Skip to main content

Container platform for IDM-Tools

Project description

Staging: idmtools-platform-container

Idmtools platform container

Table of Contents

Introduction

ContainerPlatform is a platform designed to facilitate the execution of experiments and simulations within Docker containers. It provides a robust environment with all necessary tools and dependencies installed, allowing for seamless integration and execution of computational tasks.

Pre-requisites

  • Python 3.8/3.9/3.10/3.11/3.12 x64-bit
  • OS:
    • Windows 10 Pro or Enterprise
    • Linux
    • macOS (10.15 Catalina or later)
  • Docker or Docker Desktop(required for the container platform) On Windows, please use Docker Desktop 4.0.0 or later
  • Mac user: Only support Intel based x86_64 architecture if you want to run emodpy related disease models on Docker container platform. Apple based ARM architecture currently is not supported.

Installation

  • Install python

    Ensure you have Python 3.8+ installed on your system.

  • Create and activate a virtual environment:

    python -m venv venv
    source venv/bin/activate  # On macOS/Linux
    venv\Scripts\activate     # On Windows
    
  • Install all container platform related packages

    pip install idmtools[container] --index-url=https://packages.idmod.org/api/pypi/pypi-production/simple
    
  • Optional: Install all idmtools packages

    pip install idmtools[full] --index-url=https://packages.idmod.org/api/pypi/pypi-production/simple
    
  • To override existing idmtools container related packages after installing emodpy, run this command

    pip install idmtools[container] --index-url=https://packages.idmod.org/api/pypi/pypi-production/simple --force-reinstall --no-cache-dir --upgrade
    

    Mac user: You map need to escape the square brackets with a backslash like \[container\] or \[full\] in above command.

  • Extra steps for Windows user:

    • Enable Developer Mode on Windows

      If you are running the script on Windows, you need to enable Developer Mode. To enable Developer Mode, go to Settings -> Update & Security -> For developers and select Developer Mode on, or refer to this guide.

    • Enable long file path for Windows

      Due to the file/folder structure design outlined in the section below, if running the script on Windows, be aware of the file path length limitation (less than 255 characters).

      To allow longer file paths, you can enable Long Path Support in the Windows Group Policy Editor. Refer to this guide for detailed instructions.

Examples for container platform

Initialize platform

  • This is the example using Container Platform
    from idmtools.core.platform_factory import Platform
    platform = Platform('CONTAINER', job_directory='<user job directory>')
    
  • To trigger MPI, use ntasks >=2:
    from idmtools.core.platform_factory import Platform
    platform = Platform('CONTAINER', job_directory='<user job directory>', ntasks=2)
    
  • More options for container platform initialization: refer to ContainerPlatform attributes

Container Examples

Check result with CLI commands

idmtools container status <experiment id>

Check result files

- on host: `<job_directory>/<suite_path>/<experiment_path>/<simulation_path>/`
- in container: `/home/container-data/<suite_path>/<experiment_path>/<simulation_path>/`

Folder structure

By default, idmtools now generates simulations with the following structure: job_directory/s_<suite_name>_<suite_uuid>/e_<experiment_name>_<experiment_uuid>/simulation_uuid

  • job_directory — The base directory that contains all suite, experiment, and simulation folders.
  • s_<suite_name>_<suite_uuid> — The suite directory, where the suite name (truncated to a maximum of 30 characters) is prefixed with s_, followed by its unique suite UUID.
  • e_<experiment_name>_<experiment_uuid> — The experiment directory, where the experiment name (also truncated to 30 characters) is prefixed with e_, followed by its unique experiment UUID.
  • simulation_uuid — The simulation folder identified only by its UUID.

Suite is optional. If the user does not specify a suite, the folder will be: job_directory/e_<experiment_name>_<experiment_uuid>/simulation_uuid

Examples:

If you create a suite named: my_very_long_suite_name_for_malaria_experiment

and an experiment named: test_experiment_with_calibration_phase idmtools will automatically truncate both names to a maximum of 30 characters and apply the prefixes s_ for suites and e_ for experiments, resulting in a path like:

job_directory/
└── s_my_very_long_suite_name_for_m_12345678-9abc-def0-1234-56789abcdef0/
    └── e_test_experiment_with_calibrati_abcd1234-5678-90ef-abcd-1234567890ef/
        └── 7c9e6679-7425-40de-944b-e07fc1f90ae7/

Or for no suite case:

job_directory/
└── e_test_experiment_with_calibrati_abcd1234-5678-90ef-abcd-1234567890ef/
    └── 7c9e6679-7425-40de-944b-e07fc1f90ae7/

Users can customize this structure through the idmtools.ini configuration file:

  • name_directory = False — Excludes the suite and experiment names (and their prefixes) from the simulation path.
  • sim_name_directory = True — Includes the simulation name in the simulation folder path when name_directory = True.

Basic CLI commands

ContainerPlatform provides several CLI commands to manage and monitor experiments and simulations. Below are some basic commands:

List running jobs

To list running experiment or simulation jobs:

idmtools container jobs [<container-id>] [-l <limit>] [-n <next>]

Check status

To check the status of an experiment or simulation:

idmtools container status <item-id> [-c <container_id>] [-l <limit>] [--verbose/--no-verbose]

Cancel job

To cancel an experiment or simulation job:

idmtools container cancel <item-id> [-c <container_id>]

View Experiments history

To view experiments history:

idmtools container history [<container-id>] [-l <limit>] [-n <next>]

Note

  • WorkItem is not supported on the Container Platform as it is not needed in most cases since the code already runs on user's local computer.
  • AssetCollection creation or referencing to an existing AssetCollection are not supported on the Container Platform with current release. If you've used the COMPS Platform, you may have scripts using these objects. You would need to update these scripts without using these objects in order to run them on the Container Platform.
  • Run with Singularity is not needed with Container Platform. If you take existing COMPS example and try to run it with Container Platform, you may need to remove the code that setups the singularity image.

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

idmtools_platform_container-0.0.4.tar.gz (67.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

idmtools_platform_container-0.0.4-py3-none-any.whl (69.3 kB view details)

Uploaded Python 3

File details

Details for the file idmtools_platform_container-0.0.4.tar.gz.

File metadata

File hashes

Hashes for idmtools_platform_container-0.0.4.tar.gz
Algorithm Hash digest
SHA256 74bac06e7a6e3372a9ce542ecaa21c4c5886f5a35fe02a7cd3d5faff57c17d64
MD5 af6cd869b616a65651c258b20bc64ec5
BLAKE2b-256 de24cb682a4fd4b476e745606843e75823c767e57b025c860753f9d5fc62812b

See more details on using hashes here.

Provenance

The following attestation bundles were made for idmtools_platform_container-0.0.4.tar.gz:

Publisher: deploy.yml on shchen-idmod/idmtools-1

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file idmtools_platform_container-0.0.4-py3-none-any.whl.

File metadata

File hashes

Hashes for idmtools_platform_container-0.0.4-py3-none-any.whl
Algorithm Hash digest
SHA256 283368f73d74dfb20484ed86e081ab0f6ed764a6745a4e5dfe25811c4dde7113
MD5 593234a4c1b9a886961696ffc637e04e
BLAKE2b-256 4e62267dd2ef1f759c72e05c2a3eef1b688b502762040b59a8336c79a167d523

See more details on using hashes here.

Provenance

The following attestation bundles were made for idmtools_platform_container-0.0.4-py3-none-any.whl:

Publisher: deploy.yml on shchen-idmod/idmtools-1

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

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