@jseeio/jsee

v0.3.4

Published

5 days ago

JavaScript Execution Environment

Downloads

0High
0Medium
0Low

zemlyansky

Focus on code, not UI

Minimal example:

<html>
  <div id="jsee-container">
  <script src="https://cdn.jsdelivr.net/npm/@jseeio/jsee@latest/dist/jsee.runtime.js"></script>
  <script>
    function mul (a, b) {
      return a * b
    }
    new JSEE(mul, '#jsee-container')
  </script>
</html>

↳ Result

JavaScript Execution Environment

JSEE is a browser-based environment for processing tasks. It creates a graphical interface, executes code in a web worker or via API, bridges all pieces together into a user-friendly web app. In some cases, JSEE does all of that automatically, without any configuration. And when the configuration is required, it's just one JSON file with an intuitive structure.

Inputs and outputs

JSEE works best with functional tasks and one-way flow from inputs to outputs (i.e., inputs → processing → outputs). You can also extend it to more complex scenarios, like inputs → preprocessing → updated inputs → processing → outputs or inputs → processing → outputs → custom renderer. Even though many computational tasks have a functional form, some problems require more complex interactions between a user interface and code. For such cases, JSEE is probably too constrained. That makes it not as universal as R's shiny or Python's streamlit.

How it works

Instead of dealing with raw HTML tags, input elements or charts, JSEE makes it possible to work on a higher level and describe only inputs and outputs in a JSON schema. It similarly handles code execution, by checking the model part of that JSON object. Those three parts are the most important for the future app. In many cases JSEE can generate a new schema automatically by analyzing the code alone. For example, it's possible to extract a list function arguments and use them as model inputs. When JSEE receives the JSON schema it creates a new Vue app based on it and initializes a new worker for code execution. The final app can take inputs from a user, parse files, load needed libraries, orchestrate communication between code and GUI, use Web Workers to run everything smoothly

            Schema   Model   Render*
   DEV  -►   json    js/py     js
              |        |        |
           ┌──▼────────▼────────▼───┐
           │      new JSEE (...)    │
           └────────────────────────┘
              |               |
           ┌──▼──┐     ┌──────▼─────┐ ◄~ tf.js
 USER  ◄-► │ GUI │ ◄-► │    Model   │ ◄~ pyodide
           └─────┘     └────────────┘ ◄~ wasm
             Vue³        WebWorker*

 * - optional

JSEE takes a schema object that contains three main blocks:

model - describes a model/script/API (its location, is it a function or class, should it be called automatically on every GUI change or not)
inputs - list of inputs and their descriptions
outputs - list of outputs and their descriptions

Extra blocks can be provided for further customization

render - visualization part (optional). Defines custom rendering code.
design - overall appearance (optional). Defines how the app looks overwriting defaults.

imports - a list of urls (optional). Defines a list of scripts to load before the model is initialized.

"imports": [
  "https://cdn.jsdelivr.net/npm/@tensorflow/tfjs",
  "https://cdn.jsdelivr.net/pyodide/v0.17.0/full/pyodide.js"
]

examples - a list of examples (optional). Defines a list of examples that can be used to overwrite inputs.
```
"examples": [
  { "input": "My name is Anton and I am" },
]
```

Playground

Codepen

Schema

model - Contains main parameters of the model/script
- url (string) - URL of a JS/Python script or POST/GET API
- code (function) - It's possible to pass code directly to JSEE instead of using an URL
- name (string) - Name of the executed object. Default value is taken from url or code
- type (string, default - function) - What kind of script is loaded. Influences how the code is initializated. Possible values:
  - function
  - class
  - async-function
  - async-init
  - py
  - tf
- method (string) - If type is class, method defines the name of the class method to call during evaluation
- container (string) - How input values are passed to the function/method:
  - object (default) - Pass inputs wrapped in an object, i.e. {'x': 1, 'y': 2}
  - args - Pass inputs as separate arguments
- worker (boolean) - If true, JSEE initializes a Web Worker to run the script
render - Custom rendering script. Instead of relying on JSEE for output visualization, you can provide a custom script that visualizes the results. That can be useful if you rely on custom libs for plotting.
design - Design parameters
- layout - Layout for the model/input/output blocks. If it's empty and the JSEE container is not, JSEE uses inner HTML as a template. If the container is empty too, it uses the default blocks template.
- framework - Design framework to use. If a JavaScript object with the same name is present in a global context, JSEE loads it too (using Vue's use method).
inputs - Inputs definition.
- name* - Name of the input
- type* - Type. Possible types:
  - int, float or number - Number
  - string - String
  - text - Textarea
  - checkbox or bool - Checkbox
  - select or categorical - Select (one of many options)
  - file - File Input
  - action or button - Button (its name will be passed as a caller to the model)
- default - Default value
outputs - Outputs definition
- name* - Name of the output
- type* - Type. Possible types:
  - object - JavaScript Object
  - html or svg - SVG element
  - code - Code block
  - function - Render function. Rather than returning a value, a model returns a function that JSEE will call passing the container element.
examples - List of examples
autorun (boolean, default: false) - Defines if the script should be evaluated on each input change event
interval (number, default: 0) - Defines the interval between script evaluations (in milliseconds). If set to 0, the script is evaluated only once.

JSEE is a reactive branch of StatSim's Port. It's still work in progress. Expect API changes.

Changelog

0.3.4

JSEE:

[x] Add columns parameter to the inputs, outputs blocks (making it possible to create multi-column layouts, like simple dashboards)
[x] Add function output type (for custom renderers which take a container element as an argument)
[x] Add dom-to-image library for exporting dynamic output blocks to PNG
[x] Support for inputs to be set with url parameters (e.g. ?input1=1&input2=2)

HTML Generator:

[x] Add latex and table output in the markdown renderer
[x] Cache import scripts to avoid multiple loads when --fetch is used
[x] Infer description from the markdown and update html <head> with it
[x] Update social, org, ga blocks
[x] Small layout fixes

0.3.1

[x] Add download method to jsee object
[x] Add bin folder with cmd.js for easier project building

0.2.9

[x] Add examples
[x] Add imports
[x] Add caller field to the model input (can be: run, autorun or a button name)
[x] Add title field (for buttons rn)
[x] If display field is false the input is not shown
[x] If autorun is true, then actually autorun the model initially

0.2.8

[x] Fix no input case

0.2.7

[x] Show output when result is 0
[x] Updated style for buttons and inputs

0.2.6

[x] Tests
[x] Load schema from query (loader)
[x] Reset button appears only after data change
[x] Default input type (string)
[x] Directly load code when running in a window (not code to text)
[x] Passing code directly

Published

Vulnerabilities

Links

Maintainers

Keywords

Readme

Focus on code, not UI

JavaScript Execution Environment

Inputs and outputs

How it works

Playground

Schema

Changelog

0.3.4

JSEE:

HTML Generator:

0.3.1

0.2.9

0.2.8

0.2.7

0.2.6