Skip to main content

Lightweight and lazy experiment logging

Project description

codecov PyPI PyPI - Python Version Documentation Status

Lightweight, lazy experiment logging

lazyscribe is a lightweight package for model experiment logging. It creates a single JSON file per project, and an experiment is only added to the file when code finishes (errors won't result in partially finished experiments in your project log).

lazyscribe also has functionality to allow for multiple people to work on a single project. You can merge projects together and update the list of experiments to create a single, authoritative view of all executed experiments.

Installation

$ python -m pip install lazyscribe

Basic Usage

The basic usage involves instantiating a Project and using the context manager to log an experiment:

from lazyscribe import Project

project = Project(fpath="project.json")
with project.log(name="My experiment") as exp:
    exp.log_metric("auroc", 0.5)
    exp.log_parameter("algorithm", "lightgbm")

You've created an experiment! You can view the experimental data by using list:

from pprint import pprint

pprint(list(project))
[
    {
        "name": "My experiment",
        "author": "<AUTHOR>",
        "last_updated_by": "<AUTHOR>",
        "metrics": {"auroc": 0.5},
        "parameters": {"algorithm": "lightgbm"},
        "created_at": "<CREATED_AT>",
        "last_updated": "<LAST_UPDATED>",
        "dependencies": [],
        "short_slug": "my-experiment",
        "slug": "my-experiment-<CREATED_AT>"
    }
]

Once you've finished, save the project to project.json:

project.save()

Later on, you can read the project back in read-only mode ("r"), append mode ("a"), or editable mode ("w+"):

project = Project("project.json", mode="r")
with project.log(name="New experiment") as exp:  # Raises a RuntimeError
    ...

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

lazyscribe-0.3.1-py3-none-any.whl (15.2 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page