A small, simple kernel of PyScript, made for testing purposes in the spirit of a code spike.
This is the way:
- Obvious code.
- Simple is good.
- No dependencies.
- Vanilla JavaScript.
- Pluggable.
- Comments.
- (Some) tests.
- Build for change.
This is a foundation for lightweight testing of Python interpreters that target WASM. Inspired by code in the "real" PyScript website and our plans for plugins and simple event based coordination.
Complexity, edge cases and customization is (hopefully) confined to plugins and bespoke interpreters.
That is all.
In order to compile MicroPython you'll need to ensure you have the expected dev tools described here:
https://docs.micropython.org/en/latest/develop/gettingstarted.html
Otherwise, common tasks are scripted by a Makefile (tested on Linux):
$ make
There's no default Makefile target right now. Try:
make setup - clone the required repositories.
make update - update the emsdk compiler.
make mp - compile MicroPython for WASM into the mpbuild directory.
make serve - serve the project at: http://0.0.0.0:8000/
make test - while serving the app, run the test suite in browser.
To get a working development environment with MicroPython run:
$ make setup
$ make update
$ make mp
To check things are working:
$ make serve
Then point your browser to http://0.0.0.0:8000/ to see the first page of an
interactive technical report about using MicroPython. You should be able to
change the interpreter from micropython
to pyodide
and things should just
work as before, but with a different interpreter at the bottom of the PyScript
stack.
TESTS ARE CURRENTLY BROKEN
For the sake of simplicity (and familiarity) we use the Jasmine test framework to exercise the JavaScript aspects of our code.
Ensure the project is being served (make serve
) and in a different shell, in
the root of this project, type make test
. Your default browser should open
and run the Jasmine based test suite.
The PyScript core only loads configuration, starts the Python interpreter, allows the registration of plugins and adds files to the interpreter's filesystem. All other logic, capabilities and features are contained in the plugins.
Currently, only two plugins are provided:
- One built into PyScript that implements the core
<py-script>
tag. - The other (in
customtags.js
) implements the<py-repl>
tag to demonstrate a "third party" plugin.
The story of PyScript's execution is roughly as follows:
- Configuration is loaded from the
<py-config>
tag. Once complete thepy-configured
event is dispatched, containing theconfig
object based upon default values overridden by the content of the<py-config>
tag. - When the
py-configured
event is dispatched three things happen:- The interpreter is loaded via injecting a
<script>
tag that references the interpreter's URL. Once loaded thepy-interpreter-loaded
event is dispatched. - Plugins are registered and have their
configure
function called. For each plugin registered apy-plugin-registered
event is dispatched, containing the (potentially changed)config
, and a reference to the newly registered plugin. - The content of the files to be added to the interpreter's filesystem are
fetched. Once downloaded each file causes a
py-file-fetched
event to be dispatched with the path and content of the file attached to it.
- The interpreter is loaded via injecting a
- When
py-interpreter-loaded
is dispatched two things happen:- The interpreter is instantiated / started. Once complete the
py-interpreter-ready
event is dispatched. - All registered plugins have their
start
function called and apy-plugin-started
event is dispatched for each plugin.
- The interpreter is instantiated / started. Once complete the
- When the
py-interpreter-ready
event is dispatched all plugins have theironInterpreterReady
function called with theconfig
andinterpreter
objects. At this point all files are copied onto the interpreter's filesystem. When all the files are copied thepy-files-loaded
event is dispatched. - When both the interpreter and filesystem are finished setting up and in a
ready state, the
py-finished-setup
event is dispatched to signal PyScript is ready to evaluate user's code. - Any plugins registered after the interpreter is ready immediately have their
configure
,start
andonInterpreterReady
functions called, with thepy-plugin-registered
andpy-plugin-started
events being dispatched.
That's it!
When pyscript.js
is run, it creates a window.PyScript
object that contains
read-only references to the config
, registered plugins
,
availableInterpreters
, the interpreter
used on the page, a
n isInterpreterReady
flag, a registerPlugin
function (see below) and a
runPython(code)
function that takes a string of Python.
There are copious comments in the pyscript.js
file. My intention is for
simplicity, lack of onerous dependencies (bye-bye npm
), and
understandability. This code is good if it's easy to understand what's going
on. To this end, it's laid out in a "literate" manner, so the code "tells the
story" of this implementation of PyScript by reading it from top to bottom.
Plugins are inspired by Antonio's suggestion found here, and should be relatively self explanatory.
Since simplicity is the focus, plugins are simply JavaScript objects.
Such objects are expected they have a name
attribute referencing a string
naming the plugin (useful for logging purposes). Plugins should also provide
one or more of the following functions attached to them, called in the
following order (as the lifecycle of the plugin):
configure(config)
- Gives the plugin early access to theconfig
object. Potentially, the plugin can modify it, and modifications will be visible to later steps and other plugins. Plugins must only modify the config at this point in their life-cycle. Examples of things which plugins might want to do at this point:- Early sanity check about their own options.
- Rename/remap some options.
- Add new packages to install.
- Modify options for other plugins (e.g. a debugger plugin might set the
option
show_terminal
).
start(config)
- The main entry point for plugins. At this point, config should not be modified by the plugin. Example use cases:- Define custom HTML elements.
- Start fetching external resources.
onInterpreterReady(config, interpreter)
- Called once the interpreter is ready to evaluate Python code. Example use cases:pip install
packages.- Import/initialize Python plugins.
The following events, dispatched by PyScript itself, are related to plugins:
py-plugin-registered
- Dispatched when a plugin is registered (and the event contains a reference to the newly registered plugin). This happens immediately after the plugin'sconfigure
function is called.py-plugin-started
- Dispatched immediately after a plugin'sstart
function is called. The event contains a reference to the started plugin.py-interpreter-ready
- causes each plugin'sonInterpreterReady
function to be called.
If a plugin is registered after the interpreter is ready, all three functions are immediately called in the expected sequence, one after the other.
The recommended way to create and register plugins is:
const myPlugin = function(e) {
/*
Private and internal logic, event handlers and event dispatch can happen
within the closure defined by this function.
e.g.
const FOO = "bar";
function foo() {
const myEvent = new CustomEvent("my-event", {detail: {"foo": FOO}});
document.dispatchEvent(myEvent);
}
function onFoo(e) {
console.log(e.detail);
}
document.addEventListener("my-event", onFoo);
...
*/
const plugin = {
name: "my-plugin",
configure: function(config) {
// ...
},
start: function(config) {
// ...
foo();
},
onInterpreterReady: function(config, interpreter) {
// ...
}
};
window.pyScript.registerPlugin(plugin);
}();
document.addEventListener("py-configured", myPlugin);
Then in your HTML file:
<script src="myplugin.js"></script>
<script src="pyscript.js" type="module"></script>
The Interpreter
class abstracts away all the implementation details of the
various Python interpreters we might use.
To see a complete implementation see the MicroPythonInterpreter
class that
inherits from Interpreter
. There is also an incomplete PyodideInterpreter
class so I was able to compare and contrast the differences between
implementations and arrive at a general abstraction (still very much a work in
progress). Comments in the code should explain what's going on in terms of the
life-cycle and capabilities of a "interpreter".
The afore mentioned MicroPythonInterpreter
, CPythonInterpreter
and
PyodideInterpreter
all, to a greater or lesser extent, define a uniform shim
around their respective interpreter. The MicroPython one is most complete, but
still needs work as I make changes to how MicroPython itself exposes stdout
,
stderr
and consumes stdin
.
Who knows..? But this is a good scaffold for testing different Python interpreters.
Next steps:
- More comprehensive tests.
CPythonInterpreter
fully implemented.PyodideInterpreter
finished.MicroPythonInterpreter
refactored after making MicroPython play nicer withstdout
andstderr
.- A uniform way to
pip install
packages in each interpreter. - A uniform JavaScript gateway from within each interpreter.
- A uniform
navigator
object through which to access the DOM from within each interpreter. - Running in web-workers (and associated message passing work), for each interpreter.
That's it..! ;-)