Additions - JohnnyRS/PandaCrazy-Max GitHub Wiki
Some big changes in the background and new stats page has been added.
-
- Chrome has announced that all extension must be using manifest version 3 in the future. They will be phasing out the MV2 support next year and in 2023 all extension must be running on MV3. When I made this extension I used many things from MV2 that was removed in MV3 so I needed to change it as soon as possible to find any big problems. Now it is running very well under MV3 which means this extension will run for many years as is. Many javascript libraries being used have been updated to a newer version for security. Also in the future this extension may find it's way to firefox too.
-
- This page may be changing as time goes on but for now it will allow you to download the search results HITs that it has seen on MTURK. You must have the search advanced option on to have All MTURK HITs be stored for the extension to save anything. This option may slow things down but with all the saving being done in one transaction it shouldn't be a big problem. When there is data to be downloaded the page will display all the dates it has saved and show how many HITs it has recorded. For now it's limited to 7 days.
- The downloaded file will be in a CSV format which can be read by any database program. Because it will have all the HITs seen while the search page is running you can see trends or the most active requesters. At this time you will have to use the data on your own. The extension only records the data and lets you download it but maybe in the future I can make something to show the stats more clearly.
-
- This should make the extension faster when saving data to the database. This is implemented in the history, search and stats database. It will save all the data changes and save them all at once in one transaction at a later time, usually around every 40 seconds to 90 seconds. These times can be changed in the search advanced options if you want. These caches will be saved if you close the panda crazy page as well but may not be saved if there is a crash. This should make things faster by lowering CPU usage and disk drive usage.
-
- On the search job card you will now see stats: Acc: for accepted HITs, Hits: for HITs found and Fetch: for number of fetches. This will help you know how productive a search job is. Search triggers will show the accepted and found HITs when showing the stats by clicking on a trigger card.
-
- There is now an advanced menu on the panda crazy page which will allow you to change the debug messages shown on the browser console. By default it is set to 0 to show only major failure messages. The different levels will show you every time it sends a request to MTURK or shows the queue timer process. This can help if something is wrong and it will be easier for me to find what might be causing the problem.
This version has changed some actions and added more features. I will try to explain these so it can be easier to use.
-
Can now force a trigger or a search job to use a requester search rather than only looking for newer HITs. This will allow older HITs to be found.
- You can force a RID search trigger to search for ANY HITs by enabling the trigger when Panda Crazy Search is ON. Any time a trigger is enabled it will do a requester search first and then go back to searching for only NEW HITs after. So if a trigger is enabled and you want to check for older HITs you can disable it first by double clicking on it and then reenable it by double clicking it again. This will do a requester search once and grab any HITs it finds if none of them were found before.
- Search jobs on the Panda Crazy Page can also be forced to use a requester search in the same way by enabling the job. It will do a requester search once and then do a regular search for only NEW HITs.
- Search jobs also have a way of permanently using requester search with an advanced option. This only works with search jobs on Panda Crazy Page and not search triggers for Search Crazy Page. To use this new feature you must turn on the advanced search jobs option in the General Options. Make sure it is changed to TRUE. Save the options and now all RID search jobs will have a R button after the details button. This button will be off at the start with a white background. When this button is clicked and turned on with a green background then the search job is in requester search mode. When you enable this job it will search for any HITs under this requester and try to grab them with the highest paid HIT first. This is a similar way the old script worked. It will take a spot under the search timer so it will slow down any search triggers or jobs. This should only be used for requesters who are dropping many HITs with different group ID's currently. If the requester stops dropping HITs this should be toggled back to regular search job by clicking off the R button. This does not slow down other panda's because it takes it's turn from the search timer.
-
- This should only be used if you are looking for a specific $0.00 HIT or qualifications. This could cause the script to miss higher paying HITs if there are a lot of $0.00 HITs dropped all at once. The most important thing to have this work correctly is to change the Minimum Reward for MTURK search Page to $0.00 on the General Search Options. Once that option is set then any triggers or search jobs that has the min reward at $0.00 will correctly work. There are warnings to notify users about this important detail.
-
- MTURK approval rates of requesters are shown in the Custom Triggered HITs tab on the search page. You can also turn off displaying the rate by setting the option for Display MTURK Rate For Requesters to false in the general search options.