Thanh Vinh Vo commited on
Commit
d6e8476
·
1 Parent(s): 09756ae
Files changed (2) hide show
  1. NOTES +1 -0
  2. app.py +1 -3
NOTES CHANGED
@@ -2,3 +2,4 @@
2
  - Provide tool is better than prompt
3
  - Don't give the master any tool, since it will try to delegate smaller work to the code agent, miss context
4
  - Temperature to 0
 
 
2
  - Provide tool is better than prompt
3
  - Don't give the master any tool, since it will try to delegate smaller work to the code agent, miss context
4
  - Temperature to 0
5
+ - BeautifulSoup too bad
app.py CHANGED
@@ -240,7 +240,6 @@ class BasicAgent:
240
  1. Take the question literally! Do not add any additional information or assumptions.
241
  2. `wikipedia` Python library is provided that makes it easy to to interact with Wikipedia pages.
242
  3. `pandas` Python package is provided that makes it easy to extract table data from Wikipedia HTML pages.
243
- 4. Only use BeautifulSoup to parse HTML at last resort!
244
  """,
245
  verbosity_level=0,
246
  max_steps=10,
@@ -251,7 +250,7 @@ class BasicAgent:
251
  model_id="Qwen/Qwen2.5-32B-Instruct",
252
  temperature=0.0,
253
  ),
254
- tools=[get_file, audio_to_text],
255
  managed_agents=[
256
  self.multimodal_agent,
257
  self.code_agent],
@@ -289,7 +288,6 @@ class BasicAgent:
289
  Please follow rules below:
290
  1. Take the question literally! Do not add any additional information or assumptions.
291
  2. `pandas` Python package is provided that makes it easy to extract table data from Wikipedia HTML pages.
292
- 3. Only use BeautifulSoup to parse HTML at last resort!
293
  """
294
  result = self.manager_agent.run(prompt)
295
  print(f"Agent responded with: {result}")
 
240
  1. Take the question literally! Do not add any additional information or assumptions.
241
  2. `wikipedia` Python library is provided that makes it easy to to interact with Wikipedia pages.
242
  3. `pandas` Python package is provided that makes it easy to extract table data from Wikipedia HTML pages.
 
243
  """,
244
  verbosity_level=0,
245
  max_steps=10,
 
250
  model_id="Qwen/Qwen2.5-32B-Instruct",
251
  temperature=0.0,
252
  ),
253
+ tools=[VisitWebpageTool(), GoogleSearchTool("serper"), get_file, audio_to_text, WikipediaSearchTool(), extract_table_from_html],
254
  managed_agents=[
255
  self.multimodal_agent,
256
  self.code_agent],
 
288
  Please follow rules below:
289
  1. Take the question literally! Do not add any additional information or assumptions.
290
  2. `pandas` Python package is provided that makes it easy to extract table data from Wikipedia HTML pages.
 
291
  """
292
  result = self.manager_agent.run(prompt)
293
  print(f"Agent responded with: {result}")