Unable to read history while using playground

[rk-yen]

Using version 2.7.8.

The following is how my Playground is setup.

duckDb = DuckDbTools(db_path=("/tmp/foo.db"))
agent = PythonAgent(
    name="My Agent",
    add_datetime_to_instructions=True,
    add_history_to_messages=True,
    num_history_responses=3,
    base_dir=tmp,
    tools=[
      duckDb
    ],
    description="...",
    instructions=[
      "Use tables to display data",
      "..."
    ],
    pip_install=True,
    show_tool_calls=False,
    markdown=True,
    debug_mode=True
)

app = Playground(agents=[agent]).get_app()
if __name__ == "__main__":
    serve_playground_app("my_agent:app", reload=True)

When invoking two consecutive messages, I see the following logs:

DEBUG    **************** METRICS END ******************                                                                             
DEBUG    ---------- OpenAI Async Response End ----------                                                                             
DEBUG    Added 2 Messages to AgentMemory                                                                                             
DEBUG    Added AgentRun to AgentMemory                                                                                               
DEBUG    --**-- Logging Agent Run (Async)                                                                                            
DEBUG    *********** Async Agent Run End: 05c52ccf-67da-4587-a07e-34a5d7be91fe ***********                                           
DEBUG    AgentRunRequest: <my question> d512153f-5407-4c6f-bd20-0b2f4fc83d11 rk           
         aa21a13c-77ec-427a-b47e-b75f43ce7f6a                                                                                        
DEBUG    Continuing session: d512153f-5407-4c6f-bd20-0b2f4fc83d11                                                                    
WARNING  Failed to deepcopy field: tools - cannot pickle 'PyCapsule' object                                                          
DEBUG    *********** Agent ID: 48cd3094-4beb-4cad-b27b-f0f651a29c9a ***********                                                      
DEBUG    *********** Session ID: d512153f-5407-4c6f-bd20-0b2f4fc83d11 ***********                                                    
DEBUG    Debug logs enabled                                                                                                          
DEBUG    Created new Agent: agent_id: 48cd3094-4beb-4cad-b27b-f0f651a29c9a | session_id: d512153f-5407-4c6f-bd20-0b2f4fc83d11        
INFO:     127.0.0.1:55588 - "POST /v1/playground/agent/run HTTP/1.1" 200 OK
DEBUG    *********** Async Agent Run Start: 09f69f78-e07e-45e2-830e-ea247b1e153c ***********                                         
DEBUG    Function show_tables from duckdb_tools added to model.                                                                      
DEBUG    Function describe_table from duckdb_tools added to model.                                                                   
DEBUG    Function run_query from duckdb_tools added to model.                                                                        
DEBUG    Function create_table_from_path from duckdb_tools added to model.                                                           
DEBUG    Function summarize_table from duckdb_tools added to model.                                                                  
DEBUG    Function save_to_file_and_run from python_tools added to model.                                                             
DEBUG    Function pip_install_package from python_tools added to model.                                                              
DEBUG    Function get_tool_call_history added to model.                                                                              
DEBUG    Building the system prompt for the PythonAgent.                                                                             
DEBUG    Getting messages from last 3 runs                                                                                           
DEBUG    Messages from last 3 runs: 0                                                                                                
DEBUG    ---------- OpenAI Async Response Start ----------   

In the logs above, I see:

DEBUG    Added 2 Messages to AgentMemory
DEBUG    Added AgentRun to AgentMemory

for the first run.
But, for the second run, I see:

DEBUG    Getting messages from last 3 runs
DEBUG    Messages from last 3 runs: 0

Is this a known issue?

[ysolanky]

Hello @rk-yen No, this is not a known issue, starting today, we have rebranded to Agno. The new library includes a number of improvements. Can you please migrate your code to Agno and then try again. Thank you!

[rk-yen]

@ysolanky ... I migrated to agno. And, see similar behavior.

DEBUG    ---------- OpenAI Async Response End ----------                                                                     
DEBUG    ---------- OpenAI Async Response End ----------                                                                     
DEBUG    Added 12 Messages to AgentMemory                                                                                    
DEBUG    Added AgentRun to AgentMemory                                                                                       
DEBUG    Could not create agent event: first argument must be callable or None                                               
DEBUG    *********** Agent Run End: 3add6a00-55af-4c5b-96b5-5d12d2506380 ***********  
...
...
...
DEBUG    *********** Agent ID: aa10d95d-fcb8-4408-bd58-d97bab80b3f5 ***********                                              
DEBUG    *********** Session ID: c99a137c-a1de-4b0f-bf94-25151704727e ***********                                            
DEBUG    *********** Async Agent Run Start: 6d4d634a-3259-42d9-8178-8bbdb188537a *********** 
DEBUG    Getting messages from last 3 runs                                                                                   
DEBUG    Messages from last 3 runs: 0                                                                                        
DEBUG    ---------- OpenAI Async Response Start ----------                                                                   
DEBUG    ============== developer ==============  

The code looks like the following:

from pathlib import Path
import tempfile

from agno.agent import Agent
from agno.tools.python import PythonTools
from agno.playground import Playground, serve_playground_app
from dotenv import load_dotenv

load_dotenv()

with tempfile.TemporaryDirectory() as f:
    tmp = Path(f)


my_agent = Agent(
    name="MyTest",
    add_datetime_to_instructions=True,
    add_history_to_messages=True,
    num_history_responses=3,
    tools=[
      PythonTools(base_dir=tmp, pip_install=True, run_code=True)
    ],
    description="""
      You are an expert in Python and can accomplish any task that is asked of you.
    """,
    show_tool_calls=False,
    markdown=True,
    debug_mode=True
)

app = Playground(agents=[my_agent]).get_app()
if __name__ == "__main__":
    serve_playground_app("my_agent:app", reload=True)