pyforgejo 1.0.4

A client library for accessing the Forgejo API

pyforgejo

A client library for accessing the Forgejo API.

:warning: Package refactoring

This package is being rewritten to improve code quality, development velocity, and user experience.

The current package is production-ready. To ensure a smooth transition, we recommend pinning the package to 1.0.

The upcoming 2.0 release (expected in a few months) will introduce significant changes, including updated API calls (for example, client.repository.repo_get(repo, owner) will replace repo_get.sync_detailed(repo=repo, owner=owner, client=client)). You can try it in the 2.0 branch

We'll provide upgrade instructions and documentation for the 2.0 release.

Usage

Create a client:

from pyforgejo import AuthenticatedClient

client = AuthenticatedClient(base_url='https://codeberg.org/api/v1', token='API_TOKEN')

Call an endpoint:

from pyforgejo.api.user import user_get_current

response = user_get_current.sync_detailed(client=client)

# async version
response_async = user_get_current.asyncio_detailed(client=client)
response = await response_async

print(response)

Installation

pip install pyforgejo

Forgejo API

Resources:

• API Usage | Forgejo – Beyond coding. We forge.: user guide for the Forgejo API
• Forgejo API | Codeberg: API reference for Codeberg
• Forgejo API Swagger spec | Codeberg: Codeberg's Forgejo API Swagger spec
• openapi-generators/openapi-python-client: repo for the generator library that was used to generate pyforgejo
• About Swagger Specification | Documentation | Swagger: docs for Swagger spec
• The OpenAPI Specification Explained | OpenAPI Documentation: docs for OpenAPI spec

The structure of the import statement for interacting with a specific endpoint follows this pattern:

from pyforgejo.api.<root_path> import <operation_id>

Here, <tag> is the root path or tag for the endpoint in the Swagger spec, and <operation_id> is the operationId for the specific function you want to call, converted to snake case.

For example, for the endpoint /repos/search, the Swagger spec is:

"/repos/search": {
    "tags": ["repository"],
    "operationId": "repoSearch",
    ...
}

So to hit that endpoint, the import statement will be:

from pyforgejo.api.repository import repo_search

Every path/operation combo becomes a Python module with four functions:

• sync: Blocking request that returns parsed data (if successful) or None
• sync_detailed: Blocking request that always returns a Request, optionally with parsed set if the request was successful
• asyncio: Like sync but async instead of blocking
• asyncio_detailed: Like sync_detailed but async instead of blocking

Currently, Forgejo's API spec does not provide the response schemas for every endpoints, so most endpoints will not return parsed data and only have a detailed method.

All path/query parameters and bodies become method arguments.

Development

openapi-python-client

pyforgejo is generated with openapi-python-client, with as of now very little modification.

If you run into any issues, please create an issue in this repo.

If you want to work on a PR, please consider making a PR to openapi-python-client rather than to this repo.

openapi-python-client was chosen to generate this client over openapi-generator and fern because of the following reasons:

• openapi-python-client is Python-specific, which allows it to leverage specific language features, have a clearer code, and offer a better developer experience, compared to openapi-generator's one-size-fits-all approach
• openapi-python-client is written in Python, so users of pyforgejo will be more likely to be able to make contributions and fix bugs in the generator's code itself, while openapi-generator is written in Java, which represents a higher barrier to contributions
• openapi-python-client supports more authentication options, including access tokens, than fern
• the documentation is limited, but clearer than for openapi-generator

Generating the client with openapi-python-client

1. Convert Forgejo's Swagger spec to OpenAPI with swagger-converter, as Swagger is not supported by openapi-python-client.
2. Install openapi-python-client:

   pip install openapi-python-client
   

3. Create a config.yaml file with the following content:

   project_name_override: "pyforgejo"
   

4. Generate the client (this will create a pyforgejo/ dir):

   openapi-python-client generate --path /path/to/forgejo_openapi.json --config /path/to/config.yaml
   

5. Alternatively, update the client:

   git clone https://codeberg.org/harabat/pyforgejo
   
   openapi-python-client update --path /path/to/forgejo_openapi.json --config ./pyforgejo/config.yaml
   

6. Navigate to the pyforgejo/ dir and call the API:

   from pyforgejo.client import AuthenticatedClient
   from pyforgejo.api.user import user_get_current
   
   client = AuthenticatedClient(base_url='FORGEJO_URL' + '/api/v1', token='ACCESS_TOKEN')
   response = user_get_current.sync_detailed(client=client)
   
   print(response)
   # Response(status_code=<HTTPStatus.OK: 200>, ...)
   

Because merging of PRs on openapi-python-client can be slow, the fork at https://github.com/harabat/openapi-python-client, which is where I work on pyforgejo-related PRs to openapi-python-client, might be more up-to-date. In this case, replace step 1 above with the following:

git clone https://github.com/harabat/openapi-python-client.git
pip install ./openapi-python-client --upgrade

Modifying openapi-python-client

1. Clone and modify openapi-python-client

   git clone https://github.com/openapi-generators/openapi-python-client.git
   nvim openapi-python-client/openapi_python_client/parser/openapi.py
   # make your changes
   

2. Create and activate new env
3. Install (or upgrade) modified local package

   pip install ./openapi-python-client
   # pip install ./openapi-python-client --upgrade  # after making further changes
   

4. Generate a new client the regular way

   openapi-python-client generate --path /path/to/forgejo_openapi.json --config /path/to/config.yaml
   

Testing

We use pytest for testing.

The tests are in the tests dir. Run them with:

pytest ./tests/endpoints.py