Skip to main content

Run a command many times with different combinations of its inputs.

Project description

argsearch

argsearch is a simple and composable tool for running the same command many times with different combinations of arguments. It aims to make random search and grid search easy for things like hyperparameter tuning and setting simulation parameters, while only requiring that your program accepts command line arguments in some form.

Example

$ argsearch 'echo {a} {b}' grid 3 --a 0.0 1.5 --b X Y
[
{"args": {"a": "0.0", "b": "X"}, "command": "echo 0.0 X", "stdout": "0.0 X\n", "stderr": "", "returncode": 0},
{"args": {"a": "0.0", "b": "Y"}, "command": "echo 0.0 Y", "stdout": "0.0 Y\n", "stderr": "", "returncode": 0},
{"args": {"a": "0.75", "b": "X"}, "command": "echo 0.75 X", "stdout": "0.75 X\n", "stderr": "", "returncode": 0},
{"args": {"a": "0.75", "b": "Y"}, "command": "echo 0.75 Y", "stdout": "0.75 Y\n", "stderr": "", "returncode": 0},
{"args": {"a": "1.5", "b": "X"}, "command": "echo 1.5 X", "stdout": "1.5 X\n", "stderr": "", "returncode": 0},
{"args": {"a": "1.5", "b": "Y"}, "command": "echo 1.5 Y", "stdout": "1.5 Y\n", "stderr": "", "returncode": 0}
]

Installation

pip install argsearch

Usage

argsearch takes 4 kinds of arguments:

  • A command string with templates designated by bracketed names (e.g. 'python my_script.py --flag {value}'.
  • A search strategy (random or grid).
  • A number N defining the search extent (for random search, the number of trials to make; for grid search, the number of intervals to divide each range into).
  • A range for each template in the command string (e.g. --value 1 100).

Then, argsearch runs the command string several times, each time replacing the templates with values from their associated ranges.

Ranges

3 kinds of ranges are supported:

  • Float ranges are specified by a minimum and maximum floating-point value (e.g. --value 0.0 1.0).
  • Integer ranges are specified by a minimum and maximum integer (e.g. --value 1 100). Integer ranges are guaranteed to only yield integer values.
  • Categorical ranges are specified by a list of non-numeric categories, or more than two numbers (e.g. --value A B C, --value 2 4 8 16). Categorical ranges only draw values from the listed categories, and are not divided up during a grid search.

Note that values are sampled uniformly, so if you think of your range as log-uniform this sampling behavior may not work well.

Output

The output is JSON, which can be wrangled with jq or other tools, or dumped to a file. The run is a list of mappings, one per command call, each of which has the following keys:

  • args: a mapping of argument names to values.
  • command: the command string with substitutions applied.
  • stdout: a string containing the command's stdout.
  • stderr: a string containing the command's stderr.
  • returncode: an integer return code for the command.

Project details


Download files

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

Source Distribution

argsearch-0.1.1.tar.gz (5.6 kB view hashes)

Uploaded Source

Built Distribution

argsearch-0.1.1-py3-none-any.whl (6.5 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