Many logs are printed to stdout directly rather than via logger

[enricorotundo]

There're loads of print statements in the codebase writing to stdout. See example snippet below. This results in a rather verbose output.

2025-02-04 11:53:04,339 - naptha_sdk.client.node - INFO - Node URL: http://node.naptha.ai:7001
2025-02-04 11:53:04,339 - naptha_sdk.inference - INFO - Node URL: http://node.naptha.ai:7001
2025-02-04 11:53:04,339 - naptha_sdk.client.hub - INFO - Hub URL: ws://node.naptha.ai:3001/rpc
Parsed parameters: {'tool_name': 'chat', 'tool_input_data': 'what is an ai agent?'}
2025-02-04 11:53:05,031 - naptha_sdk.client.hub - INFO - Attempting authentication for user: enricorotundo
2025-02-04 11:53:05,181 - naptha_sdk.client.hub - INFO - Successfully authenticated user: enricorotundo
Found user... {'public_key': '..', 'id': 'user:....', 'is_registered': True}
Running Agent...

[...]

Agent run input: {'consumer_id': 'user:...', 'inputs': {'tool_name': 'chat', 'tool_input_data': 'what is an ai agent?'}, 'deployment': {'node': {'ip': 'node.naptha.ai', 'user_communication_port': 7001, 'user_communication_protocol': 'http'}, 'name': None, 'module': {'id': 'agent:simple_chat_agent', 'name': 'simple_chat_agent', 'module_type': 'agent'}, 'config': None, 'data_generation_config': None, 'tool_deployments': [], 'environment_deployments': [], 'kb_deployments': [], 'memory_deployments': []}, 'signature': '.....'}

Ideally, we'd want to use a logger object instead and control the logging level (INFO, DEBUG, etc.) and let user set the desired logging level.

[dishit7]

Hey, I’d like to work on this issue and here’s what I’ve observed so far:

🔹 The logging system is already in place using get_logger() in utils.py, but it defaults to DEBUG, leading to verbose output.
🔹 Some parts of the codebase use this logger, while others rely on print() statements, which makes log management inconsistent.
🔹 There's currently no way for users to control the verbosity of logs through the CLI.

Planned Approach:
Modify get_logger() in utils.py to check an environment variable (NAPTHA_LOG_LEVEL) for the log level.
Add a --log-level argument in cli.py that sets this environment variable, allowing users to control verbosity.
Replace print() statements with proper logger calls where necessary.
Would love to get your thoughts on this! If this approach looks good, could you assign me the issue? Let me know if there are any specific conventions I should follow. Thanks!

[enricorotundo]

Hi @dishit7 thanks for stepping forward! I'd say feel free to go ahead and submit a PR with your contribution, we'll be glad to review it!

One note. The idea is to have a limited number of INFO prints to log general info like the Node URL or HUB URL. Other verbose internal/technical stuff like logging user request payload, public key, etc. should probably be DEBUG.
So please use some common sense to evaluate logger prints and use either INFO or DEBUG.