Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,61 +1,65 @@
|
|
| 1 |
-
import streamlit as st
|
| 2 |
import os
|
|
|
|
| 3 |
import tempfile
|
| 4 |
from pymongo import MongoClient
|
| 5 |
from datetime import datetime
|
| 6 |
from pathlib import Path
|
| 7 |
from document_chunker import DocumentChunker
|
| 8 |
-
from dotenv import load_dotenv
|
| 9 |
-
|
| 10 |
-
load_dotenv()
|
| 11 |
|
| 12 |
-
# MongoDB connection
|
| 13 |
-
mongo_uri = os.
|
| 14 |
-
db_name = os.
|
| 15 |
client = MongoClient(mongo_uri)
|
|
|
|
| 16 |
|
|
|
|
| 17 |
st.set_page_config(page_title="Doc Chunker", layout="wide")
|
| 18 |
st.title("π Document Chunker & Uploader")
|
| 19 |
|
| 20 |
with st.sidebar:
|
| 21 |
st.header("Settings")
|
| 22 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
is_grant_app = st.toggle("Is this a Grant Application?", value=True)
|
| 24 |
-
if st.button("Connect to Collection"):
|
| 25 |
-
collection = client[db_name][selected_collection]
|
| 26 |
-
st.success(f"Connected to `{selected_collection}` in `{db_name}`")
|
| 27 |
|
| 28 |
uploaded_file = st.file_uploader("Upload a DOCX or TXT file", type=["docx", "txt"])
|
| 29 |
|
| 30 |
if uploaded_file:
|
| 31 |
-
# Save file to temp path
|
| 32 |
temp_path = Path(tempfile.gettempdir()) / uploaded_file.name
|
| 33 |
with open(temp_path, "wb") as f:
|
| 34 |
f.write(uploaded_file.getbuffer())
|
| 35 |
|
| 36 |
st.success(f"Uploaded `{uploaded_file.name}`")
|
| 37 |
|
| 38 |
-
# Check if file already exists in collection
|
| 39 |
modified_time = datetime.now().isoformat()
|
| 40 |
-
collection =
|
|
|
|
| 41 |
if collection.find_one({"metadata.title": uploaded_file.name}):
|
| 42 |
st.warning("β οΈ This file already exists in the collection. Skipping...")
|
| 43 |
else:
|
| 44 |
-
st.write("β³ Processing...")
|
| 45 |
chunker = DocumentChunker()
|
| 46 |
-
chunks = chunker.process_document(str(temp_path))
|
| 47 |
|
| 48 |
if chunks:
|
| 49 |
for chunk in chunks:
|
| 50 |
chunk['metadata'].update({
|
| 51 |
"title": uploaded_file.name,
|
| 52 |
"uploaded_at": modified_time,
|
|
|
|
| 53 |
})
|
| 54 |
collection.insert_one(chunk)
|
| 55 |
|
| 56 |
st.success(f"β
{len(chunks)} chunks inserted into `{selected_collection}`")
|
| 57 |
|
| 58 |
-
# Show
|
| 59 |
for i, c in enumerate(chunks[:3]):
|
| 60 |
st.subheader(f"Chunk {i+1}: {c['metadata'].get('header') or 'No Header'}")
|
| 61 |
st.markdown(c['text'][:400] + "...")
|
|
|
|
|
|
|
| 1 |
import os
|
| 2 |
+
import streamlit as st
|
| 3 |
import tempfile
|
| 4 |
from pymongo import MongoClient
|
| 5 |
from datetime import datetime
|
| 6 |
from pathlib import Path
|
| 7 |
from document_chunker import DocumentChunker
|
|
|
|
|
|
|
|
|
|
| 8 |
|
| 9 |
+
# === MongoDB connection via Hugging Face secrets ===
|
| 10 |
+
mongo_uri = os.environ["MONGO_URI"]
|
| 11 |
+
db_name = os.environ.get("MONGO_DB", "grant_docs")
|
| 12 |
client = MongoClient(mongo_uri)
|
| 13 |
+
db = client[db_name]
|
| 14 |
|
| 15 |
+
# === Streamlit UI ===
|
| 16 |
st.set_page_config(page_title="Doc Chunker", layout="wide")
|
| 17 |
st.title("π Document Chunker & Uploader")
|
| 18 |
|
| 19 |
with st.sidebar:
|
| 20 |
st.header("Settings")
|
| 21 |
+
|
| 22 |
+
# Fetch collection names for dropdown
|
| 23 |
+
try:
|
| 24 |
+
existing_collections = db.list_collection_names()
|
| 25 |
+
selected_collection = st.selectbox("Choose MongoDB Collection", existing_collections, index=existing_collections.index("doc_chunks_cat") if "doc_chunks_cat" in existing_collections else 0)
|
| 26 |
+
except Exception as e:
|
| 27 |
+
st.error(f"Failed to list collections: {e}")
|
| 28 |
+
selected_collection = "doc_chunks_cat"
|
| 29 |
+
|
| 30 |
is_grant_app = st.toggle("Is this a Grant Application?", value=True)
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
uploaded_file = st.file_uploader("Upload a DOCX or TXT file", type=["docx", "txt"])
|
| 33 |
|
| 34 |
if uploaded_file:
|
|
|
|
| 35 |
temp_path = Path(tempfile.gettempdir()) / uploaded_file.name
|
| 36 |
with open(temp_path, "wb") as f:
|
| 37 |
f.write(uploaded_file.getbuffer())
|
| 38 |
|
| 39 |
st.success(f"Uploaded `{uploaded_file.name}`")
|
| 40 |
|
|
|
|
| 41 |
modified_time = datetime.now().isoformat()
|
| 42 |
+
collection = db[selected_collection]
|
| 43 |
+
|
| 44 |
if collection.find_one({"metadata.title": uploaded_file.name}):
|
| 45 |
st.warning("β οΈ This file already exists in the collection. Skipping...")
|
| 46 |
else:
|
| 47 |
+
st.write("β³ Processing with DocumentChunker...")
|
| 48 |
chunker = DocumentChunker()
|
| 49 |
+
chunks = chunker.process_document(str(temp_path))
|
| 50 |
|
| 51 |
if chunks:
|
| 52 |
for chunk in chunks:
|
| 53 |
chunk['metadata'].update({
|
| 54 |
"title": uploaded_file.name,
|
| 55 |
"uploaded_at": modified_time,
|
| 56 |
+
"is_grant_app": is_grant_app,
|
| 57 |
})
|
| 58 |
collection.insert_one(chunk)
|
| 59 |
|
| 60 |
st.success(f"β
{len(chunks)} chunks inserted into `{selected_collection}`")
|
| 61 |
|
| 62 |
+
# Show a few previews
|
| 63 |
for i, c in enumerate(chunks[:3]):
|
| 64 |
st.subheader(f"Chunk {i+1}: {c['metadata'].get('header') or 'No Header'}")
|
| 65 |
st.markdown(c['text'][:400] + "...")
|