This is a pre-production deployment of Warehouse, however changes made here WILL affect the production instance of PyPI.
Latest Version Dependencies status unknown Test status unknown Test coverage unknown
Project Description
Experimenter [![pypi](https://img.shields.io/pypi/v/experimenter.svg)](https://pypi.python.org/pypi/experimenter)
=======
Use git tags to log experiments and the exact code that was used to run those experiments. The goal is to make every experiment completely reproducible. To install use
```
pip install experimenter
```
All contributions are welcome. Experimenter was inspired by [Ali's](http://arkitus.com/patterns-for-research-in-machine-learning/) and [Charles'](http://www.theexclusive.org/2012/08/principles-of-research-code.html) blog posts.

Use Cases
-----
* It is the case that you might make some changes in your machine learning code, not necessarily worth committing, and spawn a process to run an experiment that uses these changes. However, by the time the experiment has finished you don't know what changes you were testing (presumably because you made more changes in the meanwhile).
* You need a distributed way of collecting all the experiments (parameters and results), making sure that they have been run on the exact same version of the code.

Usage
-----
Create an `ExperimentLogger` object, passing the parameters of the experiment. When the experiment is finished, call
the `record_results()` method of that object, ie:

```python
with ExperimentLogger(name="NameOfExperiment", parameters=parameters_dict) as experiment_logger:
...
experiment_logger.record_results(dict_of_results)
```

Behind the scenes, a git tag will be created (committing any changes you may have in the working tree, into a different branch). The tag will have a name of the form `exp_NameOfExperiment_timestamp` and in the message it will have a JSON representation of the parameters and the results (when/if recorded). The working state of the current branch will seemingly remain unaffected. Note that this is *not* thread-safe.

If no result is recorded (ie `record_results` is not called) within the `with` then the experiment will be deleted upon exit from that block. This is useful when stopping experiments or when experiments fail before finishing.

There is a command-line tool that helps with retrieving tests. In the git folder, run the command
```
> experimenter -c SHAofCodeState
```
to retrieve all the experiments that happened with the give code version. If `-c` is not provided then all experiments will be shown. If `-s` is provided only experiments that have results will be shown. The strict command is off by default. Use `--help` for more information.

TODO
------
* A command line tool for starting/stopping experiments.
* Auto-push tags method in `ExperimentLogger`

Release History

Release History

0.1.6

This version

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1.5

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1.4

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1.3

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1.2

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1.1

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

0.1

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

Download Files

Download Files

TODO: Brief introduction on what you do with files - including link to relevant help section.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
experimenter-0.1.6.tar.gz (5.0 kB) Copy SHA256 Checksum SHA256 Source May 8, 2016

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS HPE HPE Development Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting