Let me start out by noting that tonight’s update does not contain the solution I’m talking about in this post, but I am working on the solution right now and it will be available very soon. Tonight’s update contains a fix for the search page command
There has been a lot of frustration lately about the utter failure that has become of decaptcher and the high prices of ITT. I’ve been thinking on how to provide a sensible solution in a way that would quick and painless and not clutter up the interface of ubot with a bunch of redundant commands. Looking through the various captcha services, I noticed something interesting. Many of them, if not most of them, have a form based system, in which you navigate to a web page, upload the image file and an api key or something of the sort, hit submit, and it sends you back a solution. This forces me to think of how easy it would be to actually make a captcha reader in ubot. all one would have to do is utilize the “save element image” command and simply make a script to fill out the form.
The only problem is that in order to do this, you would have to actually leave the page that the captcha is on, which would lose the captcha, and thus lose the purpose of decoding it. It’s kind of a shame, because it would be pretty easy to use a number of new services if there were only a way to navigate to a different page without losing your place on the page you’re decoding. This leads to my proposal.
What I’d like to do is have a command that can contain other commands (the way that a loop, for instance, does). When the script gets to this container command, it will load a new webbrowser in a new window. Any command that is put into the container will be executed in the new webbrowser. This means you can run an entire mini script completely separately without disturbing your main script. Besides any number of other uses that I’m certain people will come up with, one fantastic use for this will be, of course, uploading a captcha image to a captcha reading service.
What I’m essentially doing is enabling ubotters to make their own captcha solving commands using any service they’d like (as long as they have a web based system, which many do). I think this will be the least cluttered, least limiting solution to our captcha solving woes, and at the same time, it will enable our script to do more things that I probably haven’t even thought of. I’ll be working on this and I should have an implementation within a couple days. Let me know your thoughts.