pip install gestures, WKWebView, UI animations and more

Anxietier

Hi, I tested one demo code in readme of pythonista-wkwebview

class MagicWebView(WKWebView):
  
  def on_magic(self, message):
    print('WKWebView magic ' + message)
    
html = '''
<body>
<button onclick="window.webkit.messageHandlers.magic.postMessage(\'spell\')">
Cast a spell
</button>
</body>
'''

v = MagicWebView()
v.load_html(html)

Well it works good, but when I try to load html file, the 'on_magic' doesn't work, here's my code in py file

class MagicWebView(WKWebView):
	def on_magic(self, message):
		print('WKWebView magic ' + message)

v = MagicWebView()
v.present()
# v.load_html(html)
v.load_url('test.html', no_cache=False, timeout=5)
v.clear_cache()

I don't know how to fix it

cvp

@Anxietier try

import os

w = WKWebView()
f = os.path.abspath('test.html')
print(f)
w.load_url(f)
w.present()

mikael

@Anxietier, I see the online documentation covers this poorly. Here’s the method docstring:

Loads the contents of the given url asynchronously.

If the url starts with file://, loads a local file. If the remaining url starts with /, path starts from Pythonista root.

For remote (non-file) requests, there are two additional options:

Set no_cache to True to skip the local cache, default is False
Set timeout to a specific timeout value, default is 10 (seconds)

mikael

@Anxietier, you can get this and similar documentation on other methods using the help function, e.g. in Pythonista console:

>>> import wkwebview
>>> help(wkwebview.WKWebView.load_url)

Anxietier

@cvp
thanks, but it seems doesn’t work, here’s my test code

file_name = 'test.copy(2).html'
file_abs_path = os.path.abspath(file_name)
url_file_name = 'file://' + file_name
url_file_abs_path = 'file://' + file_abs_path

with open(file_name, 'r', encoding='utf8') as f:
	html = ''.join(f.readlines())

v = MagicWebView()
v.present()

# v.load_html(html) # can load but button doesn't work

# v.load_url(file_name, no_cache=False, timeout=5) # can load but button doesn't work

# v.load_url(file_abs_path, no_cache=False, timeout=5) # can load but button doesn't work

# v.load_url(url_file_name, no_cache=False, timeout=5) # would let ista crash or load faild

# v.load_url(url_file_abs_path, no_cache=False, timeout=5) # doesn't load

v.clear_cache()

Anxietier

@mikael
I’m not sure if I described my issue clearly, and here’s my another test code, perhaps that would be clearly

file_name = 'test.copy(2).html'
file_abs_path = os.path.abspath(file_name)
url_file_name = 'file://' + file_name
url_file_abs_path = 'file://' + file_abs_path

with open(file_name, 'r', encoding='utf8') as f:
	html = ''.join(f.readlines())

v = MagicWebView()
v.present()

# v.load_html(html) # can load but button doesn't work

# v.load_url(file_name, no_cache=False, timeout=5) # can load but button doesn't work

# v.load_url(file_abs_path, no_cache=False, timeout=5) # can load but button doesn't work

# v.load_url(url_file_name, no_cache=False, timeout=5) # would let ista crash or load faild

# v.load_url(url_file_abs_path, no_cache=False, timeout=5) # doesn't load

v.clear_cache()

mikael

@Anxietier, I will have to look at this, maybe something has changed in recent iOS versions.

Meanwhile, this:

html = ''.join(f.readlines())

... looks funny, breaking and joining with no change. Just html = f.read() should work fine.

mikael

@Anxietier, I placed this script:

import wkwebview

class MyWebView(wkwebview.WKWebView):
    
    def on_magic(self, message):
        print('WKWebView magic ' + message)

wv = MyWebView()
wv.present('fullscreen')

wv.load_url('load_url.html')

... and this HTML as the file load_url.html in the same directory:

<body>
<button onclick="window.webkit.messageHandlers.magic.postMessage('spell')">
Cast a spell
</button>
</body>

And the button works fine. Note that I removed the backlashes in the onclick handler as escaping not necessary in a file.

I suspect you might have some kind of non-printing character issue.

stephen

@mikael said:

@Anxietier, have to look at this, maybe something has changed in recent iOS versions.

Meanwhile, this:

html = ''.join(f.readlines())

... looks funny, breaking and joining with no change. Just html = f.read() should work fine.

much cleaner than i would of. lol


html = ''.join([line for line in f])

Anxietier

@mikael
oh I see, I usually use r_string in py, so I didn't notice \' in doc_string
now it works fine
thanks for help

Anxietier

hi, im back again :D
we know that some browser has non-pic mode (i dont know right name,just like that mean), is it possible to add that feature to wkwebview? for some reason, i just need the full html content (dom-tree maybe?)

mikael

@Anxietier, if you mean ”Reader Mode”, WKWebView does not include it. But you can use a SafariViewController, see @cvp’s code around the middle of this thread.

JonB

If you just want to parse the HTML, use requests. Unless you need JavaScript to run I guess.

@mikael isn't there a delegate that would let you filter by uri? I forget if the should_load_uri or whatever gets called for <img>'s -- in which case he could just return false for non-html urls.

mikael

@JonB, no it does not. Would be good to understand the need a bit more before trying to bend WKWebView to it. For the most part, just fetching the page with requests, then filtering imgs and javascript before giving it to WKwebView might work just as well.

Anxietier

@JonB
yeah i need js in script tag to run. some requests is not easy to fake them you know, so i was wandering if i can let webview run in 'read mode' to simulate 'selenium', cause some imgs spend many data traffic, and i dont wanna take double traffic for one img

Anxietier

@mikael
thanks, it is helpful, i would try to understand them.

mikael

@Anxietier, if web scraping is your thing, then the SafariViewController will not help you. WKWebView is the best tool around for Pythonista, as far as I know.

mikael

@Anxietier, hey, there’s something called content rules that can be configured to just stop loading images, see here. I will give it a try a bit later.

Anxietier

@mikael
well, theres an open source android app names 'H-viewer', using regex&selector to select out imgs, and re-layout them, so that we can just watch imgs. also, theres an app in appstore names 'yealico', it has similar function.
I was trying to fake(reproduce?) that by python in ista, I found one way to do that, 'use two wkwebviews, only present one, the other one load url in background, when it did finish load, use js selector to get imgs url, then show imgs on presented webview', im not sure will background-webview cost data-usage, but I dont want to cost double 'img data-usage'

mikael

@Anxietier, the content rules, which might help or not, require at least two completion blocks, doable but unfortunately I do not have the kind of time I had earlier to spend on these projects.

I would be happy to provide background support if you want to give it a go. I would expect it to be mostly versions of code already used in this component.