ECMAScript interpreter for Cython & Python built for
- Solving & Decrypting Annoying Puzzles and Captchas at immense speeds.
- Calling upon different Javascript libraries Examples might include:
- llparse (Even though I maintain a python version but you could easily call upon llparse from javascript using this extension)
- yt-dlp-ejs (Decrypting challenges might take a while but that was why I wrote this library so we could start optimizing these things )
- Being tiny and easy to use
- Having ECMA6 Support
- Having a maintained backend (QuickJS-NG)
- Being a good companion alongside selectolax or beautiful-soup the choice is yours...
- License friendly, after abandoning pyduktape due to the backend no longer being maintained but also having a pretty poor license all together, It inspired me to try something new for a change that could run newer HTML5 Javascript for any puzzle that is thrown your way.
from cyjs import Context, Runtime
def main():
# You can also provide a runtime if needed it's usage before making multiple contexts however
# is completely optional
rt = Runtime()
ctx = Context(rt)
ctx.eval("function add(a, b){ return a + b; }; globalThis.add = add;")
add_func = ctx.get_global().get("add")
# 3
print(add_func(1, 2))
if __name__ == "__main__":
main()Example of use with external html parser tools.
from cyjs import Context
from selectolax.lexbor import LexborHTMLParser
# This example demonstates ways of cracking javascript out of webpages
# with an expernal HTML Parser and cyjs to handle the javascript logic.
# Know that A True HTML5 Dom-API Might require you to make your own
# functions and imagination but also reverse engineering the target page.
HTML_PAGE = b"""
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Document</title>
</head>
<body>
<div class="captcha">
<script>
function fake_captcha(name) {
return name + "-key";
}
</script>
<!-- Use your imagination a little... -->
</div>
<div>
<script>
this.captcha = fake_captcha('123')
</script>
</div>
</body>
</html>
"""
def main():
html = LexborHTMLParser(HTML_PAGE)
captcha = html.css_first("div.captcha > script").text(strip=True)
fake_solver = Context()
fake_solver.eval(captcha + "\nglobalThis.fake_captcha = fake_captcha;")
result = fake_solver.get_global().invoke("fake_captcha", "123")
# should print
# "123-key"
print(result)
if __name__ == "__main__":
main()There will be more examples in a future update.
| Library | ECMA-6 Support | Size | Performance | Typestubs And Readability | Cython cimportable |
|---|---|---|---|---|---|
| js2py | has to translate ECMA6 to 5 | Medium | Decent at ECMA5 and html5 scripts that use Emca-5 but starts suffering when trying to do ECMA-6 | Lacks proper typehinting making everything into an unwanted guessing game | No and it's pure python |
| quickjs (python library) | Yes | Small | Has to convert some objects back and forth but lacks typestubs and documentation | Lacks typestubs | CPython Extension that could've used a C-API Capsule to help it gain a bit of lubricant with other projects. |
| pyduktape | No | Small | somewhat fast but the backend is unmaintained and lacks proper type-hinting | None | Impossible |
| pyduktape2 | No | Small | somewhat fast but the backend is unmaintained and lacks proper type-hinting | Because this one didn't have that and my pull requests kept being laid dormant I wrote pyduktape3 | can't be done |
| pyduktape3 | No | Small | Very fast but the backend made by me but the C backend is unmaintained which is why the project was soon abandoned in favor of cyjs | I did add typestubs to this library. | The Last cherry on top this was cimporting pyducktape3 |
| strv8 | I haven't tried this one yet (Might be due to lack of windows support but I don't remeber) | Large due to V8 | Probably very fast becuase v8 is built by google. | I don't know, I sure hope a project like this has that. | I haven't seen the sourcecode yet... |
There's a few things I didn't get to becuase they are more or less puzzles to implement than they need to be but if anybody can figure these out feel free to fork and send a pull request along with a test added to pytest for eatch to ensure it works correctly. I may be uploading this library to pypi after the other things are implemented but these seem more of a chore for me to solve than really anything else.
-
If anybody finds a smarter approch to anything that has already been written throw me an issue or pull request.
-
Reporting and fixing bugs.
-
JSClass cdef class extension that can be subclassed in python and cython along with the hooks for all the JSClassExoticMethods (we need an approch to passing off a cdef class as an opaque value which I have not figured out how to do yet)
- JSClassFinalizer Hook
- JSClassGCMark Hook
- A safe approch for handling JSClass to python object conversion and vice versa (if possible)