2.6 KiB

Исходник Постоянная ссылка Ответственный История

Contributing (CLI)

Azure/azureml-examples overall contributing guide.

Pull Requests

Pull requests (PRs) to this repo require review and approval by the Azure Machine Learning team to merge. Please follow the pre-defined template and read all relevant sections below.

Important: PRs from forks of this repository are likely to fail automated workflows due to access to secrets. PRs from forks will be considered but may experience additional delay for testing.

Assets

Example asset YAML files should be placed under the assets directory in the appropriate subdirectory. The test for assets currently runs az ml {asset} create.

Jobs

The jobs directory is structured by scenario then tool then project. The project directory should contain at least one job YAML file and a source code directory named src. A docker context or file can also be placed at the root of the project directory for use in the job file, although using prebuilt docker images is preferred.

Jobs should:

use public cloud data
use inline environment definitions
use inline data definitions
have a good description
use a pre-existing compute target (defined in setup.sh)
follow the YAML section order code > command > inputs > environment > compute > experiment_name > description

Endpoints

Endpoints are currently only tested through documentation scripts.

Scripts

Scripts are bash scripts with the .sh extension at the root of the directory. They are often used in the MicrosoftDocs/azure-docs repository as the source for code snippets.

Important: Suppress printing confidential data/secrets in your scripts

Note that all the shell scripts are executed with bash -x script.sh. This will print output of each line in the workflow output, which will be useful for debugging. However if you have sensitive information, that will get printed too. For e.g.:

# get endpoint access key using CLI command - this is confidential data
ENDPOINT_CREDENTIALS=$(az ml online-endpoint get-credentials -n $ENDPOINT_NAME -o tsv --query primaryKey)

in this case the output in the workflow run will have endpoint key printed.

To disable printing the output for selected lines by enclosing them within set +x and set -x. Example

set +x
ENDPOINT_CREDENTIALS=$(az ml online-endpoint get-credentials -n $ENDPOINT_NAME -o tsv --query primaryKey)
set -x

You can also add the environment variable as a github secret to ensure it is not printed elsewhere.

2.6 KiB Исходник Постоянная ссылка Ответственный История