Compyler (Python 3)

A Python 3 conversion of zTrix's webpage2html. Instructions on performing the port are provided below. This conversion works and functions for compiling Slate builds for distribution.

The Python 3 version is in compyler.py while the original Python 2 version is left in webpage2html.py.

Conversion Process

$ 2to3 -w webpage2html.py

Search for lines utilizing base64.b64encode(). Append .decode("ascii") to these lines to get something like:

`base64.b64encode(data).decode("ascii")`

See lines 141, 204, 222.

Replace js_str.find(<str>) with js_str.find(bytes(<str>,"utf-8")) on lines 221, 223.

Replace js_str with js_str.decode("utf-8") on line 224.

Use a buffer on line 322:

# see: https://stackoverflow.com/questions/3597480/how-to-make-python-3-print-utf8
sys.stdout.buffer.write(rs)

To ensure that htmlentities are not incorrectly encoded and lost, modify the return of compile() starting on line 270:

# see: https://stackoverflow.com/questions/701704/convert-html-entities-to-unicode-and-vice-versa
if prettify:
    text = soup.prettify(formatter='html')
else:
    text = str(soup)
return text.encode("ascii", "xmlcharrefreplace")

This causes compile() to return a bytes object, which should be written to a file or stdout using Buffered I/O.

Additional Changes

Function prototype of generate changed from:

def generate(index, verbose=True, comment=True, keep_script=False, prettify=False, full_url=True, verify=True, errorpage=False):

To:

def compile(index, verbose=True, comment=True, keep_scripts=True, prettify=False, full_url=True, verify=True, errorpage=False):

keep_scripts=True ensures that JavaScript is saved by default. See line 176.

Line 321 is changed to use the updated prototype:

rs = compile(args.url, **kwargs)

The -s, --script arg behavior on line 308 is inverted:

parser.add_argument('-s', '--scripts', action='store_true', help="don't embed JavaScript in the generated html ")

Finally, change line 315, 316 to:

if args.scripts:
    kwargs['keep_scripts'] = False

Webpage2html: Original Documentation

Webpage2html: Save web page to a single html file

This is a simple script to save a web page to a single html file. No mhtml or pdf stuff, no xxx_files directory, just one single readable and editable html file.

The basic idea is to insert all css/javascript files into html directly, and use base64 data URI for image data.

Usage and Example

save webpage directly from url(recommended way):

$ python2 webpage2html.py http://www.google.com > google.html

or save webpage first using browsers such as chrome, to something.html with something_files directory beside.

$ python2 webpage2html.py /path/to/something.html > something_single.html

But note that, the second method may not always work as expected, because there may be urls like //ssl.gstatic.com/gb/images/v1_c69d5271.png (from google index page), but the file is missing in Google_files directory saved by browsers.

Enable javascript, for example, save 2048 game page into a single html for offline playing

$ python2 webpage2html.py -s http://gabrielecirulli.github.io/2048/ > 2048.html

Dependency

BeautifulSoup4, lxml, termcolor(optional)

$ pip install -r requirements.txt

or install them manually

$ pip install lxml BeautifulSoup4 requests termcolor

I have tried the default HTMLParser and html5lib as the backend parser for BeautifulSoup, but both of them are buggy, HTMLParser handles self closing tags (like <br> <meta>) incorrectly(it will wait for closing tag for <br>, so If too many <br> tags exist in the html, BeautifulSoup will complain RuntimeError: maximum recursion depth exceeded), and html5lib will encode encoded html entities such as < again to &lt;, which is definitly unacceptable. I have tested many cases, and lxml works perfectly, so I choose to use lxml now.

The termcolor package is for colored log output support if you like.

Cases still does not work

browser side less compiling

The page embeds less css directly and use less.js to compile in browser. In this case, I still cannot find a way to embed the less code into generated html to make it work.

<link rel="stylesheet/less" type="text/css" href="http://dghubble.com/blog/theme/css/style.less">
<script src="http://dghubble.com/blog/theme/js/less-1.5.0.min.js" type="text/javascript"></script>

Thanks

Thanks lukin.a.i who submitted patch to fix not recognised css link (rel=stylesheet) issue

License

webpage2html use SATA License (Star And Thank Author License), so you have to star this project before using. Read the license carefully.

Reference

Java port of this project. https://github.com/cedricblondeau/webpage2html-java

afeique / compyler