pyasi / pybuildkite Goto Github PK

Let's add the functionality provided in the documentation for the Teams endpoint here

The create_yaml_pipeline method is missing many parameters listed at https://buildkite.com/docs/apis/rest-api/pipelines#create-a-yaml-pipeline. These all need to be added. Alternatively, I would actually recommend just accepting kwargs for all parameters that don't need any further formatting. Then the Python package can be robust to changes to the Buildkite API with minimal maintenance. This comes at the cost of a more self-documenting Python API, but given the lack of activity on this repo I think it would be better to just point to the Buildkite documentation as the source of truth.

Create a method for the Add a Webhook functionality

• Create a new method in pipelines.py to use the endpoint described above
• Write unit tests for the new code

Sometimes I see my jobs fail to submit due to a network transient error. If you use requests.Session with a reasonable default of retries (3?) it should address the vast majority of these issues.

Pipelines, Builds, Jobs, and Agents are all currently only using GET endpoints. We need full CRUD ability, adding POST, PUT, and DELETE functionality to these to make them feature complete would be great. Doesn't have to be all in one PR.

It would be great if pybuildkite shipped with type stubs included so that consuming projects could use mypy to ensure that they are using valid fields/functions on pybuildkite objects

I was inspired by PyGithub's most recent release shipping with pyi stub files included with the package. They are still in the process of updating their CI process to keep the pyi files up-to-date with the source code.

I'd be willing to help out with getting the type information in pybuildkite, but was wondering what kind of approach would you want to take? .pyi files, or in-line type hints -- any desire to keep python 2 compatibility?

Latest GitHub release is 1.2.1, latest pypi release is 1.2.2 - these should be brought in sync.

Hi,

I am trying to list builds based on the value of meta_data variables:

builds = buildkite.builds().list_all_for_pipeline(org_name,
                                                  pipeline_name,
                                                  meta_data={'origin_id': 0}
                                                  )

but it seems that pybuildkite is building the url as ...?metadata={'origin':0}&... instead of ...?metadata[origin]=0&... as defined in Buildkite API documentation.

I checked the unit tests but there doesn't seem to have anything using it.

If this is indeed a bug I would be happy to make a PR to fix it.

Create a new file and class to incorporate the meta API.

• Create a new file called meta.py
• Create a class called Meta
• Return the Meta class from a method in buildkite.py
• Write unit tests for the new code

Just curious if it's intentionally not public

https://buildkite.com/pybuildkite/pybuildkite

Due to following line the state query param looks like passed instead of state=passed

pls add: state=+

If you use a datetime object including tzinfo e.g.

from datetime import datetime, deltatime
from datetime impor timezone

UTC = timezone.utc

finished_from  = datetime.now(UTC)-deltatime(days=1)

then you'll get an error because the api endpoint will be wrong generated

It appears that the list_annotations() wrapper is not paginated, since it does not accept page or with_pagination parameters as the other paginated wrappers do.

Related to #85, the code in the 1.2.2 PyPi release doesn't actually match the code from the 1.2.2 GitHub release. PyPi 1.2.2 looks like it was released December 20th, whereas the GitHub release was January 13th. Could you create a new release on PyPi and GitHub (either 1.2.3 or 1.3.0) so that the codebase is aligned? 🙏

The API provides support for updating a pipeline via PATCH: https://buildkite.com/docs/apis/rest-api/pipelines#update-a-pipeline

This doesn't seem to be supported by the library yet (the base client also doesn't provide a PATCH method, so the only way to update a pipeline currently is using the rest API directly as far as I can tell).

https://buildkite.com/changelog/99-introducing-the-yaml-steps-editor

After migrating to YAML steps, create_pipeline() no longer works.

Buildkite throws an HTTP 500 error when trying to use

buildkite.pipelines().create_pipeline(
    organization = "org",
    pipeline_name = "pipeline",
    git_repository = "[email protected]:pyasi/pybuildkite.git"
)

Error

requests.exceptions.HTTPError: 500 Server Error: Internal Server Error for url: https://api.buildkite.com/v2/organizations/dropbox/pipelines/?per_page=100

The same error can be reproduced using curl.

In order to create a pipeline, we must use the new param configuration in https://buildkite.com/docs/apis/rest-api/pipelines#create-a-pipeline-yaml-pipelines

I believe that the functions list_artifacts_for_job and list_artifacts_for_build are inverted:

The list_artifacts_for_build should not need a job, only the build. The function that gets a job should be list_artifacts_for_job.

This is also in line with the API docs: https://buildkite.com/docs/apis/rest-api/artifacts

I would also be happy to open a PR for this one

Hi, I am trying to load all our builds via
buildkite.builds().list_all
This returns the first 100 results (as expected)
but I can't figure out if there is a way to get the next 100 results.

It would be great if a new release could be issued, as I'm looking forward to using this improvement 🙏 #82

A majority of the docstrings in these classes are not filled out. It would be great to document each class and method.

https://buildkite.com/changelog/215-upcoming-change-to-the-buildkite-api

Will there be a SDK release that includes the changes?

tests/test_client.py::TestClientRequest can fail when running with pytest --randomly-seed=1234 or pytest --randomly-seed=123.

Create a new file and class to incorporate the Access Token API

• Create a new file called access_token.py
• Create a class called AccessToken
• Return the AccessToken class from a method in buildkite.py
• Write unit tests for the new code

Let's get this repo to be complete with all opened API functionality provided by Buildkite. Does not have to be a single PR.

Let's add the functionality provided in the documentation for the Users endpoint here

This repository needs tests for each of the Buildkite API functions i.e agents.py or jobs.py.

Create methods for the Archive and Unarchive functionality

• Create new methods in pipelines.py to use the endpoints described above
• Write unit tests for the new code