Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -6,6 +6,9 @@ import os
|
|
6 |
import logging
|
7 |
from typing import Dict, List
|
8 |
from datetime import datetime
|
|
|
|
|
|
|
9 |
|
10 |
# Set up logging
|
11 |
logging.basicConfig(
|
@@ -18,11 +21,51 @@ class RaindropSearchBot:
|
|
18 |
def __init__(self):
|
19 |
self.openai_api_key = os.getenv('openaikey')
|
20 |
self.raindrop_api_token = os.getenv('raindroptoken')
|
21 |
-
|
|
|
|
|
|
|
22 |
raise EnvironmentError(
|
23 |
-
"Missing required environment variables. Please ensure
|
24 |
)
|
|
|
25 |
self.client = OpenAI(api_key=self.openai_api_key)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
|
27 |
def generate_search_query(self, user_request: str) -> str:
|
28 |
"""Convert user request to optimized search terms."""
|
@@ -105,31 +148,53 @@ class RaindropSearchBot:
|
|
105 |
logger.error(f"Search error: {e}")
|
106 |
return []
|
107 |
|
108 |
-
def analyze_results(self, results: List[Dict],
|
109 |
-
|
110 |
-
|
|
|
111 |
return "No relevant results found. Try modifying your search terms."
|
112 |
|
113 |
# Create context for analysis
|
114 |
context = f"Based on the search query: '{user_query}'\n\n"
|
115 |
-
context += "Analyze these
|
116 |
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
122 |
|
123 |
try:
|
124 |
prompt = f"""
|
125 |
-
Based on
|
126 |
|
127 |
Requirements:
|
128 |
-
1. Focus
|
129 |
-
2. Organize the response by key themes
|
130 |
3. Include only factual information from the sources
|
131 |
4. Highlight any significant developments or changes
|
132 |
-
5. Note any
|
|
|
133 |
|
134 |
Context:
|
135 |
{context}
|
@@ -148,28 +213,53 @@ class RaindropSearchBot:
|
|
148 |
logger.error(f"Analysis generation error: {e}")
|
149 |
return "Error generating analysis."
|
150 |
|
151 |
-
def format_results(self, results: List[Dict],
|
152 |
-
|
153 |
-
|
154 |
-
return "No results found. Try modifying your search terms."
|
155 |
-
|
156 |
output = f"{analysis}\n\n"
|
157 |
output += "-------\n\n"
|
158 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
159 |
|
160 |
-
|
161 |
-
|
162 |
-
|
|
|
163 |
output += f"{idx}. {item.get('title', 'No Title')}\n"
|
164 |
if item.get('link'):
|
165 |
output += f" Link: {item['link']}\n"
|
166 |
-
if item.get('
|
167 |
-
output += f"
|
168 |
-
|
169 |
-
|
170 |
-
|
171 |
-
|
172 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
173 |
output += "\n"
|
174 |
|
175 |
return output
|
@@ -183,15 +273,20 @@ class RaindropSearchBot:
|
|
183 |
search_query = self.generate_search_query(user_request)
|
184 |
logger.info(f"Using search query: {search_query}")
|
185 |
|
186 |
-
#
|
187 |
-
|
188 |
-
|
|
|
|
|
|
|
|
|
|
|
189 |
|
190 |
# Generate analysis
|
191 |
-
analysis = self.analyze_results(
|
192 |
|
193 |
# Format and return results
|
194 |
-
return self.format_results(
|
195 |
|
196 |
except Exception as e:
|
197 |
logger.error(f"Error processing request: {e}", exc_info=True)
|
@@ -205,10 +300,13 @@ def chatbot_interface(user_input: str) -> str:
|
|
205 |
return bot.process_request(user_input)
|
206 |
|
207 |
# Define and launch the interface
|
208 |
-
with gr.Blocks(title="
|
209 |
gr.Markdown("""
|
210 |
-
# π
|
211 |
-
Enter your search request in natural language, and I'll find and analyze
|
|
|
|
|
|
|
212 |
""")
|
213 |
|
214 |
with gr.Row():
|
@@ -224,7 +322,7 @@ with gr.Blocks(title="Raindrop.io Search Assistant", theme=gr.themes.Soft()) as
|
|
224 |
with gr.Row():
|
225 |
output_text = gr.Textbox(
|
226 |
label="Analysis and Results",
|
227 |
-
lines=
|
228 |
interactive=False
|
229 |
)
|
230 |
|
|
|
6 |
import logging
|
7 |
from typing import Dict, List
|
8 |
from datetime import datetime
|
9 |
+
from bs4 import BeautifulSoup
|
10 |
+
from serpapi import GoogleSearch
|
11 |
+
from newsapi import NewsApiClient
|
12 |
|
13 |
# Set up logging
|
14 |
logging.basicConfig(
|
|
|
21 |
def __init__(self):
|
22 |
self.openai_api_key = os.getenv('openaikey')
|
23 |
self.raindrop_api_token = os.getenv('raindroptoken')
|
24 |
+
self.serpapi_key = os.getenv('serpapikey')
|
25 |
+
self.newsapi_key = os.getenv('newsapikey')
|
26 |
+
|
27 |
+
if not all([self.openai_api_key, self.raindrop_api_token, self.serpapi_key, self.newsapi_key]):
|
28 |
raise EnvironmentError(
|
29 |
+
"Missing required environment variables. Please ensure all API keys are set."
|
30 |
)
|
31 |
+
|
32 |
self.client = OpenAI(api_key=self.openai_api_key)
|
33 |
+
self.newsapi = NewsApiClient(api_key=self.newsapi_key)
|
34 |
+
|
35 |
+
def get_google_results(self, query: str, num_results: int = 5) -> List[Dict]:
|
36 |
+
"""Get Google search results using SerpAPI."""
|
37 |
+
try:
|
38 |
+
params = {
|
39 |
+
"q": query,
|
40 |
+
"num": num_results,
|
41 |
+
"api_key": self.serpapi_key
|
42 |
+
}
|
43 |
+
search = GoogleSearch(params)
|
44 |
+
results = search.get_dict()
|
45 |
+
|
46 |
+
if 'organic_results' in results:
|
47 |
+
return results['organic_results']
|
48 |
+
return []
|
49 |
+
|
50 |
+
except Exception as e:
|
51 |
+
logger.error(f"Google search error: {e}")
|
52 |
+
return []
|
53 |
+
|
54 |
+
def get_news_results(self, query: str, num_results: int = 5) -> List[Dict]:
|
55 |
+
"""Get news articles using NewsAPI."""
|
56 |
+
try:
|
57 |
+
news_results = self.newsapi.get_everything(
|
58 |
+
q=query,
|
59 |
+
language='en',
|
60 |
+
sort_by='relevancy',
|
61 |
+
page_size=num_results
|
62 |
+
)
|
63 |
+
|
64 |
+
return news_results.get('articles', [])
|
65 |
+
|
66 |
+
except Exception as e:
|
67 |
+
logger.error(f"News API error: {e}")
|
68 |
+
return []
|
69 |
|
70 |
def generate_search_query(self, user_request: str) -> str:
|
71 |
"""Convert user request to optimized search terms."""
|
|
|
148 |
logger.error(f"Search error: {e}")
|
149 |
return []
|
150 |
|
151 |
+
def analyze_results(self, results: List[Dict], google_results: List[Dict],
|
152 |
+
news_results: List[Dict], user_query: str) -> str:
|
153 |
+
"""Generate an analysis of all search results."""
|
154 |
+
if not any([results, google_results, news_results]):
|
155 |
return "No relevant results found. Try modifying your search terms."
|
156 |
|
157 |
# Create context for analysis
|
158 |
context = f"Based on the search query: '{user_query}'\n\n"
|
159 |
+
context += "Analyze these sources:\n\n"
|
160 |
|
161 |
+
# Add Raindrop results
|
162 |
+
if results:
|
163 |
+
context += "Bookmarked Sources:\n"
|
164 |
+
for item in results:
|
165 |
+
context += f"Title: {item.get('title', 'No title')}\n"
|
166 |
+
if item.get('excerpt'):
|
167 |
+
context += f"Content: {item['excerpt'][:500]}...\n"
|
168 |
+
context += f"Created: {item.get('created', 'No date')}\n\n"
|
169 |
+
|
170 |
+
# Add Google results
|
171 |
+
if google_results:
|
172 |
+
context += "Web Sources:\n"
|
173 |
+
for item in google_results:
|
174 |
+
context += f"Title: {item.get('title', 'No title')}\n"
|
175 |
+
if item.get('snippet'):
|
176 |
+
context += f"Content: {item['snippet']}\n\n"
|
177 |
+
|
178 |
+
# Add News results
|
179 |
+
if news_results:
|
180 |
+
context += "News Sources:\n"
|
181 |
+
for item in news_results:
|
182 |
+
context += f"Title: {item.get('title', 'No title')}\n"
|
183 |
+
if item.get('description'):
|
184 |
+
context += f"Content: {item['description']}\n"
|
185 |
+
context += f"Published: {item.get('publishedAt', 'No date')}\n\n"
|
186 |
|
187 |
try:
|
188 |
prompt = f"""
|
189 |
+
Based on all available sources, provide a comprehensive analysis of {user_query}.
|
190 |
|
191 |
Requirements:
|
192 |
+
1. Focus on the most relevant and recent information
|
193 |
+
2. Organize the response by key themes
|
194 |
3. Include only factual information from the sources
|
195 |
4. Highlight any significant developments or changes
|
196 |
+
5. Note any conflicts or differences between sources
|
197 |
+
6. Prioritize official sources and recent developments
|
198 |
|
199 |
Context:
|
200 |
{context}
|
|
|
213 |
logger.error(f"Analysis generation error: {e}")
|
214 |
return "Error generating analysis."
|
215 |
|
216 |
+
def format_results(self, results: List[Dict], google_results: List[Dict],
|
217 |
+
news_results: List[Dict], analysis: str) -> str:
|
218 |
+
"""Format all search results with analysis."""
|
|
|
|
|
219 |
output = f"{analysis}\n\n"
|
220 |
output += "-------\n\n"
|
221 |
+
|
222 |
+
# Format Raindrop results
|
223 |
+
if results:
|
224 |
+
output += "π Bookmarked Sources:\n\n"
|
225 |
+
for idx, item in enumerate(results, 1):
|
226 |
+
if item.get('title') or item.get('link'):
|
227 |
+
output += f"{idx}. {item.get('title', 'No Title')}\n"
|
228 |
+
if item.get('link'):
|
229 |
+
output += f" Link: {item['link']}\n"
|
230 |
+
if item.get('tags'):
|
231 |
+
output += f" Tags: {', '.join(item['tags'])}\n"
|
232 |
+
if item.get('excerpt'):
|
233 |
+
output += f" Description: {item['excerpt'][:200]}...\n"
|
234 |
+
if item.get('created'):
|
235 |
+
created_date = item['created'][:10]
|
236 |
+
output += f" Created: {created_date}\n"
|
237 |
+
output += "\n"
|
238 |
|
239 |
+
# Format Google results
|
240 |
+
if google_results:
|
241 |
+
output += "π Web Sources:\n\n"
|
242 |
+
for idx, item in enumerate(google_results, 1):
|
243 |
output += f"{idx}. {item.get('title', 'No Title')}\n"
|
244 |
if item.get('link'):
|
245 |
output += f" Link: {item['link']}\n"
|
246 |
+
if item.get('snippet'):
|
247 |
+
output += f" Description: {item['snippet']}\n"
|
248 |
+
output += "\n"
|
249 |
+
|
250 |
+
# Format News results
|
251 |
+
if news_results:
|
252 |
+
output += "π° Recent News:\n\n"
|
253 |
+
for idx, item in enumerate(news_results, 1):
|
254 |
+
output += f"{idx}. {item.get('title', 'No Title')}\n"
|
255 |
+
if item.get('url'):
|
256 |
+
output += f" Link: {item['url']}\n"
|
257 |
+
if item.get('description'):
|
258 |
+
output += f" Description: {item['description']}\n"
|
259 |
+
if item.get('publishedAt'):
|
260 |
+
output += f" Published: {item['publishedAt'][:10]}\n"
|
261 |
+
if item.get('source', {}).get('name'):
|
262 |
+
output += f" Source: {item['source']['name']}\n"
|
263 |
output += "\n"
|
264 |
|
265 |
return output
|
|
|
273 |
search_query = self.generate_search_query(user_request)
|
274 |
logger.info(f"Using search query: {search_query}")
|
275 |
|
276 |
+
# Get results from all sources
|
277 |
+
raindrop_results = self.search_raindrop(search_query)
|
278 |
+
google_results = self.get_google_results(search_query)
|
279 |
+
news_results = self.get_news_results(search_query)
|
280 |
+
|
281 |
+
logger.info(f"Found {len(raindrop_results)} Raindrop results")
|
282 |
+
logger.info(f"Found {len(google_results)} Google results")
|
283 |
+
logger.info(f"Found {len(news_results)} News results")
|
284 |
|
285 |
# Generate analysis
|
286 |
+
analysis = self.analyze_results(raindrop_results, google_results, news_results, user_request)
|
287 |
|
288 |
# Format and return results
|
289 |
+
return self.format_results(raindrop_results, google_results, news_results, analysis)
|
290 |
|
291 |
except Exception as e:
|
292 |
logger.error(f"Error processing request: {e}", exc_info=True)
|
|
|
300 |
return bot.process_request(user_input)
|
301 |
|
302 |
# Define and launch the interface
|
303 |
+
with gr.Blocks(title="Enhanced Search Assistant", theme=gr.themes.Soft()) as demo:
|
304 |
gr.Markdown("""
|
305 |
+
# π Enhanced Search Assistant
|
306 |
+
Enter your search request in natural language, and I'll find and analyze information from multiple sources:
|
307 |
+
- Your bookmarked content
|
308 |
+
- Web search results
|
309 |
+
- Recent news articles
|
310 |
""")
|
311 |
|
312 |
with gr.Row():
|
|
|
322 |
with gr.Row():
|
323 |
output_text = gr.Textbox(
|
324 |
label="Analysis and Results",
|
325 |
+
lines=20,
|
326 |
interactive=False
|
327 |
)
|
328 |
|