Logging

The AEA framework supports flexible logging capabilities with the standard Python logging library.

In this tutorial, we configure logging for an AEA.

First of all, create your AEA.

aea create my_aea
cd my_aea

The aea-config.yaml file should look like this.

agent_name: my_aea
author: fetchai
version: 0.1.0
description: ''
license: Apache-2.0
aea_version: 0.6.0
fingerprint: {}
fingerprint_ignore_patterns: []
connections:
- fetchai/stub:0.21.0
contracts: []
protocols:
- fetchai/default:1.1.0
skills:
- fetchai/error:0.18.0
default_connection: fetchai/stub:0.21.0
default_ledger: fetchai
required_ledgers:
- fetchai
logging_config:
  disable_existing_loggers: false
  version: 1
private_key_paths: {}

By updating the logging_config section, you can configure the loggers of your application.

The format of this section is specified in the logging.config module.

At this section you'll find the definition of the configuration dictionary schema.

Below is an example of the logging_config value.

logging_config:
  version: 1
  disable_existing_loggers: False
  formatters:
    standard:
      format: '%(asctime)s [%(levelname)s] %(name)s: %(message)s'
  handlers:
    logfile:
      class: logging.FileHandler
      formatter: standard
      level: DEBUG
      filename: logconfig.log
    console:
      class: logging.StreamHandler
      formatter: standard
      level: DEBUG
  loggers:
    aea:
      handlers:
      - logfile
      - console
      level: DEBUG
      propagate: False

This configuration will set up a logger with name aea. It prints both on console and on file with a format specified by the standard formatter.

Streaming to browser

It is possible to configure the AEA to stream logs to a browser.

First, add the following configuration to your AEA:

logging_config:
  version: 1
  disable_existing_loggers: false
  formatters:
    standard:
      format: '%(asctime)s [%(levelname)s] %(name)s: %(message)s'
  handlers:
    http:
      class: logging.handlers.HTTPHandler
      formatter: standard
      level: INFO
      host: localhost:5000
      url: /stream
      method: POST
  loggers:
    aea:
      handlers:
      - http
      level: INFO
      propagate: false

Second, create a log server:

# -*- coding: utf-8 -*-
"""A simple flask server to serve logs."""

import datetime
import itertools
import queue

from flask import Flask, Response, request, stream_with_context


def format_log(log_dict):
    """Format a log record."""
    date = datetime.datetime.fromtimestamp(float(log_dict["created"]))
    formatted_log = f"[{date.isoformat()}] [{log_dict['levelname']}] {log_dict['name']}: {log_dict['msg']}"
    return formatted_log


def create_app():
    """Create Flask app for streaming logs."""
    all_logs = []
    unread_logs = queue.Queue()
    app = Flask(__name__)

    @app.route("/")
    def index():
        """Stream logs to client."""
        def generate():
            # stream old logs
            div = "<div>{}</div>"
            for old_row in all_logs:
                yield div.format(old_row)

            # stream unread logs
            while True:
                row = unread_logs.get()
                all_logs.append(row)
                yield f"<div>{row}</div>"

        rows = generate()
        title = "<p>Waiting for logs...</p>"
        return Response(stream_with_context(itertools.chain([title], rows)))

    @app.route("/stream", methods=["POST"])
    def stream():
        """Save log record from AEA."""
        log_record_formatted = format_log(dict(request.form))
        unread_logs.put(log_record_formatted)
        return {}, 200

    app.run()


if __name__ == "__main__":
    create_app()

Save the script in a file called server.py, install flask with pip install flask and run the server with python server.py.

Third, run your AEA and visit localhost:5000 in your browser.