Forum Archive

Hello,

Can anyone tell me how write a code to auto fill a login and password without selenium or mechaniz. This is to be ru in pythonista on an iPhone. Thanks

Are you trying to fill passwords on mobile apps, or on websites?

@space-dane, use the Pythonista WebView. It has full support for Javascript, frames etc.

Since Javascript is a pain to write and test, I wrote a JSWrapper that lets you manipulate js in Python. You can find it here.

With it, you can fill in the fields by name like this:

js = JSWrapper(webview)
js.set_field('username', 'your username')
js.set_field('passwd', 'your password')

And submit the first form on the page:

js.submit()

... or, if there are multiple forms, by name or id of the form element:

js.by_name('form1').submit()

When you get to a content page, you can pick out items by id, by name, or with XPath. A simple example of the last option:

js.xpath('head/title').value()

Or you can collect a list of repeating elements, e.g. name attributes of all input elements:

js.list_each('input/@name')

More complicated alternative collects a dictionary of several elements, e.g. different cells from the rows of a table:

cell_values = js.for_each('table//tr').map(
  key='td[1]',
  some_value='td[3]'
)
for key in cell_values:
  print(key, cell_values[key]['some_value'])

WebView is asynchronous, and the actual scraping on a page can start when the webview's webview_did_finish_load callback fires. As this is fired for every page, we need a kind of a state machine to keep track where we are in the flow of pages. To help with this, the file includes a ready-made delegate where you map urls to handler functions and manage the flow by always setting the next expected handler. Below is a "full" example for a hypothetical login & content scrape. Note the example on retrieving whole body HTML content, useful when developing the scraper, and that the present` on the last line is optional, depending on whether you want to watch things happening live or not.

import jswrapper

class DemoScraper(jswrapper.WebScraper):

  def __init__(self, webview):
    super().__init__(webview)
    self.url_map = {
      self.login_page: 'starting_url',
      self.content_page: 'pther_url'
    }
    self.handler = self.login_page

  def login_page(self):
    self.set_field('username', 'your username')
    self.set_field('passwd', 'your password')

    self.handler = self.content_page
    self.by_name('form1').submit()

  def content_page(self):
    print(self.xpath('body').html())
    print(self.xpath('head/title').value())

wv = ui.WebView()
ds = DemoScraper(wv)
wv.load_url('starting_url')
wv.present()

@mikael, I’m testing jsui but I have a problem with AttributeWrapper, it is not defined. Are you still developing it? or from where I can download the code. Thanks, very good job.

@augusto, I am just starting with it, and do not yet consider it fit for any kind of use. I am currently refactoring the initial exploratory version and moving it to a separate repository with a more sensible file structure.

That said, as an immediate fix, you can replace any mention of AttributeWrapper with JSWrapper, and I will let you know as soon as the new version is in some workable shape.

@mikael
For what its worth, you might want to look at
https://forum.omz-software.com/topic/1804/debugging-in-javascript-in-pythonista

Your JSWrapper is a lot nicer than what I had, but what you might find useful was the on_error logging -- basically having javascript errors automatically get logged back to the console (via a custom uri scheme that the webview is watching for). the silent nature of javascript errors on ios made developing a pain.

@JonB, thanks, I did remember your code from back when, and it is included in the latest version. It is simply a godsend.

@JonB, still, can be difficult to tell where the error really comes from, when you have these bits and pieces of eval_js flying around.