Skip to content

Save text captcha details to file when detected

m3ownzm3ownz
edited August 2012 in Feature Requests
The feature to find and skip text question captchas when in auto mode (which is how gsa is run 99% of the time i guess) is a good feature to save on captcha costs.

I understand that if i set it to 'if all else fails, ask user' it will show these questions, and save answers in the dat file if the subsequent submission is succesful. This is also a nice feature, but the prospect of sitting in front of it for hours answering questions doesnt appeal

If would be very useful to save text questions to a file (csv?) that could be latter opened and answered in one sitting (or sent to others to answer for $), then added to the dat file, either manually or via a [import text question and answers] option.

I am happy to spend time and money doing this, and happy for the results to be pooled for the benefit of other users, but im not likely to site in front of a running copy of gsa for hours typing into popup boxes!

Comments

  • s4nt0ss4nt0s Houston, Texas
    That's an awesome idea. :P
  • BytefakerBytefaker
    I think perhaps it's a problem with the session lifetime which differs from site to site. It isn't possible to answer the questions for example 10 hours later if the session lifetime is set to 4 (the webmaster can configure this).
  • SvenSven www.GSA-Online.de
    sounds useful. I can add something like that in advanced options
  • m3ownzm3ownz
    @Sven, great, i look forward to seeing it implemented when you get the time

    This purpose isnt to use the answers in the same run, just a method to grow the dat file for latter use, so if the same question is encountered again in future it can be auto-solved.
    Obviously some questions like "what colour is last word in our logo" can not be solved like this, but others like "which is a fruit: bag, apple, jump" can be.
  • BytefakerBytefaker
    edited August 2012
    Aaaah I got it. Sounds very useful, and if we work together we should fast have a comprehensive DB of questions/answers. I missunderstood it and thought you mean solving normal captchas a little bit later but all together. This makes sense of course.
  • AlexRAlexR Cape Town
    Agreed. This is a great idea....let's see it happen! :-)
  • GiorgosKGiorgosK Greece
    +1 for this, I have already send such a feature request some time ago
    I would love it if successful answers of mine go into the .dat file or even to other user's .dat files
    so we can all benefit from that

    lots of times I sit on my computer and answer question and I do tend to find
    lots of similar questions ...
  • SvenSven www.GSA-Online.de
    I have just added this for the upcoming version. All unanswered questions get saved to a file if wanted.
  • m3ownzm3ownz
    Fantastic, i have a few questions but will wait to see how its implemented before i ask.
  • SvenSven www.GSA-Online.de
    It's a plain text file, nothing else and I don'T think its worth adding some editor around it as most people will not use it.
  • m3ownzm3ownz
    I agree completely.  I expect i can convince a few heavy users to pool on this one im sure, at least in collecting the questions for me to get answered.

    How would i go about using these question/answers in later runs? Just paste them into the .dat file?
    Related, if a field is marked as %question% in an engine, and it is encountered on a site, is the .dat file always checked for an answer before the site is skipped?
  • SvenSven www.GSA-Online.de

    I just made up my mind and added a small gui to edit the questions/answers in the new version. There you can add your filled answers to the DB.

    And yes that dat file is used before querying the user about anything.

  • BytefakerBytefaker
    edited August 2012
    Good questions m3ownz. Beside this: If a editor will save time, I'm sure someone will code one. It should be only possible to verificate the entered answers if they are really correct.

    EDIT: It sounds great Sven :)
  • m3ownzm3ownz
    @Sven great, i will make good use of it and encourage others to do so.
    Just to clarify:

    And yes that dat file is used before querying the user about anything.

    It will also check that file for an answer before skipping a site if the [if all else fails...] option is turned off correct? As i never have that option on unless im debugging.

    @Bytefaker cant really see a way of checking that - if an algorithm could know if the answer is correct we wouldnt need this function in the 1st place. Also checking the answer against submission status would likely create a bunch of false negatives due to other errors.
    Its still a numbers game, building this database will help, but it will never be 100% correct for every site.
  • BytefakerBytefaker
    edited August 2012
    I see your point m3ownz. But it would be a possibility to let different people solve the questions and compare the answers. The more have the same answer the bigger chance that it is correct. Additionaly when multiple people run them and compare the results again, it would improove the quality again.
  • m3ownzm3ownz
    Yeah, i understand. Sven probably already does some checking if they are sent to him.

    If this gets a lot of response i will look into setting up a central database on a server somewhere that can pool and compare results for contributors.
  • SvenSven www.GSA-Online.de

    >It will also check that file for an answer before skipping a site if the [if all else fails...] option is turned off correct? As i never have that option on unless im debugging.

    I don't really get what you are asking. It is like this: Before the user is queried with anythign, all other filters have already been checked against the site (e.g. content filter, url filter, PR filter, country filter, OBL filter and so on). Else it would be really stupid to query the user and later skip that site due to some other things. The recorded answers get also checked before asking the user of course.

  • m3ownzm3ownz
    Thanks Sven, I was referring to the inbuilt feature that skips sites with text captchas if the ask user option is disabled. It shows in the log as 'Text Catcha Found and Skipped', and i was unsure if it checked the dat file in this instance, but i can test this myself.

    Thanks for the useful update, appreciated.
  • BytefakerBytefaker
    I'll do a manual answer-session too after my "search site lists" is finished and send the results to Sven...

    @Sven: Thanks for adding it.
  • SvenSven www.GSA-Online.de
    Oh thats what you meant. Yes it is using that file anyway even if "ask user" is not chosen because if it finds it, the user is not asked.
  • m3ownzm3ownz
    Great, thanks a bunch
  • BytefakerBytefaker
    Would it be perhaps an idea to add access to this? 
    http://sickmarketing.com/forum/showthread.php?t=4765

Sign In or Register to comment.