Solve Manual Captcha with ChromeDriver on Demand (rebased... again (... and again)) #1582
Conversation
|
Was thinking -- is it possible to add the functionality of swapping the captcha-locked accounts after a (defined) period of time? This will allow the map users to solve it initially if they are online/around, and then just swap the account out to continue the scan (in lieu of -bh) |
|
How do I install ChromeDriver in the Python scripts directory? I have ChromeDriver installed via pip but that doesn't seem to be enough |
|
I have used this PR, it works very well. It just needs a bit more documentation on how to get it running for newbies like me. Other than that i am impressed. Thank you for doing this. |
|
Don't know if this is the appropriate place, but I know there are a lot of people (including me) that want to implement this on their scanner but have no clue how to actually do this. Im a relative noob with some computer knowledge, but how do I go ahead and add this to my program? Do I replace files or something? Help would be greatly appreciated, sorry for my noobish question! |
|
@Nonkel I got this working by doing a pull to x:/.../PokemonGoMap/PokemonGo-Map/ with below commands: git fetch origin pull/1582/head:ManualCaptcha Then you just need to download chromedriver and paste it into your C:\Python27\Scripts directory. I'm note sure if this is the correct way to do it, or if it's inappropriate to post instructions like this here (if so, someone please let me know or remove the post), but this is what I did to get it to work. |
|
Tested and work 100% not any problem. |
|
if you tested and like it I only ask for one thing, you go to my patreon an support me .... just kidding
same apply for every PR you test |
|
@Thyllz thanks for the info, but you lost me in the first sentence. What is a pull and how to do this? I tried Google ofcourse, but unfortunately it looks like Chinese to me... Lot of people in the same boat as me, seems there are solutions but you need deeper knowledge on how to install these. Can you or someone make a small step by step guide for dummies? ;-) |
|
@Nonkel 1st- install the default branch after dat, git fetch origin pull/1582/head:PR#-manualcaptcha |
|
@ic4roswings can you made for mozzila too,using geckodriver? |
|
@Nonkel
That should be it. You could find me on Discord (same username as here) if you need anymore help. Hope this help. |
|
Is it possible to use this PR for some workers and still use 2captcha for other workers? or is the 2captcha service completely removed if I install this? |
|
Will this only appear on the machine running the server or will it be presented to anyone logged into the map site? |
|
@chromer90 only on server. |
|
@m3g4mInD and @Thyllz thanks, got a lot of stuff done. But only 1 detail missing. How do I install a copy of the develop branch? |
|
@Nonkel for linux For windows git clone https://github.com/PokemonGoMap/PokemonGo-Map.git |
|
@m3g4mInD followed your steps to get a copy in a Github folder on my hard drive. Last step must be to fire up the program. But can't work it out. So now I have the original scanner installed. If I click on the desktop shortcut, it starts up the original. I have a Github directory with this other build in it. How do I start this build? Sorry for all the questions and thanks for all the help so far. Really want to start catching em all again... ;-) |
|
@Nonkel go to dat path and open cmd and run python runserver.py |
|
@Nonkel Glad I can help out :) The Documentation and basic tutorial can be found here: https://pgm.readthedocs.io/en/develop/index.html The documentation have helped me a lot. |
|
how do i run it in ubuntu vs? |
|
I'm using this for a while now and there's 1 thing I'm bothered about. If an account encounters a captcha and you manually solve it, it scans the previous position with empty and scans a new location. Can you fix that behavior? |
|
working great for me :-) |
|
New to this whole git thing but I think it needs another rebase due to #1628 |
| response = api.verify_challenge(token=captcha_token) | ||
| if 'success' in response['responses']['VERIFY_CHALLENGE']: | ||
| status['message'] = "Account {} successfully uncaptcha'd".format(account['username']) | ||
| log.info(status['message']) |
There was a problem hiding this comment.
Lines 642-645 should be reinserted here:
scan_date = datetime.utcnow()
# Make another request for the same coordinate since the previous one was captcha'd
response_dict = map_request(api, step_location, args.jitter)
status['last_scan_date'] = datetime.utcnow()
| response = api.verify_challenge(token=captcha_token) | ||
| if 'success' in response['responses']['VERIFY_CHALLENGE']: | ||
| status['message'] = "Account {} successfully uncaptcha'd".format(account['username']) | ||
| log.info(status['message']) |
There was a problem hiding this comment.
Lines 642-645 should also be reinserted here:
scan_date = datetime.utcnow()
# Make another request for the same coordinate since the previous one was captcha'd
response_dict = map_request(api, step_location, args.jitter)
status['last_scan_date'] = datetime.utcnow()
|
@Hydrargyrum-Wells this pr is for solving the captcha not for rescanning on 0/0/0 which is already being done in -speed |
|
I'm aware. Everytime @Artifice merged one of his speed scanning PRs, you would need to rebase in part because your changes to The automated diff check fails to recognize that your PR moved the pre-existing captcha-handling code into an I don't know enough about the code to tell if those lines should really be there; I don't know if your omission of them actually makes the program better. However, such an omission seems irrelevant to the much-needed manual captcha-solving this PR implements. |
|
I agree with those 4 lines @Hydrargyrum-Wells is talking about. Further, I'd like to recommend the following change to manual_captcha.py to avoid the initial large window: from selenium.webdriver.chrome.options import Options def captcha_verifier(captcha_url, status): |
Added Manual Captcha to commited PR 1566 Travis Added Manual Captcha to commited PR 1566 - Travis Added Selenium to the requirements Added Chrome verifier Removed Chrome Verification Travis Travis Avelarry's whim Avelarry's whim 2 I hate you travis Travis are you happy now ?
| @@ -64,7 +64,8 @@ def jitterLocation(location=None, maxMeters=10): | |||
|
|
|||
| # Thread to handle user input. | |||
| def switch_status_printer(display_type, current_page, mainlog, loglevel): | |||
| # Disable logging of the first handler - the stream handler, and disable it's output. | |||
| # Disable logging of the first handler - the stream handler, and disable | |||
| # it's output. | |||
There was a problem hiding this comment.
If we're being picky, it should be "its".
There was a problem hiding this comment.
I have used this pr for quite a while, and overall I think it's a good boon to the project. However, I think it also falls short on some important areas:
It works quite well if you have few capthas you need to solve and a fast computer. However, if you don't have a lot of cpu/gpu power, and a lot of captchas, the constant opening of new chromedriver processes will grant the whole process to a halt. I don't think this is acceptable tbh, but I suspect this issue is not noticeable if you just test the code on a 'good' system. To overcome this, it would be quite easy to add some kind of cap to number of open chromedriver processes.
Also a minor addition: it would be nice to randomise the location of each new window, as they tend to pop up in front of the one you have in focus.
|
Using database to store : data for captcha url | status | captcha response | date That way all captcha solving solutions would be compatible. |
|
@springjools I guess I disagree. There's a 60 second timeout on the chrome windows. If you have too many captchas to clear them in 60 seconds then manual captcha solving is relatively useless anyway. I don't think a limit to the number of windows is bad -- but I don't think it should keep this PR from getting merged. (Plus I don't know how to code that kind of limit -- but maybe @ic4roswings does.) |
|
@averylarry I understand what you're saying, but I'm arguing that if you have more than, say 5, windows open, you won't have the time to solve all those captchas anyway, so there would not be a huge idea of opening more windows. |
There was a problem hiding this comment.
@springjools @averylarry This PR needs a queue for manual captchas.
- Pop a single window at a time off the queue, the user won't be filling in more than one screen at once.
- Spawn all of the windows at the exact same location, so the user can keep the mouse in the right spot for the captcha.
- Add a nice warning in console when queue size > 10 to 20.
After that, a final review and you've got my +1.
|
I agree with this.
Although I think the limit of open windows should be configurable.
I'd prefer to have a maximum of 4 windows open at once (I'd go higher but
I'm limited by display resolution), that way I can tile them and solve them
at the same time.
The timeout should probably be configurable as well, I increased it to 2
minutes for me as I frequently have to solve multiple and the default of 1
minute wasn't giving me enough time.
@springjtools seems to not understand how the timeout works
the timeout is for *each* chromedriver as it's starting a new one for each
captcha, so it's fairly easy to end up with 20+ windows open at once
On 6 Jan 2017 20:54, "sebastienvercammen" <notifications@github.com> wrote:
*@sebastienvercammen* requested changes on this pull request.
@springjools <https://github.com/springjools> @averylarry
<https://github.com/averylarry> This PR needs a queue for manual captchas.
- Pop a single window at a time off the queue, the user won't be filling
in more than that one screen.
- Spawn all of them at the exact same location as well, so the user can
keep the mouse in the right spot for the captcha.
- Add a nice warning in console when queue size > 10 to 20.
After that, you've got my +1.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<https://github.com/PokemonGoMap/PokemonGo-Map/pull/1582#pullrequestreview-15564603>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ADEjAE-pElc2nQ_2z6CKn_x4fafitP0Cks5rPqoPgaJpZM4K75hR>
.
|
|
@Phoenix09 I think if a queue gets used, then there doesn't need to be any timeout. Add a check_login process into the manual captcha code and leave the window (or windows) up indefinitely. |
|
Please fix and merge this already....it's taken so long. Also, ran this for quite some time, worked great. |
|
@lluisd Github is not customer support. Join the Discord's #help channel if you have questions. |
|
Any chance this is getting rebased again, or are we waiting on queueing and consensus with how to proceed? |
|
Closing, bookmarklet based solving already implemented. |
Provides a way to solve captchas without 2captcha payed service
Description
After 2 consecutive 0/0/0 Returns checks the account in order to obtain the URL of the captcha if it exist opens a Chrome instance to solve the captcha It waits 30 sec or Until the captcha was solved to retrive the information
EDIT: CHROMEDRIVER IS REQUIRED, IN THE PATH, IT CAN BE NEXT TO YOUR RUNSERVER.PY
EDIT2: The way it works is by using the same tag as the 2captcha solving -cs or captcha-solving: true in the config.ini file without the 2captcha api key you'll be prompted with chromedriver
Motivation and Context
Some people like myself can't make a payment in the 2captcha website
It solves https://github.com/PokemonGoMap/PokemonGo-Map/issues/1470
How Has This Been Tested?
Running an instance waiting for the captcha to raise the script itself was tested with at least 50 captcha accounts that I Have
Screenshots (if appropriate):
Types of changes
Checklist: