Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -8,19 +8,22 @@ import os
|
|
| 8 |
import subprocess
|
| 9 |
import io
|
| 10 |
import time
|
|
|
|
| 11 |
from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
|
| 12 |
from langchain.vectorstores import FAISS
|
| 13 |
from langchain.text_splitter import CharacterTextSplitter
|
| 14 |
from langchain.chains import ConversationalRetrievalChain
|
| 15 |
from langchain.memory import ConversationBufferMemory
|
| 16 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
# Ensure Playwright installs required browsers and dependencies
|
| 19 |
subprocess.run(["playwright", "install"])
|
| 20 |
nest_asyncio.apply()
|
| 21 |
|
| 22 |
-
|
| 23 |
-
|
| 24 |
GOOGLE_API_KEY = os.environ["GOOGLE_API_KEY"]
|
| 25 |
|
| 26 |
graph_config = {
|
|
@@ -31,7 +34,6 @@ graph_config = {
|
|
| 31 |
"max_results": 8,
|
| 32 |
"verbose": True,
|
| 33 |
"headless": True
|
| 34 |
-
|
| 35 |
}
|
| 36 |
|
| 37 |
def get_data(search_term):
|
|
@@ -60,7 +62,7 @@ def get_data(search_term):
|
|
| 60 |
)
|
| 61 |
result = search_graph.run()
|
| 62 |
if not result or not result.get("grants"):
|
| 63 |
-
st.error(f"No results returned for {search_term}. Please try again with a different
|
| 64 |
return {}
|
| 65 |
return result
|
| 66 |
except Exception as e:
|
|
@@ -82,8 +84,47 @@ def get_data(search_term):
|
|
| 82 |
st.error(f"Retry failed: {e2}. Please try again later.")
|
| 83 |
return {}
|
| 84 |
else:
|
| 85 |
-
st.error(f"An error occurred for search term
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 86 |
return {}
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
|
| 88 |
def process_multiple_search_terms(search_terms):
|
| 89 |
"""
|
|
@@ -182,27 +223,52 @@ def main():
|
|
| 182 |
if "chat_interface_active" not in st.session_state:
|
| 183 |
st.session_state.chat_interface_active = False
|
| 184 |
|
| 185 |
-
# Sidebar:
|
| 186 |
-
|
| 187 |
-
"
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
placeholder="e.g.,\nRenewable energy \nclimate change research\nAgriculture in Africa"
|
| 191 |
)
|
| 192 |
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 199 |
st.session_state.scraped_data = result
|
| 200 |
if result.get("grants"):
|
| 201 |
-
st.sidebar.success(f"✅ Found {len(result['grants'])} grant opportunities from
|
| 202 |
else:
|
| 203 |
-
st.sidebar.warning("⚠️ Please enter valid
|
| 204 |
-
else:
|
| 205 |
-
st.sidebar.warning("⚠️ Please enter at least one search term to begin.")
|
| 206 |
|
| 207 |
# Sidebar: Download & Share Controls
|
| 208 |
if st.session_state.scraped_data and st.session_state.scraped_data.get('grants'):
|
|
@@ -254,18 +320,15 @@ def main():
|
|
| 254 |
|
| 255 |
if st.session_state.chat_history:
|
| 256 |
st.subheader("Chat History")
|
| 257 |
-
# Reverse the chat history to show the latest messages first
|
| 258 |
for chat in reversed(st.session_state.chat_history):
|
| 259 |
-
# User message: dark grey background with white text
|
| 260 |
st.markdown(
|
| 261 |
f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 5px; background-color:#444444; color: white;'><strong>You:</strong> {chat['query']}</div>",
|
| 262 |
unsafe_allow_html=True)
|
| 263 |
-
# Bot message: blue background with white text
|
| 264 |
st.markdown(
|
| 265 |
f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 10px; background-color:#007BFF; color: white;'><strong>Grants Bot:</strong> {chat['response']}</div>",
|
| 266 |
unsafe_allow_html=True)
|
| 267 |
else:
|
| 268 |
-
st.info("⬅️ Enter search terms in the sidebar and click
|
| 269 |
|
| 270 |
st.sidebar.markdown("---")
|
| 271 |
st.sidebar.markdown(
|
|
@@ -278,4 +341,4 @@ def main():
|
|
| 278 |
)
|
| 279 |
|
| 280 |
if __name__ == "__main__":
|
| 281 |
-
main()
|
|
|
|
| 8 |
import subprocess
|
| 9 |
import io
|
| 10 |
import time
|
| 11 |
+
import urllib.parse
|
| 12 |
from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
|
| 13 |
from langchain.vectorstores import FAISS
|
| 14 |
from langchain.text_splitter import CharacterTextSplitter
|
| 15 |
from langchain.chains import ConversationalRetrievalChain
|
| 16 |
from langchain.memory import ConversationBufferMemory
|
| 17 |
+
|
| 18 |
+
# Import Supadata and initialize the client
|
| 19 |
+
from supadata import Supadata, SupadataError
|
| 20 |
+
SUPADATA_API_KEY = os.getenv("SUPADATA")
|
| 21 |
+
supadata = Supadata(api_key=SUPADATA_API_KEY)
|
| 22 |
|
| 23 |
# Ensure Playwright installs required browsers and dependencies
|
| 24 |
subprocess.run(["playwright", "install"])
|
| 25 |
nest_asyncio.apply()
|
| 26 |
|
|
|
|
|
|
|
| 27 |
GOOGLE_API_KEY = os.environ["GOOGLE_API_KEY"]
|
| 28 |
|
| 29 |
graph_config = {
|
|
|
|
| 34 |
"max_results": 8,
|
| 35 |
"verbose": True,
|
| 36 |
"headless": True
|
|
|
|
| 37 |
}
|
| 38 |
|
| 39 |
def get_data(search_term):
|
|
|
|
| 62 |
)
|
| 63 |
result = search_graph.run()
|
| 64 |
if not result or not result.get("grants"):
|
| 65 |
+
st.error(f"No results returned for {search_term}. Please try again with a different search term.")
|
| 66 |
return {}
|
| 67 |
return result
|
| 68 |
except Exception as e:
|
|
|
|
| 84 |
st.error(f"Retry failed: {e2}. Please try again later.")
|
| 85 |
return {}
|
| 86 |
else:
|
| 87 |
+
st.error(f"An error occurred for search term: {search_term}, error: {e}. Please try again.")
|
| 88 |
+
return {}
|
| 89 |
+
|
| 90 |
+
def get_data_from_url(url):
|
| 91 |
+
"""
|
| 92 |
+
Scrape the provided URL using Supadata and pass the page content directly to the Gemini model
|
| 93 |
+
(using ChatGoogleGenerativeAI) to extract grant data in a JSON structure.
|
| 94 |
+
"""
|
| 95 |
+
try:
|
| 96 |
+
web_content = supadata.web.scrape(url)
|
| 97 |
+
page_content = web_content.content
|
| 98 |
+
full_prompt = (
|
| 99 |
+
"Extract the following grant data from the provided web content. "
|
| 100 |
+
"List me all grants or funds with:\n"
|
| 101 |
+
"- Grant name/title\n"
|
| 102 |
+
"- Short summary\n"
|
| 103 |
+
"- Funding organization\n"
|
| 104 |
+
"- Grant value (numeric only)\n"
|
| 105 |
+
"- Application deadline\n"
|
| 106 |
+
"- Eligible countries\n"
|
| 107 |
+
"- Sector/field\n"
|
| 108 |
+
"- Eligibility criteria\n"
|
| 109 |
+
"Return in JSON format.\n\n"
|
| 110 |
+
f"Web content: {page_content}"
|
| 111 |
+
)
|
| 112 |
+
llm = ChatGoogleGenerativeAI(
|
| 113 |
+
model="gemini-2.0-flash-thinking-exp", google_api_key=GOOGLE_API_KEY, temperature=0
|
| 114 |
+
)
|
| 115 |
+
response = llm.invoke(full_prompt)
|
| 116 |
+
try:
|
| 117 |
+
result = json.loads(response.content)
|
| 118 |
+
except Exception as parse_error:
|
| 119 |
+
st.error("Error parsing JSON from Gemini model response.")
|
| 120 |
+
return {}
|
| 121 |
+
if not result or not result.get("grants"):
|
| 122 |
+
st.error("No grant opportunities found in the scraped URL.")
|
| 123 |
return {}
|
| 124 |
+
return result
|
| 125 |
+
except Exception as e:
|
| 126 |
+
st.error(f"An error occurred while scraping URL {url}: {e}")
|
| 127 |
+
return {}
|
| 128 |
|
| 129 |
def process_multiple_search_terms(search_terms):
|
| 130 |
"""
|
|
|
|
| 223 |
if "chat_interface_active" not in st.session_state:
|
| 224 |
st.session_state.chat_interface_active = False
|
| 225 |
|
| 226 |
+
# Sidebar: Input Type Selection
|
| 227 |
+
input_type = st.sidebar.radio(
|
| 228 |
+
"Select Input Type:",
|
| 229 |
+
("Search Query", "URL"),
|
| 230 |
+
key="input_type_selector"
|
|
|
|
| 231 |
)
|
| 232 |
|
| 233 |
+
# Sidebar: Input field based on selection
|
| 234 |
+
if input_type == "Search Query":
|
| 235 |
+
search_input = st.sidebar.text_area(
|
| 236 |
+
"Enter Search Terms (one per line). Maximum 2",
|
| 237 |
+
height=150,
|
| 238 |
+
help="Input search terms to discover grant opportunities. Terms can be specific or generic.",
|
| 239 |
+
placeholder="e.g.,\nRenewable energy \nclimate change research\nAgriculture in Africa"
|
| 240 |
+
)
|
| 241 |
+
else:
|
| 242 |
+
url_input = st.sidebar.text_input(
|
| 243 |
+
"Enter URL to scrape for grant opportunities",
|
| 244 |
+
placeholder="https://example.com/grants"
|
| 245 |
+
)
|
| 246 |
+
|
| 247 |
+
# Execute based on input type selection
|
| 248 |
+
if input_type == "Search Query":
|
| 249 |
+
if st.sidebar.button("🔍 Get Grant Opportunities"):
|
| 250 |
+
if search_input:
|
| 251 |
+
search_terms = [term.strip() for term in search_input.split("\n") if term.strip()]
|
| 252 |
+
if search_terms:
|
| 253 |
+
with st.spinner("Searching in progress... Please wait patiently."):
|
| 254 |
+
result = process_multiple_search_terms(search_terms)
|
| 255 |
+
st.session_state.scraped_data = result
|
| 256 |
+
if result.get("grants"):
|
| 257 |
+
st.sidebar.success(f"✅ Found {len(result['grants'])} grant opportunities from {len(search_terms)} search terms!")
|
| 258 |
+
else:
|
| 259 |
+
st.sidebar.warning("⚠️ Please enter valid search terms.")
|
| 260 |
+
else:
|
| 261 |
+
st.sidebar.warning("⚠️ Please enter at least one search term to begin.")
|
| 262 |
+
else: # URL input
|
| 263 |
+
if st.sidebar.button("🔍 Scrape URL for Grant Opportunities"):
|
| 264 |
+
if url_input:
|
| 265 |
+
with st.spinner("Scraping URL... Please wait patiently."):
|
| 266 |
+
result = get_data_from_url(url_input)
|
| 267 |
st.session_state.scraped_data = result
|
| 268 |
if result.get("grants"):
|
| 269 |
+
st.sidebar.success(f"✅ Found {len(result['grants'])} grant opportunities from the URL!")
|
| 270 |
else:
|
| 271 |
+
st.sidebar.warning("⚠️ Please enter a valid URL to scrape.")
|
|
|
|
|
|
|
| 272 |
|
| 273 |
# Sidebar: Download & Share Controls
|
| 274 |
if st.session_state.scraped_data and st.session_state.scraped_data.get('grants'):
|
|
|
|
| 320 |
|
| 321 |
if st.session_state.chat_history:
|
| 322 |
st.subheader("Chat History")
|
|
|
|
| 323 |
for chat in reversed(st.session_state.chat_history):
|
|
|
|
| 324 |
st.markdown(
|
| 325 |
f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 5px; background-color:#444444; color: white;'><strong>You:</strong> {chat['query']}</div>",
|
| 326 |
unsafe_allow_html=True)
|
|
|
|
| 327 |
st.markdown(
|
| 328 |
f"<div style='padding: 10px; border-radius: 5px; margin-bottom: 10px; background-color:#007BFF; color: white;'><strong>Grants Bot:</strong> {chat['response']}</div>",
|
| 329 |
unsafe_allow_html=True)
|
| 330 |
else:
|
| 331 |
+
st.info("⬅️ Enter search terms or a URL in the sidebar and click the appropriate button to start searching.")
|
| 332 |
|
| 333 |
st.sidebar.markdown("---")
|
| 334 |
st.sidebar.markdown(
|
|
|
|
| 341 |
)
|
| 342 |
|
| 343 |
if __name__ == "__main__":
|
| 344 |
+
main()
|