problems with kancolle-auto.py #3
Comments
|
I'm running windows 8.1 (6.3) and Sikuli 1.0.1 |
|
Sorry for late response, because I went on a trip and I could not have the Internet access there. As mentioned in README.md, I developed and tested kancolle-auto only on Ubuntu, so I'm not sure kancolle-auto runs on windows. |
|
Hello amylase. I have found what appears to be the cause of this issue. You may have already found the cause too, but please let me explain to Bararun. Let me give you an example. Imagine you are running expedition no.1 and no.2. When expedition no.1 returns, the script resupplies every fleet, even including the fleet running expedition no.2, which has not returned. The script then goes to the expedition sortie screen, and successfully sends a fleet to expedition no.1 again. However, here is where the problem is. The script also tries to send out expedition no.2 again, even though it has not yet returned. It looks for decision.png, which looks like this. But because expedition no.2 is actually still in progress, we see this image instead. This button has been added in a previous update, and allows you to cancel the expedition. Before this function was added, there was only a grayed out button indicating that the fleet is already on this expedition and cannot be sent out. This grayed out button looks like this. Currently, there are 2 solutions.
As always, there is always more than one way to fix things. I await your response, amylase. Thank you for your hard work. |
|
In addition, there are two other forks of this script. It appears that silfumus has the same problem as this version, but does not rely on the built in timer. Yukariin appears to be somewhat updated. |
|
In my fork I resupply and restart only returned fleet, not all. |
|
Yes, I see. It was working wonderfully (exp 5, 21, 38). Also, Yukariin, which version of Sikuli are you currently running? In addition, just out of curiosity, why is it recommended to use 1.0.x instead of 1.1.x? |
|
I use 1.0.1 (as I said in readme). By my sense 1.1.x is very unstable, so I use 1.0.1 and it worked great for me. |
|
Hello, Yukariin. I have encountered a new problem while using your fork. Instead of opening a new issue I'll just put it in this thread. I have been running exp 5, 21, 38. Here is the last part of the log. As you can see, it is still functioning correctly until the second half there: [log] CLICK on L(452,246)@s(0)[0,0 1366x768] [log] CLICK on L(625,573)@s(0)[0,0 1366x768] ensei id = 38, end time = 2014-08-09 20:58:30.661999 successfully started [log] CLICK on L(275,564)@s(0)[0,0 1366x768] [log] CLICK on L(290,420)@s(0)[0,0 1366x768] [log] CLICK on L(677,571)@s(0)[0,0 1366x768] Try ensei id = 21, end time = 2014-08-09 20:23:43.634000 [log] CLICK on L(625,573)@s(0)[0,0 1366x768] ensei id = 21, end time = 2014-08-09 20:23:50.971999 successfully started Restart expedition [log] CLICK on L(71,386)@s(0)[0,0 1366x768] [log] CLICK on L(77,351)@s(0)[0,0 1366x768] [error] script [ *kancolle_auto ] stopped with error in line 137 This problem has occurred several times, all while I am AFK or asleep. As such, I cannot see what the script is doing as it logs those errors. I was wondering if you had ever encountered the same errors, and whether you have a fix or workaround for them or not. However, I have a hunch. It may be because the script handles expeditions individually. As such, it may encounter problems when expeditions finish at the same time. What do you think, Yukariin? I patiently await your response. |
|
Seems like this happening in resupply stage. Also, you use any viewers? I see that script say's
Nope, after script sees that one of expeditions ends - script pausing checkings and do resupply-restart stages. |
|
Thank you for finding the issue and making the changes! I have just began using the new script and will report any problems if there are any after testing it for today. About the viewer, yes, I use KanColleViewer on Windows 8. I found the problem regarding the viewer. It seems the app focus is centered around KCTV, not KCV. I will make some slight adjustments. |
Yep, because sikuli use window title for focusing on that window. And KCV have title on Japanese, I tried focus on window with japanese letters in title and it seems like not working. |
|
I have found the cause of the problem and found a workaround. The cause is a bug in Sikuli, specifically versions pre-1.1.x only. It appears that App.focus() does not work for non-ASCII characters in versions pre-1.1.x. Japanese characters do not count as ASCII and would not work as a result. Here is the thread with the developer's response. However, this issue is related with Chinese characters (but is the same problem: non-ASCII characters do not work in App.focus): https://answers.launchpad.net/sikuli/+question/162050 It appears the developer fixed it in 1.1.x:https://bugs.launchpad.net/sikuli/+bug/1008113 However, because Sikuli 1.1.x is more unstable, I prefer not to use it. And I have found easy solution to described problem above. I found an English fork of the original KCV here: https://github.com/Yuubari/KanColleViewer This is an English fork with translations and some added features. The important thing is that this KCV version has all the features of the original KCV, but the title is no longer in Japanese. Though it is harsher to read in English, it appears that it retains the Japanese translations so I can read in Japanese again. It is simply "KanColleViewer!" right now. It is holding up fine but I will continue to monitor any errors if occurring and check the log each use. Here I have attached a portion of the log to show that it is working: [log] CLICK on L(176,248)@s(0)[0,0 1366x768] [log] CLICK on L(116,251)@s(0)[0,0 1366x768] [log] CLICK on L(702,569)@s(0)[0,0 1366x768] [log] App.focus KanColleViewer!(0) #0 [log] CLICK on L(71,386)@s(0)[0,0 1366x768] [log] CLICK on L(199,396)@s(0)[0,0 1366x768] Here is the portion in my version of your script regarding App.focus(): def check_window(): I cut if off here. It's just that one line of switchApp() which has its contents changed. Try it if you wish, and let's see if it works! |
I can just add one more string with |
|
It is working very well. Running expeditions 2, 5, 21 non stop and no problems so far. Thank you very much! |
|
No problem :3 |
|
Yep, I saw this. In most cases this happen on screen that makes request to KanColle API (high ping to API) or just animation lagging. |
|
I see! I will do some more examination to see where exactly it is slow and report soon. Thank you. |
|
So far, the logs often tell me that it cannot find sortie.png and expedition.png. Thus, I believe that it is lagging after it finishes resupply and clicks home.png, and also after it clicks sortie.png. |
|
Ok, I increase waiting time for |
|
Thanks! I leave this running for a while and will report what happened in a few hours. |
|
It fixed it! It is able to run continuously without interruption now. Thank you. |
|
I had a same problem with the sortie.png, And for the "expedition.py" you could subtract a minute from each expedition timedelta. |
|
Oh, thanks, I forgot about it... 2014-08-17 15:03 GMT+08:00 babatk notifications@github.com:
|
|
sometimes i see that the ships not properly supplied but sent to expedition it happend in this part by the way i used yukarins latest commit Expedition #3 ends, restarting i think the problem is the script tried to resupply the fleet before getting the result of the expediton it only happens once in a while this 2nd error i recieved when it tried to resupply but sent to claim the results of the expedition Expedition #6 ends, restarting |
Yeeep, sometimes
It's very strange, I already fix it in #3 (comment) |
|
For select all error, it is because sometimes sikuli clicks before the cursor is placed on top of the select all checkbox. Possible fix is to add pause. Move cursor, pause for 1 second, click. |
Not sure, I have this error when I manually resupply my fleets -___- |
|
yes i use your latest version it happens after continous use ok i tried to rerun this night i deleted and redowload the script to be sure and i run it on debug and save a log on a txt file im running 3 fleets 3,5,21 here is the one of the common errors i encounter it stopped this morning but it continued without stop from 1am to 8 am +8:00 happend on this segment log on txt file [log] CLICK on L(1109,512)@s(0)[0,0 1366x768] log on cmd ensei id = 3, end time = 2014-08-27 06:52:39.427000 successfully started the errors i encounter does always not happen i don't know why it happens the 2nd error i posted last night happen when it went to the home screen but instead of claiming the the results it tried to resupply(tried to click the ressupply button) the returned fleet but because if a fleet has returned you are forced to claim the results so it went to the claim the results page instead that's when the script stopped. the last error i encountered happens when it finished sending a fleet |
|
Well, I update script. |
|
ok ill try to run it today overnight ill post if i encounter an error |
|
@Yukariin Pinging 203.104.209.55 results in alot of time out errors but the game itself loads fine and quick for me.
After looking at it for a bit, I realized the majority of the times it's worked perfectly for me is when I launch all expeditions at once rather than launching them when 1 or 2 of the other fleets are already out, Because of that the timers would be off. So for example Fleet 1 is 5mins till completion and you launch Fleet 2 while Fleet 1 is still out, The script will take the current time and add the max duration of the expedition to its timer rather than checking the time remaining which causes the results to pile up at main screen. What ends up happening is that the script goes through one of the reward screens and then tries to go straight into resupply without clearing remaining reward screens. Which resulted in the error that I received. At Least that's what I'm seeing, I'll try running it over night for a bit fresh all fleets launched together by the script and see if that runs into the same issue. |
|
@vSwifty we can't get remain time for already launched exp. |
|
@vSwifty you can run 3 and 6 exps and see how it's working. |
|
@Yukariin Ah okay, I see. I'll run it for a little while and report any issues. |
|
D:\sikuli>runScript -d 2 -r kancolle_auto.sikuli kinda stuck here idk what happened but its a one time thing just reporting it did run normaly when i tried again |
|
@kevin01523 yep, script press |
|
yeah kinda noticed that too on earlier versions it happens once in a while but rarely but now it seems it happens more recently by the way i noticed this bug well it seems connected in a way it did resupply 2 of the fleets but not fleet 4 then when sending the expeds the scripted saw that fleet 4 was not supplied then it clicked area 1 then thats it it got stuck there happens once in a while. D:\sikuli>runScript -d 2 -r kancolle_auto.sikuli |
|
heres my fleet setup |
|
ok im just asking if is it possible for a fleet to run different expedition for example fleet number 2 runs on exped 2 but i want to run it on exped 6 next then exped 2 again and so on since fleet 2 satisy the requirements need for both exped 6 and 2 if this is possible it would be a great help and it is good for the script to be more flexible |
|
@kevin01523 this looks as very unusual usecase and I'm not sure that we really need this feature. Also I'm currently working on porting script on pure Java and make some GUI for it. But I'm stuck on jar packing because of SikuliX bug (Sikuli's lib doesn't see bundled into jar images and can't search them on screen). |
|
@Yukariin i've read some topics on your problem and i think most of them encounter the same bug well i guess you need to store the images in a directory on the users computer if on windows on the user directory like in kct well in other platform well i dont have any idea but in linux i guess in usr directory or a dirtectory that has fewer restrictions well i guess i just want to get buckets and baux and rsc at the same time without changing fleet setup or something like that |
|
@kevin01523 yep, I know about that, but this is very dirty solution and Sikuli should work with bundled into jar images -___- |
|
@Yukariin have you tested it w/ the latest 1.1.0 nightly builds of sikuli it might have a working solution for that.as it is a requested to implemented solution as i has read.well i guess its not implemented and does nothing in 1.0.1 |
|
Yep, I tested it on 1.1.0. May be this be fixed on 1.2.x or may be not... |
|
Hello, I keep getting this error with the script [error] script [ kancolle_auto ] stopped with error in line 120 I run expeditions 2, 11 and 21. |
|
https://github.com/tantinevincent/Onegai-ooyodosan i found a similar project w/ extras it works but need improvements |
|
I picked this game up recently and developed my own fork of this project as another alternative: https://github.com/mrmin123/kancolle-auto |
|
@mrmin123 ill test it on windows and ill post errors here |
|
@kevin01523 Thanks; if you find anything feel free to report it on my fork. I just opened up the Issues section of it. |
|
@mrmin123 ok noted ill test first the exped for 1 day |
|
Microsoft Windows [Version 6.0.6002] C:\Windows\system32>call C:\SikuliX\runIDE.cmd -r C:\SikuliX\kancolle_auto.sikul +++ JAVA_HOME not specified C:\Windows\system32>java -version Can anyone help me with this? It was running fine until I did a system restore. |
|
@AppleBoi86 try installing java 7 jre |
|
@kevin01523 Thank you very much. I got it to work again. Sorry if this is OT. |
|
well theres a workaround for java 8 you need to edit runsetup for sikuli for java 8 so youn need to resetup sikuli again for that |
|
@AppleBoi86 ah i forgot about this one http://stackoverflow.com/questions/27927332/sikuli-java-not-found-error well you can do this if you want sikuli to use java 8 instead of 7 |
|
i know java 8 and 7 can be installed side by side thats how it is on my pc now |
|
With the latest update, resupply all will not work anymore. We can solve it by changing the supply_all.png but it will get stuck if your other fleet is supply is full. |
|
its fixed on latest update on this repo https://github.com/mrmin123/kancolle-auto |
|
Thanks for that kevin01523. I ran into a problem and would like some help. Here's the problem I encounter.�C:\Windows\system32>call C:\Sikuli\runIDE.cmd -r C:\mrmin123\kancolle_auto.sikul |
I've managed to run kancolle-auto.py on Sikuli 1.0.1 and the expeditions run well whe they are sent for the first time. The problem comes when one of the expeditions return.
the script receives the expedition, ressuplies it but then it starts trying to send all the expeditions again, with the consequence that it can't find decision.png in any of the exeditions that haven't returned already.
this iss the log I'm receiving:
[log] CLICK on L(49,307)@s(0)[0,0 1366x768]
[log] CLICK on L(57,272)@s(0)[0,0 1366x768]
[log] CLICK on L(154,168)@s(0)[0,0 1366x768]
[log] CLICK on L(99,169)@s(0)[0,0 1366x768]
[log] CLICK on L(680,489)@s(0)[0,0 1366x768]
[log] CLICK on L(183,169)@s(0)[0,0 1366x768]
[log] CLICK on L(99,169)@s(0)[0,0 1366x768]
[log] CLICK on L(680,489)@s(0)[0,0 1366x768]
[log] CLICK on L(214,168)@s(0)[0,0 1366x768]
[log] CLICK on L(99,169)@s(0)[0,0 1366x768]
[log] CLICK on L(680,489)@s(0)[0,0 1366x768]
[log] CLICK on L(49,307)@s(0)[0,0 1366x768]
[log] CLICK on L(175,308)@s(0)[0,0 1366x768]
[log] CLICK on L(651,287)@s(0)[0,0 1366x768]
[log] CLICK on L(120,484)@s(0)[0,0 1366x768]
[log] CLICK on L(276,248)@s(0)[0,0 1366x768]
[error] script [ kancolle_auto ] stopped with error in line 122
[error] FindFailed ( can not find decision.png on the screen. )
[error] --- Traceback --- error source first line: module ( function ) statement 49: main ( shutsugeki ) click("decision.png")
114: main ( main ) running_ensei_list += shutsugeki()
107: main ( init ) main()
[error] --- Traceback --- end --------------
The text was updated successfully, but these errors were encountered: