Silarn / EDMC-BioScan

Exobiology plugin for EDMC - Species deduction engine for biological signals and analysis aid. Quickly determine targets for research. Discord: https://discord.gg/RhY7KPzTht
GNU General Public License v2.0
13 stars 1 forks source link

Latest Version Not Working For Me #1

Closed enrogae closed 1 year ago

enrogae commented 1 year ago

2.5 isn't working for me at all. I've tried installing twice - and while FSS scanning, when I find biologicals BioScan shows nothing on the EDMC window like it usually does. Additionally, when I try to run the journal scan it just sits at 0% indefinitely and makes no progress. I've tried reinstalling then FSSing without the journal scan process running at all, and still get no results in EDMC. I ended up having to revert back to the previous version which still works. EDMarketConnector-debug.log

Silarn commented 1 year ago

At least one issue appears to be that the migration didn't run. The volcanism column should have been made nullable. Would you mind attaching the db file so I can inspect it?

Silarn commented 1 year ago

Actually some of the errors I see in the log file correspond to 2.0. Let me make sure the zip files were packaged properly.

Silarn commented 1 year ago

I guess that was after you rolled back the version. If you can send me the log immediately after you start EDMC, that would help. That should show me the errors from the migration process. As it is, I'm only seeing errors from it trying and failing to run queries on a database structure that isn't correct as EDMC rotates log files over a size limit.

Silarn commented 1 year ago

Alternatively you can always delete the DB and create a fresh file, which will have the correct structure. Eventually any lost data will get restored by the journal import. The only thing you might lose is old waypoints for scans that were not completed.

Silarn commented 1 year ago

It seems I did mess up setting the version string in one part of the migration process but that would only affect running it more than once. To fix the initial migration issue I will need the db file, most likely.

Silarn commented 1 year ago

EDMC-BioScan-2.5.1-dev.zip

Please try this version. You may need to run it twice as the first run will fix the database version string. The second run should attempt the migration again and then you can send the log file for that.

My best guess is that there are null fields which were made required and I need to handle that correctly but I need info to accurately determine the problem.

This version will put the plugin on pause and announce if it encounters a migration error so it should be obvious when that happens.

enrogae commented 1 year ago

GitHub won't let me upload the Db file, says it doesn't support that file type. I'm kinda new around here, is there another way I can send? I also don't mind deleting the file and starting over. I'm not concerned about old waypoints for incomplete scans.

Silarn commented 1 year ago

If you have a dropbox / google drive / onedrive account you could share it from temporarily, that should work. It may let you upload a zipped version.

enrogae commented 1 year ago

bioscan.zip

See if this works

enrogae commented 1 year ago

EDMarketConnector-debug.log

I just ran EDMC after installing 2.5.1 and there is no BioScan panel at all. Here is debug file after that.

Silarn commented 1 year ago

Please try this.

EDMC-BioScan-2.5.1-test.zip

(That zip didn't include the sqlalchemy modules so if you erased the old directory it would have failed to start. This one does.)

I updated the migration to select the columns in order. Previously it was just inserting arbitrarily based on how it was laid out in the database file. As the distance column was inserted earlier, this was causing it to try putting data in the wrong column.

I also added some safety checks so that it would delete and recreate temporary columns and tables in case the migration failed and left it partially complete.

Silarn commented 1 year ago

EDMC-BioScan-2.5.1-test.zip

There was a potential issue with the above version if you hadn't run the migration yet as it would attempt to drop a column that didn't exist. This version should correct that.

If you can validate that it works, I will probably push a new release.

enrogae commented 1 year ago

I will try it out right now!

Silarn commented 1 year ago

The journal parsing can be slow, as I said. I just extended the progress to show one decimal place so it's easier to tell that progress is being made. But if it runs into a problem it generally will halt the process and remove the progress display. I suppose I should have it say something if there was an error.

enrogae commented 1 year ago

That might be nice. It seems to be working with this version again. Just ran an FSS in a new system and it's giving me possible species types again! Nice! Parsing still stuck at 0% but I'll let it run overnight and see how it goes.

Totally unrelated - do you think we might ever get an option to hide already scanned signals on planets (both on planet and in system) when we get all three stages on them? Especially with the limited vertical space, in systems with many signals it would be nice to be able to 'check them off' from the list as we go and shorten the list down to nothing. It would make my inner completionist very happy hehe.

Anyway, LOVE the plugin!!! I wanted something more lightweight than BioInsights and this is just fantastic. I recommend it to everyone in my squad all the time. If you ever need any testing on anything I'd be happy to guinea pig for you. :)

Silarn commented 1 year ago

Sure, I think that's a reasonable request. The new version will have improved codex display as well (so you can tell which color / species is 'marked').

I'll push a release shortly, which will include the tenth place in the progress.

enrogae commented 1 year ago

Awesome. Thanks so much! You rule!

Silarn commented 1 year ago

Turns out there was an issue on Windows with the journal threading. I've fixed this in the 2.5.1 downloads.

enrogae commented 1 year ago

Like I said, you rule!

On Wed, Jun 7, 2023, 9:57 PM Jeremy Rimpo @.***> wrote:

Turns out there was an issue on Windows with the journal threading. I've fixed this in the 2.5.1 downloads.

— Reply to this email directly, view it on GitHub https://github.com/Silarn/EDMC-BioScan/issues/1#issuecomment-1581814970, or unsubscribe https://github.com/notifications/unsubscribe-auth/BANAWHVLX66BGUKT5LEEJ2TXKE5SLANCNFSM6AAAAAAY6RKZKI . You are receiving this because you modified the open/close state.Message ID: @.***>

enrogae commented 1 year ago

Didn't want to post a new issue for this, but just let you know. 2.5.4 gave me an error on first attempt to start the Journal Parse process - said Error Parsing Journals, please submit a report. However, I clicked it a second time and I'm now getting a percentage countup (first time for me) so it looks like it is working now.

Not a big issue but since I got the error the first time, wanted to let you know. You're still my hero!

Silarn commented 1 year ago

If you still see a BioScan error in the log, feel free to attach it. It's hard to say otherwise. I tried to find a good balance point where the threads wouldn't cause a database lock timeout from too many concurrent writes but it might still happen. If that's the case, then restarting wouldn't necessarily trigger it again unless another database lockout occurs.

enrogae commented 1 year ago

If I see another error message I'll let you know. For now I'm just going to assume it was a fluke. 😎

On Thu, Jun 8, 2023, 9:46 PM Jeremy Rimpo @.***> wrote:

If you still see a BioScan error in the log, feel free to attach it. It's hard to say otherwise. I tried to find a good balance point where the threads wouldn't cause a database lock timeout from to many concurrent writes but it could still happen. If that's what happened then restarting wouldn't necessarily trigger it again unless another timeout happened.

— Reply to this email directly, view it on GitHub https://github.com/Silarn/EDMC-BioScan/issues/1#issuecomment-1583870730, or unsubscribe https://github.com/notifications/unsubscribe-auth/BANAWHQCEB7XFRZ22GR7NBTXKKE65ANCNFSM6AAAAAAY6RKZKI . You are receiving this because you modified the open/close state.Message ID: @.***>

Silarn commented 1 year ago

Fair enough but I may consider wrapping the database commits in a limited timed retry loop for the next release. In that case it should only fail if there's a serious issue.

enrogae commented 1 year ago

If it helps you there were a couple of BioScan errors logged today when I started it up. EDMarketConnector-debug.log

FWIW I just restarted and and reran the journal parse and it completed to 100% in about a second. (Since it's already completed before)

Silarn commented 1 year ago

Looks like it failed trying to write star data for Sol. I can only guess that multiple journals had Sol scan data they tried to write at the same time so the one that committed second failed because star data has to be unique. A retry loop would probably correct for that as well.