Panda Crazy Script for Panda's

[Jharkan]

The best way is to add a panda for a hit from that requester. Find out the requester ID. Go into the details of that panda and put the requester ID in the Requester ID field. Change the mode to Search and it should do a search on that requester until it sees a hit and grabs the first largest pay hit. You can switch back and forth between panda and search mode. The group id will get filled in when it find hits by searching.

Some people use the requester ID in the Add Job instead of a group id but the script uses the requester ID for the group ID which doesn't work. Have to fill in the requester ID for it to work and switch to search mode. The group ID should always be filled in with a group ID to work properly.

why does it need a group ID to work?

 

[Johnnyrs]

why does it need a group ID to work?

Group ID is only for panda's which the script is mainly for. The add Job button is only to add a panda. The search mode is just an option to make it easier to grab hits from requesters that have multiple panda's. When the script finds a hit with the search mode it fills in the group ID and does a regular panda. So the only way to use the search mode is to have a panda in the first place. You can get around it like I said by adding a requester ID instead of a groupid and then copying it to the requester ID field. No way to add a search job because it's slow and not as fast as a panda. I don't want to encourage people to have a lot of searches and then complain it's not grabbing anything. It's for special cases.

 

[Jharkan]

Group ID is only for panda's which the script is mainly for. The add Job button is only to add a panda. The search mode is just an option to make it easier to grab hits from requesters that have multiple panda's. When the script finds a hit with the search mode it fills in the group ID and does a regular panda. So the only way to use the search mode is to have a panda in the first place. You can get around it like I said by adding a requester ID instead of a groupid and then copying it to the requester ID field. No way to add a search job because it's slow and not as fast as a panda. I don't want to encourage people to have a lot of searches and then complain it's not grabbing anything. It's for special cases.

What exactly do you mean slow? What is slow about it?

 

[Johnnyrs]

What exactly do you mean slow? What is slow about it?

When doing a search the script sends a search request to mturk using up 1 cycle or 1 second. Then mturk sends the page showing all the hits. If there are hits the script then sends a panda to grab that hit taking another cycle or another second. That is 2 seconds. Also mturk has a delay when hits get seen on the search page. Panda's will grab a hit even before it gets a chance to show on the search page. Sometimes I can grab 10 hits or more with a panda even before hit monitor or hit scraper alerts me to it. So if a lot of people are using panda's it might not even show up on a search page so the search mode would see nothing. Use it for special requester's.

 

[Johnnyrs]

Today I released a new update (0.3.18) to greasyfork. Be sure to update it if greasemonkey/tampermonkey doesn't update it automatically. Should update it automatically sometime today. Be sure to backup your data by exporting it from the jobs menu at top so you won't lose anything if something goes wrong.

New Features:

• The time a hit was accepted will now show up in the accepted tab on the bottom.

Bugs Fixed:

• Fixed a bug where you couldn't search on mturk anymore with the search mturk button.
• Fixed a bug where it would not get out of ham mode when deleting a job in ham mode.

This was mainly a bug fix update and nothing major has been done.

 

[T. Leela]

First I just want to recommend a few things. Be sure to have Panda Crazy in it's own window and not a tab. It will go a lot faster and not limited to 1 second timers by the browser. Second be sure you really need a search mode for some because it does take longer to find the hits and then try to grab it. Search mode is best used for requesters who put out a lot of hits with different panda's or when they constantly change panda's. A panda will always be faster.

Two ways to find out which jobs are on and collecting. Click on the text at the top where it says "# jobs On". It will show a window of all the jobs collecting and you can stop stuff you don't want. Another place is in the Status Log on the bottom. It will show the elapsed time and the mode it is in. Now for the issues you are having I think the grouping might be causing those issues. I had problems with that when I added it in but thought I got everything sorted. There could still be a problem because a grouping can save delayed go ham mode which is probably causing the ham mode issues. The aqua color comes from creating groupings so something weird happened with those jobs while creating a grouping. You can see there are no stats for those jobs. I would suggest to delete those. Also if you can export all your jobs from the jobs menu to a file and email it to me so I can see what might be causing those issues I might be able to fix it for a future update. Thanks

Thanks so much, I did see a recommendation you posted re: putting it in its own window, I start out that way, but seem to open more tabs in the window without realizing, ill watch that. I use search for c-sats, pinterests, things with lots of hit ids but have them running only at certain times, and depending on what comes up might add a temporary panda. for it
I will check out the other things, I'm always adding and deleting and might be doing something weird. Also i will be sure I'm updated and if that fixed anything.

Also I had tons of pandas that I needed to delete (thanks for easier to access delete btw) and it definitely slowed things down, whether they were running or not, as I accidentally cleared them all out, and it is so much better now. (saw some recent posts.) Anyway thanks again, and I'll send an email, I know I need to organize things again,cluttered, repetitive mess.

 

[TurRaider8v]

Bulk deleting jobs using the Search is not working

 

[Johnnyrs]

Bulk deleting jobs using the Search is not working

Thanks. Didn't see that problem before. It deletes 1 only and then gets an error. I am actually redesigning the search jobs and search mturk buttons so they will be faster and show all jobs. The multiple delete will be working too. Still working on it so not sure when I'll have it finish just yet.

 

[Jharkan]

how exactly do I make a small donation to you, as a thanks for contributing to the community, and for the coding which you've done?

 

[ScrapingForQuarters]

On a captcha alert ... does this mean you tried to accept a HIT and it failed?

If so, do you, or could you in some way, indicate which HIT or search "failed?"

maybe a special color AND/OR and a note in the accepted log?

In some cases, I know that in a certain time window, if I get aced out due to a Captcha, I might as well turn a PandA off, cause there is unlikely to be any second chances.

But if I have mulitple PandAs and searches going, and I get a Captcha alert, I have no idea which one caused it.

Thanks for all you do.

 

[Johnnyrs]

how exactly do I make a small donation to you, as a thanks for contributing to the community, and for the coding which you've done?

I am not asking for donations but if you really want to and have some money to spare I do accept Amazon Gift Cards. You can send it to my email in the spoiler below:

Spoiler: Email Address

Log in or register now. to view Spoiler content!

 

[Johnnyrs]

On a captcha alert ... does this mean you tried to accept a HIT and it failed?
If so, do you, or could you in some way, indicate which HIT or search "failed?"

A captcha alert means the script tried to accept a hit and it was asked for a captcha which it can't do. It will pause the script so you can try to accept a hit and fill in the captcha. After you can unpause the script and it will accept hits normally. This only happens for people who are not masters. Doesn't matter which hit a captcha was received because it will show up on any hit. Some people have a hit they can accept fast so they can fill in the captcha, return the hit and then continue the script. Other people turn off the captcha alert in the general options and have another script handle the captcha because it sometimes will happen after a certain amount of hits. I don't get captcha's so I can't test or tell you the best way to handle it.

 

[ScrapingForQuarters]

A captcha alert means the script tried to accept a hit and it was asked for a captcha which it can't do. It will pause the script so you can try to accept a hit and fill in the captcha. After you can unpause the script and it will accept hits normally. This only happens for people who are not masters. Doesn't matter which hit a captcha was received because it will show up on any hit. Some people have a hit they can accept fast so they can fill in the captcha, return the hit and then continue the script. Other people turn off the captcha alert in the general options and have another script handle the captcha because it sometimes will happen after a certain amount of hits. I don't get captcha's so I can't test or tell you the best way to handle it.

Thanks, but I know all that stuff ... since I get them I know how to handle them (with aid of
https://greasyfork.org/en/scripts/10300-captcha-clear
) and then clicking the unpause PC button.

My question was about whether PC can record (say in the Accepted Log) and tell me which HIT it was attempting to accept when it happened. In some cases this is very useful information since if you miss the boat on accepting a HIT due to a Captcha https://en.wikipedia.org/wiki/Coitus_interruptus, you might as turn that PandA off for the rest of the day. As you have pointed out to us many times, active PandAs are key resources and an allocation problem, so turning off useless ones ASAP is key to greater $$ success.

The one I'm thinking of pays much more than the batch HITs I'm doing while I wait to find out if I got lucky that day. The timing is somewhat uncertain as to when it's posted. But I've had cases that I missed it due to a Captcha, and it's really irritating!

I will be happy to test this, should you choose to work it in, since you don't get them. But the code is there to detect them and give us the unique crow sound effect that so amuses children of Turking parents, so I'm hoping it's a relatively trivial enhancement.

However for mucho extra credit ....
and this is probably much harder/involved, the whole issue of knowing how many HITs till you will hit one (even if it's just your best esitmate) would be very useful. This is somewhat hepful
https://greasyfork.org/en/scripts/907-captcha-counter
though truth be told when doing batches I don't see the info. I have another script from
mmmturkeybacon https://greasyfork.org/en/scripts/3105-mmmturkeybacon-scroll-to-workspace
that I think typically hides (skips over) the count.

For those of that use your Helper Scripts, maybe something useful could be added (as an option) in visually that counts the number of HITs accepted since the last Captcha. Even just the count accepted since last captcha would be MUCH better than nothing. This could be printed in your depiction of my queue as (or ON) the first line (since one's queue during batches is likely much longer than the space allocated for that display and printing the number of each line while interesting, doesn't emphasize that the "hiden" end of queue is near a "braking point.") ... or maybe on the "title line?"

I think most of get 35 HITS between them, though I've heard claims of some much more often for unknown and unsubstantiated reasons. I think there was an option somewhere such that you could set your "default number" (e.g., 35) for the scripts. Anyway, the ultimate is a countdown to 0, but even just a countup (with reset on the Captcha) would still be a big win.

 

[catnapped]

Search seems to be tacking on the "&isSelctedBySearchExperiment=" part from the URL/preview (right after the group ID) when it finds something. Haven't caught anything lately so I don't know whether or not that's breaking it.

 

[Johnnyrs]

Search seems to be tacking on the "&isSelctedBySearchExperiment=" part from the URL/preview (right after the group ID) when it finds something. Haven't caught anything lately so I don't know whether or not that's breaking it.

I just tested a search mode on a ibotta panda I had. It found one and accepted it. It picked a bad .01 cent products but it did accept it. The "&isSelctedBySearchExperiment=" seems to not cause problems with a panda. The search url is created by the script so it doesn't have the added stuff in the url. Even a pantha seems to be working also.

 

[Johnnyrs]

Thanks, but I know all that stuff ... since I get them I know how to handle them
My question was about whether PC can record (say in the Accepted Log) and tell me which HIT it was attempting to accept when it happened. In some cases this is very useful information since if you miss the boat on accepting a HIT due to a Captcha https://en.wikipedia.org/wiki/Coitus_interruptus, you might as turn that PandA off for the rest of the day. As you have pointed out to us many times, active PandAs are key resources and an allocation problem, so turning off useless ones ASAP is key to greater $$ success.

The one I'm thinking of pays much more than the batch HITs I'm doing while I wait to find out if I got lucky that day. The timing is somewhat uncertain as to when it's posted. But I've had cases that I missed it due to a Captcha, and it's really irritating!

I understand what you are saying about a counter that could easily be added once I find a good place for it. I could further use that information to do something similar like other scripts. Having it actually do something with the counter info would be a more advanced feature for the future.

I don't understand why you would need to stop a panda after getting a captcha. If a captcha is usually done after a certain amount of hits then it shouldn't matter which panda it was going to grab unless those are penny hits and it would be better to go after the higher paying hits. I don't think a captcha is tied to a specific panda. The panda link that it got the captcha on should be in the window that comes up after the alert. Maybe I could label the link better for example instead of "Go to Mturk Page" it could be: "Go to panda link" with more information about that panda. Maybe have an option to have the link go to a throw away panda so you can click on it, fill in the captcha, return it, then unpause the script. If a window doesn't show up with a link after an alert then that's a bug I need to address.

 

[ScrapingForQuarters]

I understand what you are saying about a counter that could easily be added once I find a good place for it. I could further use that information to do something similar like other scripts. Having it actually do something with the counter info would be a more advanced feature for the future.


I don't understand why you would need to stop a panda after getting a captcha. If a captcha is usually done after a certain amount of hits then it shouldn't matter which panda it was going to grab unless those are penny hits and it would be better to go after the higher paying hits. I don't think a captcha is tied to a specific panda. The panda link that it got the captcha on should be in the window that comes up after the alert. Maybe I could label the link better for example instead of "Go to Mturk Page" it could be: "Go to panda link" with more information about that panda. Maybe have an option to have the link go to a throw away panda so you can click on it, fill in the captcha, return it, then unpause the script. If a window doesn't show up with a link after an alert then that's a bug I need to address.

I didn't realize clicking "Go to Mturk Page" was supposed to do that. I think I tried it way back when and I'm pretty sure it didn't do that, so I haven't tried it since once I found a different "default" strategy to resolve captchas. I don't recall which page it went to, it might have been even just mturk.com, but in any case, it wasn't helpful. But I agree a better label is advised. Note that if you're doing Google's Zoltars, though you could type in the captcha string and that's useful, I suspect you won't actually be able to complete that HIT instance as GOOG requires you do to them in order. But for your average "high value" PandA, it gives you some shot to grab one, though again if they're doling them out in SMALL batches, you're likely S-O-L for that "bunch" due to high demand though certainly you wouldn't want to stop the PandA in that case as you might get in on the next "small bunch."

As for stopping, my case was a very restricted "high value" and "high demand" PandA. They used the same GID every day, but If you don't get it on the first try, all the available slots (1 or 2 is my guess) will be gone for that day by the time you can type any captcha string, and it's wait till tomorrow. But this is why being better able to predict if you're near a Captcha becomes an issue. If you're doing a batch and "in the neighborhood" of the time it will HIT, and know you may ALSO hit a Capthca, you might well turn off collection for that batch for a while till time time interval has passed. You may well have a queue full that will keep you busy while you wait it out, though of course you're risking not filling your queue while they're still available.

 

[Johnnyrs]

As for stopping, my case was a very restricted "high value" and "high demand" PandA. They used the same GID every day, but If you don't get it on the first try, all the available slots (1 or 2 is my guess) will be gone for that day by the time you can type any captcha string, and it's wait till tomorrow. But this is why being better able to predict if you're near a Captcha becomes an issue. If you're doing a batch and "in the neighborhood" of the time it will HIT, and know you may ALSO hit a Capthca, you might well turn off collection for that batch for a while till time time interval has passed. You may well have a queue full that will keep you busy while you wait it out, though of course you're risking not filling your queue while they're still available.

I'll make sure that link will go to the panda link and label it better with more detail. I kind of just used the same coding from the log off alert window and that goes to your queue page because it's the easiest way to log back in.

I agree a captcha counter would definitely help out in the circumstance you described. I'll see what I can do. The biggest decision is where I should put it. Once I get the counter working then I can do something with it in the future which could make things easier.

 

[ScrapingForQuarters]

Small nit on searches: I used Add Job with a requester ID instead of a real GID (which I didn't have.) I went to details and copied over the "fake GID" into the requester ID field, but forgot to delete it from the GID field. But the system left the "fake GID" there even though it sould be easy to tell it's not a GID (I think the lengths are quite different.) So intead of a [] being shown where you click or double click it has something that if you double click for example results in "Invalid URL parameters." Actually, what would be stellar if PC would realize it must be a RID and put it in the right place and leave the GID field clear. Also, it would be really nice if in this case the double click would produce on single or double click a search page/link, not an previewandaccept page/link. I see a sample turkoptigon page produces https://www.mturk.com/mturk/searchbar?requesterId=A2IFZMK47ZIZ8X&selectedSearchType=hitgroups

 

[Johnnyrs]

Small nit on searches: I used Add Job with a requester ID instead of a real GID (which I didn't have.) I went to details and copied over the "fake GID" into the requester ID field, but forgot to delete it from the GID field. But the system left the "fake GID" there even though it sould be easy to tell it's not a GID (I think the lengths are quite different.) So intead of a [] being shown where you click or double click it has something that if you double click for example results in "Invalid URL parameters." Actually, what would be stellar if PC would realize it must be a RID and put it in the right place and leave the GID field clear. Also, it would be really nice if in this case the double click would produce on single or double click a search page/link, not an previewandaccept page/link. I see a sample turkoptigon page produces https://www.mturk.com/mturk/searchbar?requesterId=A2IFZMK47ZIZ8X&selectedSearchType=hitgroups

Yes I know I need to fix these type of errors. I just don't want to make it easy to add a search job because it will be slower than a panda. I've had a few people asking if they should change everything to a search. I need to put some error checking when a group id is not correct and not let someone panda it. I can change the way the single and double click would work too. You can always add any panda with the add job, change the requester ID to the one you want, switch it to search mode and set a friendly text for it. Once it finds a hit it will fill in all the details including the group id so doesn't matter what it started with. The fake gid trick works but it can cause problems but I don't want to stop it just yet. I'll address the search problems in the future sometime.

I have done some work on the captcha counter and hope to have an update tomorrow after some testing.