGED
|
@@ -1,5 +1,4 @@
|
|
| 1 |
import tempfile
|
| 2 |
-
import time
|
| 3 |
import streamlit as st
|
| 4 |
from PyPDF2 import PdfReader
|
| 5 |
from langchain.text_splitter import CharacterTextSplitter
|
|
@@ -11,9 +10,9 @@ from langchain.chains import ConversationalRetrievalChain
|
|
| 11 |
import os
|
| 12 |
import pickle
|
| 13 |
from datetime import datetime
|
| 14 |
-
from backend.generate_metadata import generate_metadata, ingest
|
| 15 |
-
|
| 16 |
|
|
|
|
| 17 |
css = '''
|
| 18 |
<style>
|
| 19 |
.chat-message {
|
|
@@ -58,46 +57,6 @@ user_template = '''
|
|
| 58 |
</div>
|
| 59 |
'''
|
| 60 |
|
| 61 |
-
def main():
|
| 62 |
-
|
| 63 |
-
st.set_page_config(page_title="Doc Verify RAG", page_icon=":mag:")
|
| 64 |
-
st.write('Anomaly detection for document metadata', unsafe_allow_html=True)
|
| 65 |
-
st.header("Doc Verify RAG :mag:")
|
| 66 |
-
|
| 67 |
-
if "openai_api_key" not in st.session_state:
|
| 68 |
-
st.session_state.openai_api_key = False
|
| 69 |
-
if "openai_org" not in st.session_state:
|
| 70 |
-
st.session_state.openai_org = False
|
| 71 |
-
if "classify" not in st.session_state:
|
| 72 |
-
st.session_state.classify = False
|
| 73 |
-
|
| 74 |
-
col1, col2 = st.columns(2)
|
| 75 |
-
with col1:
|
| 76 |
-
uploaded_file = st.file_uploader("Choose a PDF file", type=["pdf", "txt"])
|
| 77 |
-
|
| 78 |
-
if uploaded_file is not None:
|
| 79 |
-
try:
|
| 80 |
-
with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(uploaded_file.name)[1]) as tmp:
|
| 81 |
-
tmp.write(uploaded_file.read())
|
| 82 |
-
file_path = tmp.name
|
| 83 |
-
st.write(f'Created temporary file {file_path}')
|
| 84 |
-
|
| 85 |
-
docs = ingest(file_path)
|
| 86 |
-
st.write('## Querying Together.ai API')
|
| 87 |
-
metadata = generate_metadata(docs)
|
| 88 |
-
st.write(f'## Metadata Generated by {MODEL_NAME}')
|
| 89 |
-
st.write(metadata)
|
| 90 |
-
|
| 91 |
-
# Clean up the temporary file
|
| 92 |
-
os.remove(file_path)
|
| 93 |
-
|
| 94 |
-
except Exception as e:
|
| 95 |
-
st.error(f'Error: {e}')
|
| 96 |
-
with col2:
|
| 97 |
-
if st.button("Abbruch MFH Holzweg 13"):
|
| 98 |
-
st.session_state.user_space = "deconstruction"
|
| 99 |
-
|
| 100 |
-
|
| 101 |
|
| 102 |
def get_pdf_text(pdf_docs):
|
| 103 |
text = ""
|
|
@@ -166,13 +125,11 @@ def safe_vec_store():
|
|
| 166 |
pickle.dump(vector_store, f)
|
| 167 |
|
| 168 |
|
|
|
|
| 169 |
def main():
|
| 170 |
|
| 171 |
|
| 172 |
|
| 173 |
-
def set_pw():
|
| 174 |
-
st.session_state.openai_api_key = True
|
| 175 |
-
|
| 176 |
st.subheader("Your documents")
|
| 177 |
|
| 178 |
if st.session_state.classify:
|
|
@@ -239,6 +196,51 @@ def main():
|
|
| 239 |
|
| 240 |
if st.button("Load Embeddings"):
|
| 241 |
st.warning("this function is not in use, just upload the vectorstore")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 242 |
|
| 243 |
|
| 244 |
if __name__ == '__main__':
|
|
|
|
| 1 |
import tempfile
|
|
|
|
| 2 |
import streamlit as st
|
| 3 |
from PyPDF2 import PdfReader
|
| 4 |
from langchain.text_splitter import CharacterTextSplitter
|
|
|
|
| 10 |
import os
|
| 11 |
import pickle
|
| 12 |
from datetime import datetime
|
| 13 |
+
from backend.generate_metadata import generate_metadata, ingest
|
|
|
|
| 14 |
|
| 15 |
+
MODEL_NAME = "mixtral"
|
| 16 |
css = '''
|
| 17 |
<style>
|
| 18 |
.chat-message {
|
|
|
|
| 57 |
</div>
|
| 58 |
'''
|
| 59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
def get_pdf_text(pdf_docs):
|
| 62 |
text = ""
|
|
|
|
| 125 |
pickle.dump(vector_store, f)
|
| 126 |
|
| 127 |
|
| 128 |
+
"""
|
| 129 |
def main():
|
| 130 |
|
| 131 |
|
| 132 |
|
|
|
|
|
|
|
|
|
|
| 133 |
st.subheader("Your documents")
|
| 134 |
|
| 135 |
if st.session_state.classify:
|
|
|
|
| 196 |
|
| 197 |
if st.button("Load Embeddings"):
|
| 198 |
st.warning("this function is not in use, just upload the vectorstore")
|
| 199 |
+
"""
|
| 200 |
+
|
| 201 |
+
|
| 202 |
+
def main():
|
| 203 |
+
|
| 204 |
+
st.set_page_config(page_title="Doc Verify RAG", page_icon=":mag:")
|
| 205 |
+
st.write('Anomaly detection for document metadata', unsafe_allow_html=True)
|
| 206 |
+
st.header("Doc Verify RAG :mag:")
|
| 207 |
+
|
| 208 |
+
def set_pw():
|
| 209 |
+
st.session_state.openai_api_key = True
|
| 210 |
+
|
| 211 |
+
if "openai_api_key" not in st.session_state:
|
| 212 |
+
st.session_state.openai_api_key = False
|
| 213 |
+
if "openai_org" not in st.session_state:
|
| 214 |
+
st.session_state.openai_org = False
|
| 215 |
+
if "classify" not in st.session_state:
|
| 216 |
+
st.session_state.classify = False
|
| 217 |
+
|
| 218 |
+
col1, col2 = st.columns(2)
|
| 219 |
+
with col1:
|
| 220 |
+
uploaded_file = st.file_uploader("Choose a PDF file", type=["pdf", "txt"])
|
| 221 |
+
|
| 222 |
+
if uploaded_file is not None:
|
| 223 |
+
try:
|
| 224 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(uploaded_file.name)[1]) as tmp:
|
| 225 |
+
tmp.write(uploaded_file.read())
|
| 226 |
+
file_path = tmp.name
|
| 227 |
+
st.write(f'Created temporary file {file_path}')
|
| 228 |
+
|
| 229 |
+
docs = ingest(file_path)
|
| 230 |
+
st.write('## Querying Together.ai API')
|
| 231 |
+
metadata = generate_metadata(docs)
|
| 232 |
+
st.write(f'## Metadata Generated by {MODEL_NAME}')
|
| 233 |
+
st.write(metadata)
|
| 234 |
+
|
| 235 |
+
# Clean up the temporary file
|
| 236 |
+
os.remove(file_path)
|
| 237 |
+
|
| 238 |
+
except Exception as e:
|
| 239 |
+
st.error(f'Error: {e}')
|
| 240 |
+
with col2:
|
| 241 |
+
OPENAI_API_KEY = st.text_input("OPENAI API KEY:", type="password",
|
| 242 |
+
disabled=st.session_state.openai_api_key, on_change=set_pw)
|
| 243 |
+
classification = st.file_uploader("upload the metadata", type=["csv", "txt"])
|
| 244 |
|
| 245 |
|
| 246 |
if __name__ == '__main__':
|