How to get latest artifact created for a given branch using the Gitlab python API?

Question

Using this request it's possible to download the latest artifacts created in the Gitlab CI on a specific branch. For example:

curl "https://mygitlab.com/api/v4/projects/MYPROJECTNAME/jobs/artifacts/MYBRANCH/download" \
  --data-urlencode "job=build"

How do I do the same thing using the Gitlab python API? This page was all the documentation I could find, but it doesn't have a way of finding the latest job or artifact.

Answer 1

You can download the latest GitLab artifact from a job using the requests library or python-gitlab. You'll need to setup the following environment variables:

• GITLAB_TOKEN: If you have a paid version of GitLab, you can use the built-in CI_PROJECT_TOKEN and use PROJECT-TOKEN instead of PRIVATE-TOKEN in the request header. Otherwise, add this as a masked variable.

• CI_PROJECT_ID: If you're running from a pipeline, you get this for free. Otherwise, you can find the ID under the title in the project home page in GitLab.

python-gitlab

To download a specific file from the latest job in a branch:

import gitlab

project_id = os.getenv('CI_PROJECT_ID')     # built-in pipeline variable
gitlab_token = os.getenv('GITLAB_TOKEN')    # store in masked CI/CD variable
branch_name = 'master'
job_name = 'system-tests'

gl = gitlab.Gitlab('https://gitlab.example.com', private_token=gitlab_token)
project = gl.projects.get(project_id)
raw_data = project.artifact(ref_name=branch_name, artifact_path='path/to/filename', job=job_name)

The API docs show some examples for writing the streamed data to file: https://python-gitlab.readthedocs.io/en/stable/gl_objects/pipelines_and_jobs.html#streaming-example

To download the artifact as a zip from a job, it will be simpler to just use the requests library.

requests

The code below downloads all artifacts of the job as a zip file. To download the contents of a single artifact file instead of the whole zip, simply replace download in the URL below with raw/path/to/filename

import requests

project_id = os.getenv('CI_PROJECT_ID')     # built-in pipeline variable
gitlab_token = os.getenv('GITLAB_TOKEN')    # stored in masked CI/CD variable
branch_name = 'master'
job_name = 'system-tests'

url = f'https://gitlab.example.com/api/v4/projects/{project_id}/jobs/artifacts/{branch_name}/download'
headers = {'PRIVATE-TOKEN': gitlab_token}
params = {'job': job_name}

response = requests.get(url=url, headers=headers, params=params)

# if artifact is a text file:
print(response.text)