Gateway (gw) is a lightweight dispatcher that turns every Python function into a command line entry. It includes a recipe runner so you can compose automations with only functions and .gwr files.

Install from PyPI or from source and invoke gway on the command line. Every module inside projects/ becomes a namespace on gw and a CLI sub-command.

gway hello-world
gway awg find-awg --meters 30 --amps 60

from gway import gw
gw.hello_world()
result = gw.awg.find_awg(meters=30, amps=60)
print(result["awg"])

pip install gway pulls the latest released package from PyPI. Use this when you simply want to depend on GWAY in your own projects. To work on the framework itself clone the repository and install it in editable mode:

git clone https://github.com/arthexis/gway.git
cd gway
pip install -r requirements.txt
pip install -e .

• Gateway Object gw: entry point for all operations. Calling gw.project.func() is equivalent to gway project func.
• Projects: any .py file or directory inside projects/ is loaded on demand. Nested modules use dotted notation.
• Builtins: common utilities such as resource, run_recipe, help, test and notify are always available.
• Results & Context: return values are stored in gw.results and are referenced by name. Use sigils like [result.key] to pull values into later calls.
• Sigils: [VAR] or [object.attr] placeholders resolve from previous results, gw.context and environment variables. Automatic resolution only happens for sigils prefixed with % (e.g. %[VAR]); other sigils remain lazy until passed to gw.resolve or gw["VAR"].
• Recipes .gwr: text files listing commands. Indented lines reuse the previous command allowing very compact scripts. Run them via gway -r file or gw.run_recipe('file.gwr').
• Unquoted Kwargs: values after --key may include spaces up to the next - or -- token; quoting is optional.
• Environment Loading: envs/clients/<user>.env and envs/servers/<host>.env are read automatically. A file can specify a BASE_ENV to inherit defaults from another file.

• Async & Watchers: coroutines are executed in background threads. Use

  gw.until with file or URL watchers (and even PyPI version checks) to keep services running until a condition changes. PyPI version checks poll every 30 minutes by default.

• Resources: gw.resource resolves a file path in the workspace and can create files or directories. gw.resource_list lists files matching filters.
• Logging & Testing: gw.setup_logging configures rotating logs in logs/. gway test --coverage or gw.test() run the suite.

Recipes are .gwr files listing commands to automate tasks. Run them with gway -r file or gw.run_recipe('file.gwr').

awg awg-probe --target localhost
auth-db load sample.cdv

Folder Structure

Here's a quick reference of the main directories in a typical GWAY workspace:

The test suite verifies projects using the gw dispatcher. Install all dependencies and run gway test to execute it.

Tests are discovered recursively so directories under tests may mirror the source tree. A suggested structure is:

tests/
    gway/
    projects/

Before executing the suite, ensure the package and all dependencies are installed. Follow the commands in TESTING.rst to install requirements.txt and the editable package, then invoke gway test.

MIT License