Panda Crazy Script for Panda's

[vburnin8tor]

The fetches thing might be a bug I fixed in a future update but I will recheck this. For auto collects there are a few reductions the extension makes but that is on the search side when it finds a lot of HITs all at once. The amount of panda's it creates is limited so it won't slow the panda's down.

Is it possible to add a tooltip similar to "Stopped for Once"?

 

[turkinthailand]

In PCS, is there any way to search for a particular RID in my list? There's one I want to remove but the text name that comes up in PCM doesn't match any text name in PCS, so I'm a bit stuck how to find it.

 

[Johnnyrs]

In PCS, is there any way to search for a particular RID in my list? There's one I want to remove but the text name that comes up in PCM doesn't match any text name in PCS, so I'm a bit stuck how to find it.

You can use the list button at the top which will show you all triggers added. The RID or GID is displayed in the 'Trigger ID or Term' column. The search text input at the top only searches on the title but I'll add ID search later.

 

[Dabeech]

Is it possible to have it grab any hit I qualify for over a certain dollar amount? I’ve tried the search and it requires info like requester or task name that I obviously wouldn’t know. I’ve searched and tried messing around and can’t figure it out. Thanks!

 

[turkinthailand]

Is it possible to have it grab any hit I qualify for over a certain dollar amount? I’ve tried the search and it requires info like requester or task name that I obviously wouldn’t know. I’ve searched and tried messing around and can’t figure it out. Thanks!

nope

 

[turkinthailand]

Just wondering if adding gIDs rather than rIDs to PCS is any more efficient? Or does it make no difference? I've only added rIDs so far, but for some regular HITs where the gID is fixed, I could easily change them to gIDs.

 

[Johnnyrs]

Is it possible to have it grab any hit I qualify for over a certain dollar amount? I’ve tried the search and it requires info like requester or task name that I obviously wouldn’t know. I’ve searched and tried messing around and can’t figure it out. Thanks!

No. The closest you can get to this is to use the Panda Crazy Search custom search with a 3 letter word or more. Once you create one then turn on the Automatic Collect HITs to true. Then you have to make sure the collect auto button is turned on. Now you can make many custom searches with different terms all automatic collecting and it can get close to what you want but you may have more HITs sitting in the queue which you may not be able to finish. So just target the important high paying terms.

 

[Johnnyrs]

Just wondering if adding gIDs rather than rIDs to PCS is any more efficient? Or does it make no difference? I've only added rIDs so far, but for some regular HITs where the gID is fixed, I could easily change them to gIDs.

It makes no difference at all. The extension just does a search for any GIDs and RIDs enabled on the MTURK search page and tries to grab it. I find GIDs are more used for one time HITs like surveys or a specific HIT from a requester that sends out multiple HITs instead of trying to include or exclude HITs in the RID. Take for example: a good survey drops and you have a panda job trying to grab it. It doesn't grab it for maybe like 30 minutes but it may get returned later on in the day by someone else. You don't want to have a panda job running all day taking a spot better used for another panda. Just move it to a GID search trigger on the search page and can have it search for it all day without sacrificing a valuable panda job spot.

 

[Dabeech]

When my custom search grabs a hit and pushes it to Panda Crazy Max, it only "goes ham" for 12 fetches and then it stops all together. It does not default to the standard search after HAM. I can't find what option fixes or address that issue. Thanks in advance for the help!

 

[Johnnyrs]

When my custom search grabs a hit and pushes it to Panda Crazy Max, it only "goes ham" for 12 fetches and then it stops all together. It does not default to the standard search after HAM. I can't find what option fixes or address that issue. Thanks in advance for the help!

You're right. It's not passing the option for fetches. There should be a temporary fetch option that is missing. I'll have to fix that in the next update.

 

[turkinthailand]

So I'm apparently using the latest version, 0.9.37. The behavior when searches are triggered from PCS to PCM has changed it seems.

What I used to do since first using the software is let the search trigger do its thing, generate a job in PCM, and then turn off the trigger temporarily in PCS. This would allow me to then take control of the job in PCM. But since yesterday (at least), even after turning off the trigger in PCS, I have ZERO control of the job in PCM. It keeps following the carried over settings, regardless of what buttons I click, or try to click.

So, for example, previously, if I turned off the search trigger, and then set the job to GOHAM manually in PCM, it would keep going until I changed it again. So it would behave like PCM jobs I have in there permanently, which is exactly what I wanted.

But now I seem to have zero control of jobs that have come over from PCS. Last night I wanted a job to just keep running in GOHAM mode for several minutes. But it wouldn't stick like it used. It kept stopping after 12 seconds or whatever the setting was.... so I had to sit there bashing the button every 12 seconds.

I don't know if this is a new bug or you have fixed a bug and now it's working how you wanted it to work, but it's really, really painful to work with now. I really liked how it worked before the update i.e. once I turn the trigger off in PCS, then it let me have total control of the job in PCM if I choose to control it by hand.

Thoughts?

On a side note, I wasn't even aware I was auto-updating versions.... is this a setting in Chrome somewhere?

cheers,

 

[Johnnyrs]

What I used to do since first using the software is let the search trigger do its thing, generate a job in PCM, and then turn off the trigger temporarily in PCS. This would allow me to then take control of the job in PCM. But since yesterday (at least), even after turning off the trigger in PCS, I have ZERO control of the job in PCM. It keeps following the carried over settings, regardless of what buttons I click, or try to click.

On a side note, I wasn't even aware I was auto-updating versions.... is this a setting in Chrome somewhere?

Yes there was a bug and a misunderstanding of what is passed over from the search triggers. My thought has always been if the extension automatically creates and starts collecting a panda there should be a time limit or there will be lot of panda's running and the user may not even be aware of it. If a job is manually started it should use the options that the user has set.

So there was a lot of confusion about the panda job options tab in the triggers for the duration and fetches limits. the new update adds temporary values towards the bottom which should be used to control jobs that are created by the triggers. The collecting after duration and collecting # of fetches gets put in the permanent options of jobs now. Those should be 0 if you want a job to run with no limit when manually collecting. Those are 0 when triggers are created now but I still need to fix a few bugs with it. You can also change any options in the panda job that was created and the triggers shouldn't change those values.

Chrome automatically updates and I don't think there is an easy way of stopping it for security reasons. In fact this extension will stop running in 2023 without future updates so best not to try to stop it. It gets updates rarely at this time anyway.

 

[turkinthailand]

also, just noticed this.... just got 3 hits running, and for some reason, it's staying on one hit 3 times, then doing the other 2 hits just once (repeatedly)

Requester: ****** | Pay: $0.05 | Mode: panda | Accepted: 0 | Fetched: 94 | Elapsed: 5.9s
Requester: ****** | Pay: $0.06 | Mode: panda | Accepted: 586 | Fetched: 63614 | Elapsed: 1.0s
Requester: ******** | Pay: $0.05 | Mode: panda | Accepted: 53 | Fetched: 43810 | Elapsed: 5.9s

That can't be right, it's like "weighting" has suddenly made a reappearance. Any ideas?

update: this was solved by restarting chrome.

 

[turkinthailand]

Yes there was a bug and a misunderstanding of what is passed over from the search triggers. My thought has always been if the extension automatically creates and starts collecting a panda there should be a time limit or there will be lot of panda's running and the user may not even be aware of it. If a job is manually started it should use the options that the user has set.

So there was a lot of confusion about the panda job options tab in the triggers for the duration and fetches limits. the new update adds temporary values towards the bottom which should be used to control jobs that are created by the triggers. The collecting after duration and collecting # of fetches gets put in the permanent options of jobs now. Those should be 0 if you want a job to run with no limit when manually collecting. Those are 0 when triggers are created now but I still need to fix a few bugs with it. You can also change any options in the panda job that was created and the triggers shouldn't change those values.

Chrome automatically updates and I don't think there is an easy way of stopping it for security reasons. In fact this extension will stop running in 2023 without future updates so best not to try to stop it. It gets updates rarely at this time anyway.

I'll have to digest this further. I much prefer the way it used to work. It was perfect for me but hey ho, I will try to understand how it works now )

thanks.

 

[vburnin8tor]

So I'm apparently using the latest version, 0.9.37. The behavior when searches are triggered from PCS to PCM has changed it seems.

So there was a lot of confusion about the panda job options tab in the triggers for the duration and fetches limits. the new update adds temporary values towards the bottom which should be used to control jobs that are created by the triggers. The collecting after duration and collecting # of fetches gets put in the permanent options of jobs now.

So if I understand correctly, prior to 0.9.37, when a search trigger in PCSearch created a job it would simply create the job in PCMax and instantly attempt to collect once or twice (usually accepting with a score of Acc:1|Fetched:1, or failing with a score of Acc:0|Fetched:1-5).
If you were to go back into PCMax and manually click 'Collect' or 'Once'/'Panda' in the MTurk workspace, forker, or Custom Triggered Hits tab, the job would begin infinitely.
(Ignoring the time and fetch limits was the bug)

Now, following 0.9.37+, a job created by PCSearch will instantly attempt to collect using the "TEMPORARY" limits set. And if these temporary limits are set to 0, the job will follow the previous behavior with a score of (Acc:0|Fetched:1-5).
And manually clicking collect will follow the time and fetch limits set. (bug fixed!)

So the way to return to the previous behavior is to set all jobs time and fetch limits to 0 instead, editing the temporary limits as needed.

The issue I encounter (similar to turkinthailand) is that once the job hits the permanent limits (not at 0), they can no longer be interacted with. If my permanent fetch limit is 5, once I click "Collect" it will fetch 5 and won't let me click collect again. I'd expect to click collect again and the job do 5 more fetches.

 

[turkinthailand]

Do you know what I end up doing now (which is really sad)? When a search triggers a new job that has lots of hits in it, I end up copying the job in PCM, deleting the one created by the search, and using the copied one.

Why? Because I don't like spending half my time fighting over the job and having the settings change even when I have clicked on the job myself to change the settings. Once the job is created, and I override the settings with mouse clicks, why not let those settings stick? There is nothing confusing about it. It's confusing when the job keeps turning on and off again, etc. etc. Doesn't the user know best what they want to actually do?

It feels analogous to a self-driven car. When I click on a job, I am saying - hey, I want to take over control here - please don't fight me

 

[Johnnyrs]

So if I understand correctly, prior to 0.9.37, when a search trigger in PCSearch created a job it would simply create the job in PCMax and instantly attempt to collect once or twice (usually accepting with a score of Acc:1|Fetched:1, or failing with a score of Acc:0|Fetched:1-5).
If you were to go back into PCMax and manually click 'Collect' or 'Once'/'Panda' in the MTurk workspace, forker, or Custom Triggered Hits tab, the job would begin infinitely.
(Ignoring the time and fetch limits was the bug)

Now, following 0.9.37+, a job created by PCSearch will instantly attempt to collect using the "TEMPORARY" limits set. And if these temporary limits are set to 0, the job will follow the previous behavior with a score of (Acc:0|Fetched:1-5).
And manually clicking collect will follow the time and fetch limits set. (bug fixed!)

So the way to return to the previous behavior is to set all jobs time and fetch limits to 0 instead, editing the temporary limits as needed.

The issue I encounter (similar to turkinthailand) is that once the job hits the permanent limits (not at 0), they can no longer be interacted with. If my permanent fetch limit is 5, once I click "Collect" it will fetch 5 and won't let me click collect again. I'd expect to click collect again and the job do 5 more fetches.

The previous version would use the stop collecting after (seconds) as the temporary value and copy it to the panda job options which was wrong. It would also use the Stop collecting after # of fetches as a temporary value and copy it to the panda job options which was also wrong. Now those options are copied to the new panda job so those should be at 0 usually which will allow the panda job to run indefinitely when a user presses the collect button manually. There is still another bug with stop collecting after option which should be in minutes not seconds exactly like the panda job options.

Now the temporary options should be used only if the extension creates a new panda job or starts a panda job on it's own. Those should have a value or it will use a default value so there won't be dozens of panda's running at the same time.

What you are describing to me sounds like you are setting the daily accepted HIT limit which will only let you collect that amount of hits in one day which is rare to use instead of the accept only once value. I would recommend to create a duplicate search trigger and change no values and it should let you press the collect button and it will fetch indefinitely.

 

[Johnnyrs]

Do you know what I end up doing now (which is really sad)? When a search triggers a new job that has lots of hits in it, I end up copying the job in PCM, deleting the one created by the search, and using the copied one.

Why? Because I don't like spending half my time fighting over the job and having the settings change even when I have clicked on the job myself to change the settings. Once the job is created, and I override the settings with mouse clicks, why not let those settings stick? There is nothing confusing about it. It's confusing when the job keeps turning on and off again, etc. etc. Doesn't the user know best what they want to actually do?

It feels analogous to a self-driven car. When I click on a job, I am saying - hey, I want to take over control here - please don't fight me

Like I said before that is how it is supposed to run. If the extension creates a panda job or starts to collect a panda job then it will use the temporary option values. Then when a user presses the collect button on their own it doesn't use the temporary values but it will use the panda job options for stop collecting and stop fetches. Those values should be 0 so it will run indefinitely which is set at 0 by default. So yes the extension should let the user have control. There is either a bug happening or an option which is set and it should not be set. I would recommend creating a new search trigger and leave the options alone. Check if you can press the collect button and have it collect indefinitely.

 

[vburnin8tor]

What you are describing to me sounds like you are setting the daily accepted HIT limit which will only let you collect that amount of hits in one day which is rare to use instead of the accept only once value. I would recommend to create a duplicate search trigger and change no values and it should let you press the collect button and it will fetch indefinitely.

I'm not using the Daily Accepted limit, but I am editing an existing trigger. Daily Accepted set to 0 and only editing the temporary fetch limit and the problem persists. Creating the new trigger fixed it.

 

[Johnnyrs]

Creating the new trigger fixed it.

Great! probably the old triggers still had something set which shouldn't have been set but wasn't able to be changed for some reason. I'm preparing a new update that should zero out those two options because it's rarely used anyway. That was my mistake. I should have copied those two values over to the temporary values and zeroed out the previous options.