Skip to content

Latest commit

 

History

History
136 lines (102 loc) · 5.47 KB

DEVELOPMENT.md

File metadata and controls

136 lines (102 loc) · 5.47 KB
Raw

DANDI Client Development

Development environment

Assuming that you have python3 (and virtualenv) installed, the fastest way to establish yourself a development environment (or a sample deployment), is via virtualenv:

git clone https://github.com/dandi/dandi-cli \
    && cd dandi-cli \
    &&  virtualenv --system-site-packages --python=python3 venvs/dev3 \
    && source venvs/dev3/bin/activate \
    && pip install -e .[test]

Install and activate precommit

Install pre-commit dependency with pip install pre-commit

In the source directory

pre-commit install

Running tests locally

You can run all tests locally by running tox (you can install tox running pip install tox):

tox -e py3

In order to check proper linting and typing of your changes you can also run tox with lint and typing:

tox -e lint,typing

dandi-api instance

The dandi-api repository provides a docker-compose recipe for establishing a local instance of a fresh dandi-api. See README.md:Docker for the instructions. In a new instance, you would need to generate a new API key to be used by the dandi client for upload etc.

Relevant dandi client commands (such as upload) are aware of such an instance as dandi-api-local-docker-tests. See the note below on the DANDI_DEVEL environment variable, which is needed in order to expose the development command line options.

Environment variables

  • DANDI_DEVEL -- enables otherwise hidden command line options, such as explicit specification of the dandi-api instance. All those options would otherwise be hidden from the user-visible (--help) interface, unless this env variable is set to a non-empty value

  • DANDI_API_KEY -- avoids using keyrings, thus making it possible to "temporarily" use another account etc for the "API" version of the server.

  • DANDI_LOG_LEVEL -- set log level. By default INFO, should be an int (10 - DEBUG).

  • DANDI_CACHE -- clear persistent cache handling. Known values are clear - would clear the cache, ignore - would ignore it. Note that for metadata cache we use only released portion of dandi.__version__ as a token. If handling of metadata has changed while developing, set this env var to clear to have cache clear()ed before use.

  • DANDI_INSTANCEHOST -- defaults to localhost. Point to host/IP which hosts a local instance of dandiarchive.

  • DANDI_TESTS_PERSIST_DOCKER_COMPOSE -- When set, the tests will reuse the same Docker containers across test runs instead of creating & destroying a new set on each run. Set this environment variable to 0 to cause the containers to be destroyed at the end of the next run.

  • DANDI_TESTS_PULL_DOCKER_COMPOSE -- When set to an empty string or 0, the tests will not pull the latest needed Docker images at the start of a run if older versions of the images are already present.

  • DANDI_TESTS_NO_VCR — When set, the use of vcrpy to playback captured HTTP requests during testing will be disabled

  • DANDI_DEVEL_INSTRUMENT_REQUESTS_SUPERLEN -- When set, the upload() function will patch requests to log the results of calls to requests.utils.super_len()

Sourcegraph

The Sourcegraph browser extension can be used to view code coverage information as follows:

  1. Install the Sourcegraph browser extension in your browser (Chrome or Firefox only)

  2. Sign up for a Sourcegraph account if you don't already have one. You must be signed in to Sourcegraph for the remaining steps.

  3. Enable the Codecov Sourcegraph extension

  4. On GitHub, when viewing a dandi-cli source file (either on a branch or in a pull request diff), there will be a "Coverage: X%" button at the top of the source listing. Pressing this button will toggle highlighting of the source lines based on whether they are covered by tests or not.

Releasing with GitHub Actions, auto, and pull requests

New releases of dandi-cli are created via a GitHub Actions workflow built around auto. Whenever a pull request is merged that has the "release" label, auto updates the changelog based on the pull requests since the last release, commits the results, tags the new commit with the next version number, and creates a GitHub release for the tag. This in turn triggers a job for building an sdist & wheel for the project and uploading them to PyPI.

Labelling pull requests

The section that auto adds to the changelog on a new release consists of the titles of all pull requests merged into master since the previous release, organized by label. auto recognizes the following PR labels:

  • major — for changes corresponding to an increase in the major version component
  • minor — for changes corresponding to an increase in the minor version component
  • patch — for changes corresponding to an increase in the patch/micro version component; this is the default label for unlabelled PRs
  • internal — for changes only affecting the internal API
  • documentation — for changes only affecting the documentation
  • tests — for changes to tests
  • dependencies — for updates to dependency versions
  • performance — for performance improvements