Commit Graph

190 Commits

Author SHA1 Message Date
469745ce3b partial fix/optimisation for refimgs, only recalc if file is newer than when we build the refimg encoding 2021-01-23 21:21:29 +11:00
f253760d26 update start time of jobs to when they actually are starting to run, should also help with durations being more accurate. Also added a catch-all force job finish if one is not done already. Will stop it enetering an endless loop 2021-01-23 20:19:44 +11:00
e17d6c41fe just use python time calls for jobs, now db and python calls, as on Cams box, this can results in the last update (Cams box) occurring before the job is created in the db (maras time) - and that shows up as negative durations 2021-01-23 20:05:15 +11:00
221ab03dbb fixed BUG-8 job failures should cascade cancels. Also put back dependency for AI job to wait for GetFileDetails (might not really be needed as I think Cams code just calls exif rotate itself anyway, also took out my temporary block of running the slow AI code. Need to discuss how to optimise it so it does not re-run parts it should know, e.g. refimg stuff once / not again unless ref imgs changed, ditto file/ai scan , etc. 2021-01-23 00:32:17 +11:00
bfe0613167 fixed BUGS 12/13/14, fixed new thumbnail code, fixed duplicate db entries for directories, fixed optimising code not working, by just ditching the directory optimising anyway 2021-01-23 00:00:47 +11:00
954132cf1a BUG-2 (FixPath) solved, because we no longer fix paths :) 2021-01-22 22:33:44 +11:00
685c626fd7 updated AI job, to show progress in joblog, and fixed BUG-12 (can now forcescan again) 2021-01-22 22:27:35 +11:00
aa81d46a09 fixed BUG-9, current_file_num uninitd 2021-01-22 21:15:47 +11:00
Cam
1cf65f4c31 Merge branch 'master' of mara.ddp.net:photoassistant 2021-01-22 18:48:34 +11:00
Cam
605b26af89 implemented ai in our pa_job_manager, with refimgs, a job, and the whole deal 2021-01-22 18:46:18 +11:00
cfd4a3071f fix current_file_num and num_files being wrong in import and gen*details jobs 2021-01-22 16:32:30 +11:00
f73d488ee7 mid-way through new Import function, has buggy counts, updated bugs and todos 2021-01-22 15:33:35 +11:00
2f3d7c1ae2 ProcessFilesInDir - now is generic and takes funcs to call on files 2021-01-21 22:22:42 +11:00
22cab9b090 fix BUG-5 (job file num cnt wrong when processing unknown file type) 2021-01-21 16:09:37 +11:00
a9819c613b more path fixes 2021-01-21 15:25:19 +11:00
586c8e8edf debug back on by default 2021-01-21 15:18:27 +11:00
4cf1613949 new fix for path shenanigans... python deals with it anyway, so just leave it "linux" pathed for simplicity 2021-01-21 15:18:12 +11:00
e56b7f51cc quick fix 2021-01-21 15:04:09 +11:00
197c864624 fix bug with unknown files 2021-01-21 14:59:33 +11:00
bdef281d3b remove FixPaths until we are sure we know the bug / fix 2021-01-21 14:28:48 +11:00
a716ed6007 use FixPath on keep_dirs key - a path, as the backslashes in windows cause python errors 2021-01-21 11:04:57 +11:00
db056e08aa remove pa_job_manager table, its not needed, and fixed bug where Gen hash optimisation did not finish job when it shoud have 2021-01-21 00:58:02 +11:00
2cc023bd10 cleaned up debugs, used if debug before print debug, also ran into threading issues, so for now RunJob is non-threaded 2021-01-20 23:22:44 +11:00
5b99855cb5 call HandleJobs() when RunJob() finished to run any dependant jobs 2021-01-20 22:01:14 +11:00
574c12f7f5 cleaned up debugs, and optimised gen hash / thumbs to only process dirs with new content 2021-01-20 21:46:41 +11:00
76e37a2dc0 changed all new_file to file, also make JobForceScan work 2021-01-20 20:02:31 +11:00
30195e9597 removed filedata*, moved ProcessImportDir out into its own function, used it, also in scannow job, ensure job states are set appropriately to enforce all of the jobs are processed right 2021-01-20 16:37:00 +11:00
0de95d56d5 fixed up JobExtra printout in class, also put it in job.py, and passed extras to templates/jobs.html to show the extra details on the jobs list page so we can see what path is imported for example 2021-01-20 00:05:56 +11:00
9474c12a0d JobGetFileDetails now works - first pass, need to improve handling of files in dirs - I optimised/assumed last dir is what you are in, but that is not always true, so need to set actual dir each time 2021-01-19 23:34:12 +11:00
2bd25c3e18 set up FileType properly, select on in DB, rather than add new ones each time, ouch! Also renamed type to type_id (better as it is a new foreign key) and use type as the relationship to hold the object anyway 2021-01-19 17:32:33 +11:00
47085b7a43 cleaned up code, before rewrite for viewing 2021-01-19 11:10:25 +11:00
5285bf66ab alternate database structure with entry, (new_)file, dir works in terms of DB and SQL alchemy, tomorrow we rewrite the jobs to really use the new structure 2021-01-18 21:40:46 +11:00
4053919def comment out sleep again, also removed some debugs, and converted some to AddLogForJob 2021-01-18 13:20:15 +11:00
507336c2b2 commented out sleep which I use for testing of active jobs/job refresh page, and re-title the job list html 2021-01-17 12:38:28 +11:00
2b9dedb9b9 change how we calculate active jobs, probably will remove it from pa_job_engine, safer with threads I think. But, mostyle, added in client / server socket comms between web FE and job manager, with better job creation message (including link to job detail) and when you view job detail it auto-refreshes every few seconds until job complete) 2021-01-17 12:35:28 +11:00
abff2d8bab finished moving GenerateFileData out of files.py into pa_job_manager.py 2021-01-17 11:31:06 +11:00
a89f7fdb2b fixed bugs with wrong sequence names in DB classes, also added AddLogForJob and used it in GenerateFileData, and fixed but where I re-ran a completed job 2021-01-17 11:05:14 +11:00
c5d8811671 actually create a job for scan now, and let the job manager do the work and send back output via new PA_JobManager_FE_Message table 2021-01-17 10:50:43 +11:00
b01bdea505 had to make pa_eng global in Init* func, but that worked, now have a basic harness to run a job, and ran scannow job, by pulling in much of the files.py code, but it will be removed from there when this is finished. Will have to move to IPC (via sockets?) to allow better coordination of new job processing and updating 2021-01-16 23:54:31 +11:00
826e2fd421 switched name to pa_job_manager, feels more appropriate 2021-01-16 17:46:54 +11:00