RepoMirrors
/
hydrus
mirror of https://github.com/hydrusnetwork/hydrus.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
hydrus/hydrus/client/ClientFiles.py

3297 lines
131 KiB
Python
Raw Blame History

import collections
import os
import random
import threading
import time
import typing
from hydrus.core import HydrusConstants as HC
from hydrus.core import HydrusData
from hydrus.core import HydrusExceptions
from hydrus.core import HydrusGlobals as HG
from hydrus.core import HydrusLists
from hydrus.core import HydrusPaths
from hydrus.core import HydrusThreading
from hydrus.core import HydrusTime
from hydrus.core.files import HydrusFileHandling
from hydrus.core.files import HydrusPSDHandling
from hydrus.core.files import HydrusVideoHandling
from hydrus.core.files.images import HydrusBlurhash
from hydrus.core.files.images import HydrusImageColours
from hydrus.core.files.images import HydrusImageHandling
from hydrus.core.files.images import HydrusImageMetadata
from hydrus.core.files.images import HydrusImageOpening
from hydrus.core.networking import HydrusNetworking
from hydrus.client import ClientConstants as CC
from hydrus.client import ClientFilesPhysical
from hydrus.client import ClientImageHandling
from hydrus.client import ClientPaths
from hydrus.client import ClientSVGHandling # important to keep this in, despite not being used, since there's initialisation stuff in here
from hydrus.client import ClientPDFHandling # important to keep this in, despite not being used, since there's initialisation stuff in here
from hydrus.client import ClientThreading
from hydrus.client import ClientVideoHandling
from hydrus.client.metadata import ClientTags
REGENERATE_FILE_DATA_JOB_FILE_METADATA = 0
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL = 1
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL = 2
REGENERATE_FILE_DATA_JOB_OTHER_HASHES = 3
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES = 4
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD = 5
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD = 6
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS = 7
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP = 8
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA = 9
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP = 10
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL = 11
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL = 12
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE = 13
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD = 14
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD = 15
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE = 16
REGENERATE_FILE_DATA_JOB_PIXEL_HASH = 17
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY = 18
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA = 19
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF = 20
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD = 21
REGENERATE_FILE_DATA_JOB_BLURHASH = 22
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY = 23
regen_file_enum_to_str_lookup = {
REGENERATE_FILE_DATA_JOB_FILE_METADATA : 'regenerate file metadata',
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL : 'regenerate thumbnail',
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL : 'regenerate thumbnail if incorrect size',
REGENERATE_FILE_DATA_JOB_OTHER_HASHES : 'regenerate non-standard hashes',
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES : 'delete duplicate neighbours with incorrect file extension',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD : 'if file is missing, remove record (leave no delete record)',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD : 'if file is missing, remove record (leave a delete record)',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL : 'if file is missing, then if has URL try to redownload',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD : 'if file is missing, then if has URL try to redownload, else remove record',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY : 'if file is missing, note it in log',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD : 'if file is missing/incorrect, move file out and remove record',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL : 'if file is missing/incorrect, then move file out, and if has URL try to redownload',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD : 'if file is missing/incorrect, then move file out, and if has URL try to redownload, else remove record',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE : 'if file is incorrect, move file out',
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS : 'fix file read/write permissions',
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP : 'check for membership in the similar files search system',
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA : 'regenerate perceptual hashes',
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP : 'regenerate file modified time',
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY: 'determine if the file has transparency',
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF : 'determine if the file has EXIF metadata',
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA : 'determine if the file has non-EXIF human-readable embedded metadata',
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE : 'determine if the file has an icc profile',
REGENERATE_FILE_DATA_JOB_PIXEL_HASH : 'regenerate pixel hashes',
REGENERATE_FILE_DATA_JOB_BLURHASH: 'regenerate blurhash'
}
regen_file_enum_to_description_lookup = {
REGENERATE_FILE_DATA_JOB_FILE_METADATA : 'This regenerates file metadata like resolution and duration, or even filetype (such as mkv->webm), which may have been misparsed in a previous version.',
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL : 'This forces a complete regeneration of the thumbnail from the source file.',
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL : 'This looks for the existing thumbnail, and if it is not the correct resolution or is missing, will regenerate a new one for the source file.',
REGENERATE_FILE_DATA_JOB_OTHER_HASHES : 'This regenerates hydrus\'s store of md5, sha1, and sha512 supplementary hashes, which it can use for various external (usually website) lookups.',
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES : 'Sometimes, a file metadata regeneration will mean a new filetype and thus a new file extension. If the existing, incorrectly named file is in use, it must be copied rather than renamed, and so there is a spare duplicate left over after the operation. This jobs cleans up the duplicate at a later time.',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD : '''This checks to see if the file is present in the file system as expected. Use this if you have lost a number of files from your file structure, do not think you can recover them, and need hydrus to re-sync with what it actually has.
Missing files will have their internal file record in the database removed. This is just like a file delete except it does not leave a deletion record, so if you ever find the file again in future, you can import it again easily.
All missing files will have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD : '''This checks to see if the file is present in the file system as expected. Use this if you have manually deleted a number of files from your file structure, do not want to get them again, and need hydrus to re-sync with what it actually has. Another example of this situation is restoring an old backed-up database to a newer client_files structure--to catch the database up, you want to teach it that any files missing in the newer structure should be deleted, with a record.
Missing files will have their internal file record deleted just like a normal file delete. Normal imports that see these files again in future will ignore them as 'previously deleted'.
All missing files will have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL : '''This checks to see if the file is present in the file system as expected. If it is not, and it has known post/file URLs, the URLs will be automatically added to a new URL downloader.'
All missing files will have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD : '''THIS IS THE EASY AND QUICK ONE-SHOT WAY TO REPAIR A DATABASE WITH MISSING FILES.
This checks to see if the file is present in the file system as expected. If it is not, and it has known post/file URLs, the URLs will be automatically added to a new URL downloader.
Missing files with no URLs will have their internal file record in the database removed. This is just like a file delete except it does not leave a deletion record, so if you ever find the file again in future, you can import it again easily.
All missing files will have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY : 'This checks to see if the file is present in the file system as expected. If it is not, it records the file\'s hash, tags, and URLs to your database directory, just like the other "missing file" jobs, but makes no other action.',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD : '''This does the same check as the \'file is missing\' job, and if the file is where it is expected, it ensures its file content, byte-for-byte, is as expected. This discovers hard drive damage or other external interference. This is a heavy job, so be wary.
Missing/Incorrect files will have their internal file record in the database removed. This is just like a file delete except it does not leave a deletion record, so if you ever find the file again in future, you can import it again easily.
All incorrect files will be exported to a new folder in your database directory for later manual examination if you wish.
All missing/Incorrect files will also have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL : '''This does the same check as the \'file is missing\' job, and if the file is where it is expected, it ensures its file content, byte-for-byte, is as expected. This discovers hard drive damage or other external interference. This is a heavy job, so be wary. If the file is incorrect _and_ has known post/file URLs, the URLs will be automatically added to a new URL downloader.
All incorrect files will be exported to a new folder in your database directory for later manual examination if you wish.
All missing/Incorrect files will also have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD : '''This does the same check as the \'file is missing\' job, and if the file is where it is expected, it ensures its file content, byte-for-byte, is as expected. This discovers hard drive damage or other external interference. This is a heavy job, so be wary. If the file is incorrect _and_ has known post/file URLs, the URLs will be automatically added to a new URL downloader.
Missing/Incorrect files with no URLs will have their internal file record in the database removed. This is just like a file delete except it does not leave a deletion record, so if you ever find the file again in future, you can import it again easily.
All incorrect files will be exported to a new folder in your database directory for later manual examination if you wish.
All missing/Incorrect files will also have their hashes, tags, and URLs exported to a new folder in your database directory for later manual recovery attempts if you wish.''',
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE : 'If the file is where it is expected, this ensures its file content, byte-for-byte, is correct. This is a heavy job, so be wary. If the file is incorrect, it will be exported to your database directory along with its known URLs. The client\'s file record will not be deleted. This is useful if you have a valid backup and need to clear out invalid files from your live db so you can fill in gaps from your backup with a program like FreeFileSync.',
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS : 'This ensures that files in the file system are readable and writeable. For Linux/macOS users, it specifically sets 644. If you wish to run this job on Linux/macOS, ensure you are first the file owner of all your files.',
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP : 'This checks to see if files should be in the similar files system, and if they are falsely in or falsely out, it will remove their record or queue them up for a search as appropriate. It is useful to repair database damage.',
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA : 'This forces a regeneration of the file\'s similar-files \'phashes\'. It is not useful unless you know there is missing data to repair.',
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP : 'This rechecks the file\'s modified timestamp and saves it to the database.',
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY : 'This loads the file to see if it has an alpha channel with useful data (completely opaque/transparency alpha channels are discarded). Only works for images and animated gif.',
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF : 'This loads the file to see if it has EXIF metadata, which can be shown in the media viewer and searched with "system:image has exif".',
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA : 'This loads the file to see if it has non-EXIF human-readable metadata, which can be shown in the media viewer and searched with "system:image has human-readable embedded metadata".',
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE : 'This loads the file to see if it has an ICC profile, which is used in "system:has icc profile" search.',
REGENERATE_FILE_DATA_JOB_PIXEL_HASH : 'This generates a fast unique identifier for the pixels in a still image, which is used in duplicate pixel searches.',
REGENERATE_FILE_DATA_JOB_BLURHASH: 'This generates a very small version of the file\'s thumbnail that can be used as a placeholder while the thumbnail loads.'
}
NORMALISED_BIG_JOB_WEIGHT = 100
regen_file_enum_to_job_weight_lookup = {
REGENERATE_FILE_DATA_JOB_FILE_METADATA : 100,
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL : 50,
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL : 25,
REGENERATE_FILE_DATA_JOB_OTHER_HASHES : 100,
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES : 25,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD : 5,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD : 5,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL : 25,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD : 30,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY : 5,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD : 100,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL : 100,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD : 100,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE : 100,
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS : 25,
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP : 1,
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA : 100,
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP : 10,
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY : 25,
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF : 25,
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA : 25,
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE : 25,
REGENERATE_FILE_DATA_JOB_PIXEL_HASH : 100,
REGENERATE_FILE_DATA_JOB_BLURHASH: 15
}
regen_file_enum_to_overruled_jobs = {
REGENERATE_FILE_DATA_JOB_FILE_METADATA : [],
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL : [ REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL ],
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL : [],
REGENERATE_FILE_DATA_JOB_OTHER_HASHES : [],
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES : [],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY : [],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD : [ REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD ],
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE : [],
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS : [],
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP : [],
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA : [ REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP ],
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP : [],
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY : [],
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF : [],
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA : [],
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE : [],
REGENERATE_FILE_DATA_JOB_PIXEL_HASH : [],
REGENERATE_FILE_DATA_JOB_BLURHASH: []
}
ALL_REGEN_JOBS_IN_RUN_ORDER = [
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY,
REGENERATE_FILE_DATA_JOB_FILE_METADATA,
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL,
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL,
REGENERATE_FILE_DATA_JOB_BLURHASH,
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA,
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP,
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS,
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP,
REGENERATE_FILE_DATA_JOB_OTHER_HASHES,
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY,
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF,
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA,
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE,
REGENERATE_FILE_DATA_JOB_PIXEL_HASH,
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES
]
ALL_REGEN_JOBS_IN_HUMAN_ORDER = [
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY,
REGENERATE_FILE_DATA_JOB_FILE_METADATA,
REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL,
REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL,
REGENERATE_FILE_DATA_JOB_BLURHASH,
REGENERATE_FILE_DATA_JOB_PIXEL_HASH,
REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA,
REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP,
REGENERATE_FILE_DATA_JOB_OTHER_HASHES,
REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP,
REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY,
REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF,
REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA,
REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE,
REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS,
REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES
]
def GetAllFilePaths( raw_paths, do_human_sort = True, clear_out_sidecars = True ):
file_paths = []
paths_to_process = list( raw_paths )
while len( paths_to_process ) > 0:
next_paths_to_process = []
for path in paths_to_process:
if HG.started_shutdown:
raise HydrusExceptions.ShutdownException()
if os.path.isdir( path ):
try:
# on Windows, some network file paths return True on isdir(). maybe something to do with path length or number of subdirs
path_listdir = os.listdir( path )
except NotADirectoryError:
file_paths.append( path )
continue
subpaths = [ os.path.join( path, filename ) for filename in path_listdir ]
next_paths_to_process.extend( subpaths )
else:
file_paths.append( path )
paths_to_process = next_paths_to_process
if do_human_sort:
HydrusData.HumanTextSort( file_paths )
num_files_with_sidecars = len( file_paths )
if clear_out_sidecars:
exts = [ '.txt', '.json', '.xml' ]
def has_sidecar_ext( p ):
if True in ( p.endswith( ext ) for ext in exts ):
return True
return False
def get_base_prefix_component( p ):
base_prefix = os.path.basename( p )
if '.' in base_prefix:
base_prefix = base_prefix.split( '.', 1 )[0]
return base_prefix
# let's get all the 'Image123' in our 'path/to/Image123.jpg' list
all_non_ext_prefix_components = { get_base_prefix_component( file_path ) for file_path in file_paths if not has_sidecar_ext( file_path ) }
def looks_like_a_sidecar( p ):
# if we have Image123.txt, that's probably a sidecar!
return has_sidecar_ext( p ) and get_base_prefix_component( p ) in all_non_ext_prefix_components
file_paths = [ path for path in file_paths if not looks_like_a_sidecar( path ) ]
num_sidecars = num_files_with_sidecars - len( file_paths )
return ( file_paths, num_sidecars )
class ClientFilesManager( object ):
def __init__( self, controller ):
self._controller = controller
self._rwlock = ClientThreading.FileRWLock()
self._prefixes_to_client_files_subfolders = collections.defaultdict( list )
self._smallest_prefix = 2
self._largest_prefix = 2
self._physical_file_delete_wait = threading.Event()
self._locations_to_free_space = {}
self._bad_error_occurred = False
self._missing_subfolders = set()
self._Reinit()
self._controller.sub( self, 'Reinit', 'new_ideal_client_files_locations' )
self._controller.sub( self, 'shutdown', 'shutdown' )
def _AddFile( self, hash, mime, source_path ):
dest_path = self._GenerateExpectedFilePath( hash, mime )
if HG.file_report_mode or HG.file_import_report_mode:
HydrusData.ShowText( 'Adding file to client file structure: from {} to {}'.format( source_path, dest_path ) )
file_size = os.path.getsize( source_path )
dest_free_space = self._GetFileStorageFreeSpace( hash )
if dest_free_space < 100 * 1048576 or dest_free_space < file_size:
message = 'The disk for path "{}" is almost full and cannot take the file "{}", which is {}! Shut the client down now and fix this!'.format( dest_path, hash.hex(), HydrusData.ToHumanBytes( file_size ) )
HydrusData.ShowText( message )
self._HandleCriticalDriveError()
raise Exception( message )
try:
HydrusPaths.MirrorFile( source_path, dest_path )
except Exception as e:
message = f'Copying the file from "{source_path}" to "{dest_path}" failed! Details should be shown and other import queues should be paused. You should shut the client down now and fix this!'
HydrusData.ShowText( message )
self._HandleCriticalDriveError()
raise Exception( message ) from e
def _AddThumbnailFromBytes( self, hash, thumbnail_bytes, silent = False ):
dest_path = self._GenerateExpectedThumbnailPath( hash )
if HG.file_report_mode:
HydrusData.ShowText( 'Adding thumbnail: ' + str( ( len( thumbnail_bytes ), dest_path ) ) )
try:
HydrusPaths.TryToGiveFileNicePermissionBits( dest_path )
with open( dest_path, 'wb' ) as f:
f.write( thumbnail_bytes )
except Exception as e:
subfolder = self._GetSubfolderForFile( hash, 't' )
if not subfolder.PathExists():
raise HydrusExceptions.DirectoryMissingException( f'The directory {subfolder} was not found! Reconnect the missing location or shut down the client immediately!' )
raise HydrusExceptions.FileMissingException( 'The thumbnail for file "{}" failed to write to path "{}". This event suggests that hydrus does not have permission to write to its thumbnail folder. Please check everything is ok.'.format( hash.hex(), dest_path ) )
if not silent:
self._controller.pub( 'clear_thumbnails', { hash } )
self._controller.pub( 'new_thumbnails', { hash } )
def _AttemptToHealMissingLocations( self ):
# if a missing prefix folder seems to be in another location, lets update to that other location
correct_rows = []
some_are_unhealable = False
fixes_counter = collections.Counter()
known_base_locations = self._GetCurrentSubfolderBaseLocations()
( media_base_locations, thumbnail_override_base_location ) = self._controller.Read( 'ideal_client_files_locations' )
known_base_locations.update( media_base_locations )
if thumbnail_override_base_location is not None:
known_base_locations.add( thumbnail_override_base_location )
for missing_subfolder in self._missing_subfolders:
missing_base_location = missing_subfolder.base_location
prefix = missing_subfolder.prefix
potential_correct_base_locations = []
for known_base_location in known_base_locations:
if known_base_location == missing_base_location:
continue
potential_location_subfolder = ClientFilesPhysical.FilesStorageSubfolder( prefix, known_base_location )
if potential_location_subfolder.PathExists():
potential_correct_base_locations.append( known_base_location )
if len( potential_correct_base_locations ) == 1:
correct_base_location = potential_correct_base_locations[0]
correct_subfolder = ClientFilesPhysical.FilesStorageSubfolder( prefix, correct_base_location )
correct_rows.append( ( missing_subfolder, correct_subfolder ) )
fixes_counter[ ( missing_base_location, correct_base_location ) ] += 1
else:
some_are_unhealable = True
if len( correct_rows ) > 0 and some_are_unhealable:
message = 'Hydrus found multiple missing locations in your file storage. Some of these locations seemed to be fixable, others did not. The client will now inform you about both problems.'
self._controller.BlockingSafeShowCriticalMessage( 'Multiple file location problems.', message )
if len( correct_rows ) > 0:
summaries = sorted( ( '{} folders seem to have moved from {} to {}'.format( HydrusData.ToHumanInt( count ), missing_base_location, correct_base_location ) for ( ( missing_base_location, correct_base_location ), count ) in fixes_counter.items() ) )
summary_message = 'Some client file folders were missing, but they appear to be in other known locations! The folders are:'
summary_message += os.linesep * 2
summary_message += os.linesep.join( summaries )
summary_message += os.linesep * 2
summary_message += 'Assuming you did this on purpose, or hydrus recently inserted stub values after database corruption, Hydrus is ready to update its internal knowledge to reflect these new mappings as soon as this dialog closes. If you know these proposed fixes are incorrect, terminate the program now.'
HydrusData.Print( summary_message )
self._controller.BlockingSafeShowCriticalMessage( 'About to auto-heal client file folders.', summary_message )
HG.client_controller.WriteSynchronous( 'repair_client_files', correct_rows )
def _ChangeFileExt( self, hash, old_mime, mime ):
old_path = self._GenerateExpectedFilePath( hash, old_mime )
new_path = self._GenerateExpectedFilePath( hash, mime )
if old_path == new_path:
# some diff mimes have the same ext
return
if HG.file_report_mode:
HydrusData.ShowText( 'Changing file ext: ' + str( ( old_path, new_path ) ) )
if HydrusPaths.PathIsFree( old_path ):
try:
HydrusPaths.MergeFile( old_path, new_path )
needed_to_copy_file = False
except:
HydrusPaths.MirrorFile( old_path, new_path )
needed_to_copy_file = True
else:
HydrusPaths.MirrorFile( old_path, new_path )
needed_to_copy_file = True
return needed_to_copy_file
def _GenerateExpectedFilePath( self, hash, mime ):
self._WaitOnWakeup()
subfolder = self._GetSubfolderForFile( hash, 'f' )
hash_encoded = hash.hex()
return subfolder.GetFilePath( f'{hash_encoded}{HC.mime_ext_lookup[ mime ]}' )
def _GenerateExpectedThumbnailPath( self, hash ):
self._WaitOnWakeup()
subfolder = self._GetSubfolderForFile( hash, 't' )
hash_encoded = hash.hex()
return subfolder.GetFilePath( f'{hash_encoded}.thumbnail' )
def _GenerateThumbnailBytes( self, file_path, media ):
hash = media.GetHash()
mime = media.GetMime()
( width, height ) = media.GetResolution()
duration = media.GetDurationMS()
num_frames = media.GetNumFrames()
bounding_dimensions = self._controller.options[ 'thumbnail_dimensions' ]
thumbnail_scale_type = self._controller.new_options.GetInteger( 'thumbnail_scale_type' )
thumbnail_dpr_percent = HG.client_controller.new_options.GetInteger( 'thumbnail_dpr_percent' )
target_resolution = HydrusImageHandling.GetThumbnailResolution( ( width, height ), bounding_dimensions, thumbnail_scale_type, thumbnail_dpr_percent )
percentage_in = self._controller.new_options.GetInteger( 'video_thumbnail_percentage_in' )
try:
thumbnail_bytes = HydrusFileHandling.GenerateThumbnailBytes( file_path, target_resolution, mime, duration, num_frames, percentage_in = percentage_in )
except Exception as e:
raise HydrusExceptions.FileMissingException( 'The thumbnail for file ' + hash.hex() + ' could not be regenerated from the original file for the above reason. This event could indicate hard drive corruption. Please check everything is ok.' )
return thumbnail_bytes
def _GetCurrentSubfolderBaseLocations( self, only_files = False ):
known_base_locations = set()
for ( prefix, subfolders ) in self._prefixes_to_client_files_subfolders.items():
if only_files and not prefix.startswith( 'f' ):
continue
for subfolder in subfolders:
known_base_locations.add( subfolder.base_location )
return known_base_locations
def _GetFileStorageFreeSpace( self, hash: bytes ) -> int:
subfolder = self._GetSubfolderForFile( hash, 'f' )
base_location = subfolder.base_location
if base_location in self._locations_to_free_space:
( free_space, time_fetched ) = self._locations_to_free_space[ base_location ]
if free_space > 100 * ( 1024 ** 3 ):
check_period = 3600
elif free_space > 15 * ( 1024 ** 3 ):
check_period = 600
else:
check_period = 60
if HydrusTime.TimeHasPassed( time_fetched + check_period ):
free_space = HydrusPaths.GetFreeSpace( base_location.path )
self._locations_to_free_space[ base_location ] = ( free_space, HydrusTime.GetNow() )
else:
free_space = HydrusPaths.GetFreeSpace( base_location.path )
self._locations_to_free_space[ base_location ] = ( free_space, HydrusTime.GetNow() )
return free_space
def _GetPossibleSubfoldersForFile( self, hash: bytes, prefix_type: str ) -> typing.List[ ClientFilesPhysical.FilesStorageSubfolder ]:
hash_encoded = hash.hex()
result = []
for i in range( self._smallest_prefix, self._largest_prefix + 1 ):
prefix = prefix_type + hash_encoded[ : i ]
if prefix in self._prefixes_to_client_files_subfolders:
result.extend( self._prefixes_to_client_files_subfolders[ prefix ] )
return result
def _GetAllSubfolders( self ) -> typing.List[ ClientFilesPhysical.FilesStorageSubfolder ]:
result = []
for ( prefix, subfolders ) in self._prefixes_to_client_files_subfolders.items():
result.extend( subfolders )
return result
def _GetRebalanceTuple( self ):
# TODO: obviously this will change radically when we move to multiple folders for real and background migration. hacks for now
# In general, I think this thing is going to determine the next migration destination and purge flag
# the background file migrator will work on current purge flags and not talk to this guy until the current flags are clear
( ideal_media_base_locations, ideal_thumbnail_override_base_location ) = self._controller.Read( 'ideal_client_files_locations' )
service_info = HG.client_controller.Read( 'service_info', CC.COMBINED_LOCAL_FILE_SERVICE_KEY )
all_local_files_total_size = service_info[ HC.SERVICE_INFO_TOTAL_SIZE ]
total_ideal_weight = sum( ( base_location.ideal_weight for base_location in ideal_media_base_locations ) )
smallest_subfolder_normalised_weight = 1
largest_subfolder_normalised_weight = 0
current_base_locations_to_normalised_weights = collections.Counter()
current_base_locations_to_size_estimate = collections.Counter()
file_prefixes = [ prefix for prefix in self._prefixes_to_client_files_subfolders.keys() if prefix.startswith( 'f' ) ]
all_media_base_locations = set( ideal_media_base_locations )
for file_prefix in file_prefixes:
subfolders = self._prefixes_to_client_files_subfolders[ file_prefix ]
subfolder = subfolders[0]
base_location = subfolder.base_location
all_media_base_locations.add( base_location )
normalised_weight = subfolder.GetNormalisedWeight()
current_base_locations_to_normalised_weights[ base_location ] += normalised_weight
current_base_locations_to_size_estimate[ base_location ] += normalised_weight * all_local_files_total_size
if normalised_weight < smallest_subfolder_normalised_weight:
smallest_subfolder_normalised_weight = normalised_weight
if normalised_weight > largest_subfolder_normalised_weight:
largest_subfolder_normalised_weight = normalised_weight
smallest_subfolder_num_bytes = smallest_subfolder_normalised_weight * all_local_files_total_size
#
# ok so the problem here is that when a location blocks new subfolders because of max num bytes rules, the other guys have to take the slack and end up overweight
# we want these overweight guys to nonetheless distribute their stuff according to relative weights
# so, what we'll do is we'll play a game with a split-pot, where bust players can't get dosh from later rounds
second_round_base_locations = []
desperately_overweight_locations = []
overweight_locations = []
available_locations = []
starving_locations = []
# first round, we need to sort out who is bust
total_normalised_weight_lost_in_first_round = 0
for base_location in all_media_base_locations:
current_num_bytes = current_base_locations_to_size_estimate[ base_location ]
if not base_location.AbleToAcceptSubfolders( current_num_bytes, smallest_subfolder_num_bytes ):
if base_location.max_num_bytes is None:
total_normalised_weight_lost_in_first_round = base_location.ideal_weight / total_ideal_weight
else:
total_normalised_weight_lost_in_first_round += base_location.max_num_bytes / all_local_files_total_size
if base_location.NeedsToRemoveSubfolders( current_num_bytes ):
desperately_overweight_locations.append( base_location )
else:
second_round_base_locations.append( base_location )
random.shuffle( second_round_base_locations )
# second round, let's distribute the remainder
# I fixed some logic and it seems like everything here is now AbleToAccept, so maybe we want another quick pass on this
# or just wait until I do the slow migration and we'll figure something out with the staticmethod on BaseLocation that just gets ideal weights
# I also added this jank regarding / ( 1 - first_round_weight ), which makes sure we are distributing the remaining weight correctly
second_round_total_ideal_weight = sum( ( base_location.ideal_weight for base_location in second_round_base_locations ) )
for base_location in second_round_base_locations:
current_normalised_weight = current_base_locations_to_normalised_weights[ base_location ]
current_num_bytes = current_base_locations_to_size_estimate[ base_location ]
# can be both overweight and able to eat more
if base_location.WouldLikeToRemoveSubfolders( current_normalised_weight / ( 1 - total_normalised_weight_lost_in_first_round ), second_round_total_ideal_weight, largest_subfolder_normalised_weight ):
overweight_locations.append( base_location )
if base_location.EagerToAcceptSubfolders( current_normalised_weight / ( 1 - total_normalised_weight_lost_in_first_round ), second_round_total_ideal_weight, smallest_subfolder_normalised_weight, current_num_bytes, smallest_subfolder_num_bytes ):
starving_locations.insert( 0, base_location )
elif base_location.AbleToAcceptSubfolders( current_num_bytes, smallest_subfolder_num_bytes ):
available_locations.append( base_location )
#
if len( desperately_overweight_locations ) > 0:
potential_sources = desperately_overweight_locations
potential_destinations = starving_locations + available_locations
elif len( overweight_locations ) > 0:
potential_sources = overweight_locations
potential_destinations = starving_locations
else:
potential_sources = []
potential_destinations = []
if len( potential_sources ) > 0 and len( potential_destinations ) > 0:
source_base_location = potential_sources.pop( 0 )
destination_base_location = potential_destinations.pop( 0 )
random.shuffle( file_prefixes )
for file_prefix in file_prefixes:
subfolders = self._prefixes_to_client_files_subfolders[ file_prefix ]
subfolder = subfolders[0]
base_location = subfolder.base_location
if base_location == source_base_location:
overweight_subfolder = ClientFilesPhysical.FilesStorageSubfolder( file_prefix, source_base_location )
underweight_subfolder = ClientFilesPhysical.FilesStorageSubfolder( file_prefix, destination_base_location )
return ( overweight_subfolder, underweight_subfolder )
else:
thumbnail_prefixes = [ prefix for prefix in self._prefixes_to_client_files_subfolders.keys() if prefix.startswith( 't' ) ]
for thumbnail_prefix in thumbnail_prefixes:
if ideal_thumbnail_override_base_location is None:
file_prefix = 'f' + thumbnail_prefix[1:]
subfolders = None
if file_prefix in self._prefixes_to_client_files_subfolders:
subfolders = self._prefixes_to_client_files_subfolders[ file_prefix ]
else:
# TODO: Consider better that thumbs might not be split but files would.
# We need to better deal with t43 trying to find its place in f431, and t431 to f43, which means triggering splits or whatever (when we get to that code)
for ( possible_file_prefix, possible_subfolders ) in self._prefixes_to_client_files_subfolders.items():
if possible_file_prefix.startswith( file_prefix ) or file_prefix.startswith( possible_file_prefix ):
subfolders = possible_subfolders
break
if subfolders is None:
# this shouldn't ever fire, and by the time I expect to split subfolders, all this code will work different anyway
# no way it could possibly go wrong
raise Exception( 'Had a problem trying to find a thumnail migration location due to split subfolders! Let hydev know!' )
subfolder = subfolders[0]
correct_base_location = subfolder.base_location
else:
correct_base_location = ideal_thumbnail_override_base_location
subfolders = self._prefixes_to_client_files_subfolders[ thumbnail_prefix ]
subfolder = subfolders[0]
current_thumbnails_base_location = subfolder.base_location
if current_thumbnails_base_location != correct_base_location:
current_subfolder = ClientFilesPhysical.FilesStorageSubfolder( thumbnail_prefix, current_thumbnails_base_location )
correct_subfolder = ClientFilesPhysical.FilesStorageSubfolder( thumbnail_prefix, correct_base_location )
return ( current_subfolder, correct_subfolder )
return None
def _GetSubfolderForFile( self, hash: bytes, prefix_type: str ) -> ClientFilesPhysical.FilesStorageSubfolder:
# TODO: So this will be a crux of the more complicated system
# might even want a media result eventually, for various 'ah, because it is archived, it should go here'
# for now it is a patch to navigate multiples into our currently mutually exclusive storage dataset
# we probably need to break this guy into variants of 'getpossiblepaths' vs 'getidealpath' for different callers
# getideal would be testing purge states and client files locations max num bytes stuff
# there should, in all circumstances, be a place to put a file, so there should always be at least one non-num_bytes'd location with weight to handle 100% coverage of the spillover
# if we are over the limit on the place the directory is supposed to be, I think we are creating a stub subfolder in the spillover place and writing there, but that'll mean saving a new subfolder, so be careful
# maybe the spillover should always have 100% coverage no matter what, and num_bytes'd locations should always just have extensions. something to think about
return self._GetPossibleSubfoldersForFile( hash, prefix_type )[0]
def _HandleCriticalDriveError( self ):
self._controller.new_options.SetBoolean( 'pause_import_folders_sync', True )
self._controller.new_options.SetBoolean( 'pause_subs_sync', True )
self._controller.new_options.SetBoolean( 'pause_all_file_queues', True )
HydrusData.ShowText( 'A critical drive error has occurred. All importers--subscriptions, import folders, and paged file import queues--have been paused. Once the issue is clear, restart the client and resume your imports after restart under the file and network menus!' )
self._controller.pub( 'notify_refresh_network_menu' )
def _IterateAllFilePaths( self ):
for ( prefix, subfolders ) in self._prefixes_to_client_files_subfolders.items():
if prefix.startswith( 'f' ):
for subfolder in subfolders:
files_dir = subfolder.path
filenames = list( os.listdir( files_dir ) )
for filename in filenames:
yield os.path.join( files_dir, filename )
def _IterateAllThumbnailPaths( self ):
for ( prefix, subfolders ) in self._prefixes_to_client_files_subfolders.items():
if prefix.startswith( 't' ):
for subfolder in subfolders:
files_dir = subfolder.path
filenames = list( os.listdir( files_dir ) )
for filename in filenames:
yield os.path.join( files_dir, filename )
def _LookForFilePath( self, hash ):
for potential_mime in HC.ALLOWED_MIMES:
potential_path = self._GenerateExpectedFilePath( hash, potential_mime )
if os.path.exists( potential_path ):
return ( potential_path, potential_mime )
subfolders = self._GetPossibleSubfoldersForFile( hash, 'f' )
for subfolder in subfolders:
if not subfolder.PathExists():
raise HydrusExceptions.DirectoryMissingException( f'The directory {subfolder.path} was not found! Reconnect the missing location or shut down the client immediately!' )
raise HydrusExceptions.FileMissingException( 'File for ' + hash.hex() + ' not found!' )
def _Reinit( self ):
self._ReinitSubfolders()
if HG.client_controller.IsFirstStart():
try:
dirs_to_test = set()
for subfolder in self._GetAllSubfolders():
dirs_to_test.add( subfolder.base_location.path )
dirs_to_test.add( subfolder.path )
for dir_to_test in dirs_to_test:
try:
HydrusPaths.MakeSureDirectoryExists( dir_to_test )
except:
text = 'Attempting to create the database\'s client_files folder structure in {} failed!'.format( dir_to_test )
self._controller.BlockingSafeShowCriticalMessage( 'unable to create file structure', text )
raise
except:
text = 'Attempting to create the database\'s client_files folder structure failed!'
self._controller.BlockingSafeShowCriticalMessage( 'unable to create file structure', text )
raise
else:
self._ReinitMissingLocations()
if len( self._missing_subfolders ) > 0:
self._AttemptToHealMissingLocations()
self._ReinitSubfolders()
self._ReinitMissingLocations()
if len( self._missing_subfolders ) > 0:
self._bad_error_occurred = True
#
missing_dict = HydrusData.BuildKeyToListDict( [ ( subfolder.base_location, subfolder.prefix ) for subfolder in self._missing_subfolders ] )
missing_base_locations = sorted( missing_dict.keys(), key = lambda b_l: b_l.path )
missing_string = ''
for missing_base_location in missing_base_locations:
missing_prefixes = sorted( missing_dict[ missing_base_location ] )
missing_prefixes_string = ' ' + os.linesep.join( ( ', '.join( block ) for block in HydrusLists.SplitListIntoChunks( missing_prefixes, 32 ) ) )
missing_string += os.linesep
missing_string += str( missing_base_location )
missing_string += os.linesep
missing_string += missing_prefixes_string
#
if len( self._missing_subfolders ) > 4:
text = 'When initialising the client files manager, some file locations did not exist! They have all been written to the log!'
text += os.linesep * 2
text += 'If this is happening on client boot, you should now be presented with a dialog to correct this manually!'
self._controller.BlockingSafeShowCriticalMessage( 'missing locations', text )
HydrusData.DebugPrint( 'Missing locations follow:' )
HydrusData.DebugPrint( missing_string )
else:
text = 'When initialising the client files manager, these file locations did not exist:'
text += os.linesep * 2
text += missing_string
text += os.linesep * 2
text += 'If this is happening on client boot, you should now be presented with a dialog to correct this manually!'
self._controller.BlockingSafeShowCriticalMessage( 'missing locations', text )
def _ReinitSubfolders( self ):
subfolders = self._controller.Read( 'client_files_subfolders' )
self._prefixes_to_client_files_subfolders = collections.defaultdict( list )
for subfolder in subfolders:
self._prefixes_to_client_files_subfolders[ subfolder.prefix ].append( subfolder )
self._smallest_prefix = min( ( len( prefix ) for prefix in self._prefixes_to_client_files_subfolders.keys() ) ) - 1
self._largest_prefix = max( ( len( prefix ) for prefix in self._prefixes_to_client_files_subfolders.keys() ) ) - 1
def _ReinitMissingLocations( self ):
self._missing_subfolders = set()
for ( prefix, subfolders ) in self._prefixes_to_client_files_subfolders.items():
for subfolder in subfolders:
if not subfolder.PathExists():
self._missing_subfolders.add( subfolder )
def _WaitOnWakeup( self ):
if HG.client_controller.new_options.GetBoolean( 'file_system_waits_on_wakeup' ):
while HG.client_controller.JustWokeFromSleep():
HydrusThreading.CheckIfThreadShuttingDown()
time.sleep( 1.0 )
def AllLocationsAreDefault( self ):
with self._rwlock.read:
db_dir = self._controller.GetDBDir()
client_files_default = os.path.join( db_dir, 'client_files' )
all_base_locations = self._GetCurrentSubfolderBaseLocations()
return False not in ( location.path.startswith( client_files_default ) for location in all_base_locations )
def LocklessAddFileFromBytes( self, hash, mime, file_bytes ):
dest_path = self._GenerateExpectedFilePath( hash, mime )
if HG.file_report_mode:
HydrusData.ShowText( 'Adding file from string: ' + str( ( len( file_bytes ), dest_path ) ) )
HydrusPaths.TryToGiveFileNicePermissionBits( dest_path )
with open( dest_path, 'wb' ) as f:
f.write( file_bytes )
def AddFile( self, hash, mime, source_path, thumbnail_bytes = None ):
with self._rwlock.write:
self._AddFile( hash, mime, source_path )
if thumbnail_bytes is not None:
self._AddThumbnailFromBytes( hash, thumbnail_bytes )
def AddThumbnailFromBytes( self, hash, thumbnail_bytes, silent = False ):
with self._rwlock.write:
self._AddThumbnailFromBytes( hash, thumbnail_bytes, silent = silent )
def ChangeFileExt( self, hash, old_mime, mime ):
if old_mime == mime:
return False
with self._rwlock.write:
return self._ChangeFileExt( hash, old_mime, mime )
def ClearOrphans( self, move_location = None ):
with self._rwlock.write:
job_status = ClientThreading.JobStatus( cancellable = True )
job_status.SetStatusTitle( 'clearing orphans' )
job_status.SetStatusText( 'preparing' )
self._controller.pub( 'message', job_status )
orphan_paths = []
orphan_thumbnails = []
for ( i, path ) in enumerate( self._IterateAllFilePaths() ):
( i_paused, should_quit ) = job_status.WaitIfNeeded()
if should_quit:
return
if i % 100 == 0:
status = 'reviewed ' + HydrusData.ToHumanInt( i ) + ' files, found ' + HydrusData.ToHumanInt( len( orphan_paths ) ) + ' orphans'
job_status.SetStatusText( status )
try:
( directory, filename ) = os.path.split( path )
should_be_a_hex_hash = filename[:64]
hash = bytes.fromhex( should_be_a_hex_hash )
is_an_orphan = HG.client_controller.Read( 'is_an_orphan', 'file', hash )
except:
is_an_orphan = True
if is_an_orphan:
if move_location is not None:
( source_dir, filename ) = os.path.split( path )
dest = os.path.join( move_location, filename )
dest = HydrusPaths.AppendPathUntilNoConflicts( dest )
HydrusData.Print( 'Moving the orphan ' + path + ' to ' + dest )
try:
HydrusPaths.MergeFile( path, dest )
except Exception as e:
HydrusData.ShowText( f'Had trouble moving orphan from {path} to {dest}! Abandoning job!' )
HydrusData.ShowException( e, do_wait = False )
job_status.Cancel()
return
orphan_paths.append( path )
time.sleep( 2 )
for ( i, path ) in enumerate( self._IterateAllThumbnailPaths() ):
( i_paused, should_quit ) = job_status.WaitIfNeeded()
if should_quit:
return
if i % 100 == 0:
status = 'reviewed ' + HydrusData.ToHumanInt( i ) + ' thumbnails, found ' + HydrusData.ToHumanInt( len( orphan_thumbnails ) ) + ' orphans'
job_status.SetStatusText( status )
try:
is_an_orphan = False
( directory, filename ) = os.path.split( path )
should_be_a_hex_hash = filename[:64]
hash = bytes.fromhex( should_be_a_hex_hash )
is_an_orphan = HG.client_controller.Read( 'is_an_orphan', 'thumbnail', hash )
except:
is_an_orphan = True
if is_an_orphan:
orphan_thumbnails.append( path )
time.sleep( 2 )
if move_location is None and len( orphan_paths ) > 0:
status = 'found ' + HydrusData.ToHumanInt( len( orphan_paths ) ) + ' orphans, now deleting'
job_status.SetStatusText( status )
time.sleep( 5 )
for ( i, path ) in enumerate( orphan_paths ):
( i_paused, should_quit ) = job_status.WaitIfNeeded()
if should_quit:
return
HydrusData.Print( 'Deleting the orphan ' + path )
status = 'deleting orphan files: ' + HydrusData.ConvertValueRangeToPrettyString( i + 1, len( orphan_paths ) )
job_status.SetStatusText( status )
ClientPaths.DeletePath( path )
if len( orphan_thumbnails ) > 0:
status = 'found ' + HydrusData.ToHumanInt( len( orphan_thumbnails ) ) + ' orphan thumbnails, now deleting'
job_status.SetStatusText( status )
time.sleep( 5 )
for ( i, path ) in enumerate( orphan_thumbnails ):
( i_paused, should_quit ) = job_status.WaitIfNeeded()
if should_quit:
return
status = 'deleting orphan thumbnails: ' + HydrusData.ConvertValueRangeToPrettyString( i + 1, len( orphan_thumbnails ) )
job_status.SetStatusText( status )
HydrusData.Print( 'Deleting the orphan ' + path )
ClientPaths.DeletePath( path, always_delete_fully = True )
if len( orphan_paths ) == 0 and len( orphan_thumbnails ) == 0:
final_text = 'no orphans found!'
else:
final_text = HydrusData.ToHumanInt( len( orphan_paths ) ) + ' orphan files and ' + HydrusData.ToHumanInt( len( orphan_thumbnails ) ) + ' orphan thumbnails cleared!'
job_status.SetStatusText( final_text )
HydrusData.Print( job_status.ToString() )
job_status.Finish()
def DeleteNeighbourDupes( self, hash, true_mime ):
with self._rwlock.write:
correct_path = self._GenerateExpectedFilePath( hash, true_mime )
if not os.path.exists( correct_path ):
return # misfire, let's not actually delete the right one
for mime in HC.ALLOWED_MIMES:
if mime == true_mime:
continue
incorrect_path = self._GenerateExpectedFilePath( hash, mime )
if incorrect_path == correct_path:
# some diff mimes have the same ext
continue
if os.path.exists( incorrect_path ):
delete_ok = HydrusPaths.DeletePath( incorrect_path )
if not delete_ok and random.randint( 1, 52 ) != 52:
self._controller.WriteSynchronous( 'file_maintenance_add_jobs_hashes', { hash }, REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES, HydrusTime.GetNow() + ( 7 * 86400 ) )
def DoDeferredPhysicalDeletes( self ):
wait_period = self._controller.new_options.GetInteger( 'ms_to_wait_between_physical_file_deletes' ) / 1000
num_files_deleted = 0
num_thumbnails_deleted = 0
while not HG.started_shutdown:
with self._rwlock.write:
( file_hash, thumbnail_hash ) = self._controller.Read( 'deferred_physical_delete' )
if file_hash is None and thumbnail_hash is None:
break
if file_hash is not None:
media_result = self._controller.Read( 'media_result', file_hash )
expected_mime = media_result.GetMime()
try:
path = self._GenerateExpectedFilePath( file_hash, expected_mime )
if not os.path.exists( path ):
( path, actual_mime ) = self._LookForFilePath( file_hash )
ClientPaths.DeletePath( path )
num_files_deleted += 1
except HydrusExceptions.FileMissingException:
HydrusData.Print( 'Wanted to physically delete the "{}" file, with expected mime "{}", but it was not found!'.format( file_hash.hex(), HC.mime_string_lookup[ expected_mime ] ) )
if thumbnail_hash is not None:
path = self._GenerateExpectedThumbnailPath( thumbnail_hash )
if os.path.exists( path ):
ClientPaths.DeletePath( path, always_delete_fully = True )
num_thumbnails_deleted += 1
self._controller.WriteSynchronous( 'clear_deferred_physical_delete', file_hash = file_hash, thumbnail_hash = thumbnail_hash )
if num_files_deleted % 10 == 0 or num_thumbnails_deleted % 10 == 0:
self._controller.pub( 'notify_new_physical_file_delete_numbers' )
self._physical_file_delete_wait.wait( wait_period )
self._physical_file_delete_wait.clear()
if num_files_deleted > 0 or num_thumbnails_deleted > 0:
self._controller.pub( 'notify_new_physical_file_delete_numbers' )
HydrusData.Print( 'Physically deleted {} files and {} thumbnails from file storage.'.format( HydrusData.ToHumanInt( num_files_deleted ), HydrusData.ToHumanInt( num_files_deleted ) ) )
def GetCurrentFileBaseLocations( self ):
with self._rwlock.read:
return self._GetCurrentSubfolderBaseLocations( only_files = True )
def GetFilePath( self, hash, mime = None, check_file_exists = True ):
with self._rwlock.read:
if HG.file_report_mode:
HydrusData.ShowText( 'File path request: ' + str( ( hash, mime ) ) )
if mime is None:
( path, mime ) = self._LookForFilePath( hash )
else:
path = self._GenerateExpectedFilePath( hash, mime )
if check_file_exists and not os.path.exists( path ):
try:
# let's see if the file exists, but with the wrong ext!
( actual_path, old_mime ) = self._LookForFilePath( hash )
except HydrusExceptions.FileMissingException:
raise HydrusExceptions.FileMissingException( 'No file found at path {}!'.format( path ) )
self._ChangeFileExt( hash, old_mime, mime )
# we have now fixed the path, it is good to return
return path
def GetMissingSubfolders( self ):
return self._missing_subfolders
def GetThumbnailPath( self, media ):
hash = media.GetHash()
mime = media.GetMime()
if HG.file_report_mode:
HydrusData.ShowText( 'Thumbnail path request: ' + str( ( hash, mime ) ) )
with self._rwlock.read:
path = self._GenerateExpectedThumbnailPath( hash )
thumb_missing = not os.path.exists( path )
if thumb_missing:
self.RegenerateThumbnail( media )
return path
def LocklessHasFile( self, hash, mime ):
path = self._GenerateExpectedFilePath( hash, mime )
if HG.file_report_mode:
HydrusData.ShowText( 'File path test: ' + path )
return os.path.exists( path )
def LocklessHasThumbnail( self, hash ):
path = self._GenerateExpectedThumbnailPath( hash )
if HG.file_report_mode:
HydrusData.ShowText( 'Thumbnail path test: ' + path )
return os.path.exists( path )
def Rebalance( self, job_status ):
try:
if self._bad_error_occurred:
HydrusData.ShowText( 'A serious file error has previously occurred during this session, so further file moving will not be reattempted. Please restart the client before trying again.' )
return
with self._rwlock.write:
rebalance_tuple = self._GetRebalanceTuple()
while rebalance_tuple is not None:
if job_status.IsCancelled():
break
( source_subfolder, dest_subfolder ) = rebalance_tuple
text = f'Moving "{source_subfolder}" to "{dest_subfolder}".'
HydrusData.Print( text )
job_status.SetStatusText( text )
# these two lines can cause a deadlock because the db sometimes calls stuff in here.
self._controller.WriteSynchronous( 'relocate_client_files', source_subfolder, dest_subfolder )
self._Reinit()
rebalance_tuple = self._GetRebalanceTuple()
time.sleep( 0.01 )
finally:
job_status.SetStatusText( 'done!' )
job_status.FinishAndDismiss()
def RebalanceWorkToDo( self ):
with self._rwlock.read:
return self._GetRebalanceTuple() is not None
def RegenerateThumbnail( self, media ):
hash = media.GetHash()
mime = media.GetMime()
if mime not in HC.MIMES_WITH_THUMBNAILS:
return
with self._rwlock.read:
file_path = self._GenerateExpectedFilePath( hash, mime )
if not os.path.exists( file_path ):
raise HydrusExceptions.FileMissingException( 'The thumbnail for file ' + hash.hex() + ' could not be regenerated from the original file because the original file is missing! This event could indicate hard drive corruption. Please check everything is ok.')
thumbnail_bytes = self._GenerateThumbnailBytes( file_path, media )
with self._rwlock.write:
self._AddThumbnailFromBytes( hash, thumbnail_bytes )
return True
def RegenerateThumbnailIfWrongSize( self, media ):
do_it = False
try:
hash = media.GetHash()
mime = media.GetMime()
if mime not in HC.MIMES_WITH_THUMBNAILS:
return
( media_width, media_height ) = media.GetResolution()
path = self._GenerateExpectedThumbnailPath( hash )
if not os.path.exists( path ):
raise Exception()
thumbnail_mime = HydrusFileHandling.GetThumbnailMime( path )
numpy_image = ClientImageHandling.GenerateNumPyImage( path, thumbnail_mime )
( current_width, current_height ) = HydrusImageHandling.GetResolutionNumPy( numpy_image )
bounding_dimensions = self._controller.options[ 'thumbnail_dimensions' ]
thumbnail_scale_type = self._controller.new_options.GetInteger( 'thumbnail_scale_type' )
thumbnail_dpr_percent = HG.client_controller.new_options.GetInteger( 'thumbnail_dpr_percent' )
( expected_width, expected_height ) = HydrusImageHandling.GetThumbnailResolution( (media_width, media_height), bounding_dimensions, thumbnail_scale_type, thumbnail_dpr_percent )
if current_width != expected_width or current_height != expected_height:
do_it = True
except:
do_it = True
if do_it:
self.RegenerateThumbnail( media )
return do_it
def Reinit( self ):
# this is still useful to hit on ideals changing, since subfolders bring the weight and stuff of those settings. we'd rather it was generally synced
self._Reinit()
def UpdateFileModifiedTimestamp( self, media, modified_timestamp: int ):
hash = media.GetHash()
mime = media.GetMime()
with self._rwlock.write:
path = self._GenerateExpectedFilePath( hash, mime )
if os.path.exists( path ):
existing_access_time = os.path.getatime( path )
existing_modified_time = os.path.getmtime( path )
try:
os.utime( path, ( existing_access_time, modified_timestamp ) )
HydrusData.Print( 'Successfully changed modified time of "{}" from {} to {}.'.format( path, HydrusTime.TimestampToPrettyTime( existing_modified_time ), HydrusTime.TimestampToPrettyTime( modified_timestamp ) ))
except PermissionError:
HydrusData.Print( 'Tried to set modified time of {} to file "{}", but did not have permission!'.format( HydrusTime.TimestampToPrettyTime( modified_timestamp ), path ) )
def shutdown( self ):
self._physical_file_delete_wait.set()
def HasHumanReadableEmbeddedMetadata( path, mime ):
if mime not in HC.FILES_THAT_CAN_HAVE_HUMAN_READABLE_EMBEDDED_METADATA:
return False
if mime == HC.APPLICATION_PDF:
has_human_readable_embedded_metadata = ClientPDFHandling.HasHumanReadableEmbeddedMetadata( path )
else:
try:
pil_image = HydrusImageOpening.RawOpenPILImage( path )
except:
return False
has_human_readable_embedded_metadata = HydrusImageMetadata.HasHumanReadableEmbeddedMetadata( pil_image )
return has_human_readable_embedded_metadata
def HasTransparency( path, mime, duration = None, num_frames = None, resolution = None ):
if mime not in HC.MIMES_THAT_WE_CAN_CHECK_FOR_TRANSPARENCY:
return False
try:
if mime in HC.IMAGES:
numpy_image = HydrusImageHandling.GenerateNumPyImage( path, mime )
return HydrusImageColours.NumPyImageHasUsefulAlphaChannel( numpy_image )
elif mime in ( HC.ANIMATION_GIF, HC.ANIMATION_APNG ):
if num_frames is None or resolution is None:
return False # something crazy going on, so let's bail out
we_checked_alpha_channel = False
if mime == HC.ANIMATION_GIF:
renderer = ClientVideoHandling.GIFRenderer( path, num_frames, resolution )
else: # HC.ANIMATION_APNG
renderer = HydrusVideoHandling.VideoRendererFFMPEG( path, mime, duration, num_frames, resolution )
for i in range( num_frames ):
numpy_image = renderer.read_frame()
if not we_checked_alpha_channel:
if not HydrusImageColours.NumPyImageHasAlphaChannel( numpy_image ):
return False
we_checked_alpha_channel = True
if HydrusImageColours.NumPyImageHasUsefulAlphaChannel( numpy_image ):
return True
except HydrusExceptions.DamagedOrUnusualFileException as e:
HydrusData.Print( 'Problem determining transparency for "{}":'.format( path ) )
HydrusData.PrintException( e )
return False
return False
def add_extra_comments_to_job_status( job_status: ClientThreading.JobStatus ):
extra_comments = []
num_thumb_refits = job_status.GetIfHasVariable( 'num_thumb_refits' )
num_bad_files = job_status.GetIfHasVariable( 'num_bad_files' )
if num_thumb_refits is not None:
extra_comments.append( 'thumbs needing regen: {}'.format( HydrusData.ToHumanInt( num_thumb_refits ) ) )
if num_bad_files is not None:
extra_comments.append( 'missing or invalid files: {}'.format( HydrusData.ToHumanInt( num_bad_files ) ) )
sub_status_message = '\n'.join( extra_comments )
if len( sub_status_message ) > 0:
job_status.SetStatusText( sub_status_message, 2 )
class FilesMaintenanceManager( object ):
def __init__( self, controller ):
self._controller = controller
self._pubbed_message_about_bad_file_record_delete = False
self._pubbed_message_about_invalid_file_export = False
self._work_tracker = HydrusNetworking.BandwidthTracker()
self._idle_work_rules = HydrusNetworking.BandwidthRules()
self._active_work_rules = HydrusNetworking.BandwidthRules()
self._ReInitialiseWorkRules()
self._maintenance_lock = threading.Lock()
self._lock = threading.Lock()
self._serious_error_encountered = False
self._wake_background_event = threading.Event()
self._reset_background_event = threading.Event()
self._shutdown = False
self._controller.sub( self, 'NotifyNewOptions', 'notify_new_options' )
self._controller.sub( self, 'Wake', 'checkbox_manager_inverted' )
self._controller.sub( self, 'Shutdown', 'shutdown' )
def _AbleToDoBackgroundMaintenance( self ):
HG.client_controller.WaitUntilViewFree()
if HG.client_controller.CurrentlyIdle():
if not self._controller.new_options.GetBoolean( 'file_maintenance_during_idle' ):
return False
if not self._controller.GoodTimeToStartBackgroundWork():
return False
return self._idle_work_rules.CanStartRequest( self._work_tracker )
else:
if not self._controller.new_options.GetBoolean( 'file_maintenance_during_active' ):
return False
return self._active_work_rules.CanStartRequest( self._work_tracker )
def _CanRegenThumbForMediaResult( self, media_result ):
mime = media_result.GetMime()
if mime not in HC.MIMES_WITH_THUMBNAILS:
return False
( width, height ) = media_result.GetResolution()
if width is None or height is None:
# this guy is probably pending a metadata regen but the user forced thumbnail regen now
# we'll wait for metadata regen to notice the new dimensions and schedule this job again
return False
return True
def _CheckFileIntegrity( self, media_result, job_type ):
hash = media_result.GetHash()
mime = media_result.GetMime()
file_is_missing = False
file_is_invalid = False
path = ''
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
except HydrusExceptions.FileMissingException:
file_is_missing = True
HydrusData.DebugPrint( 'Missing file: {}!'.format( hash.hex() ) )
if not file_is_missing and job_type in ( REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE ):
actual_hash = HydrusFileHandling.GetHashFromPath( path )
if hash != actual_hash:
file_is_invalid = True
HydrusData.DebugPrint( 'Invalid file: {} actually had hash {}!'.format( hash.hex(), actual_hash.hex() ) )
file_was_bad = file_is_missing or file_is_invalid
if file_was_bad:
error_dir = os.path.join( self._controller.GetDBDir(), 'missing_and_invalid_files' )
HydrusPaths.MakeSureDirectoryExists( error_dir )
pretty_timestamp = time.strftime( '%Y-%m-%d %H-%M-%S', time.localtime( self._controller.GetBootTime() ) )
missing_hashes_filename = '{} missing hashes.txt'.format( pretty_timestamp )
missing_hashes_path = os.path.join( error_dir, missing_hashes_filename )
with open( missing_hashes_path, 'a', encoding = 'utf-8' ) as f:
f.write( hash.hex() )
f.write( '\n' )
tags = media_result.GetTagsManager().GetCurrentAndPending( CC.COMBINED_TAG_SERVICE_KEY, ClientTags.TAG_DISPLAY_STORAGE )
if len( tags ) > 0:
try:
with open( os.path.join( error_dir, '{}.tags.txt'.format( hash.hex() ) ), 'w', encoding = 'utf-8' ) as f:
for tag in sorted( tags ):
f.write( tag )
f.write( '\n' )
except Exception as e:
HydrusData.Print( 'Tried to export tags for missing file {}, but encountered this error:'.format( hash.hex() ) )
HydrusData.PrintException( e, do_wait = False )
urls = media_result.GetLocationsManager().GetURLs()
if len( urls ) > 0:
try:
with open( os.path.join( error_dir, '{}.urls.txt'.format( hash.hex() ) ), 'w', encoding = 'utf-8' ) as f:
for url in urls:
f.write( url )
f.write( '\n' )
with open( os.path.join( error_dir, 'all_urls.txt' ), 'a', encoding = 'utf-8' ) as f:
for url in urls:
f.write( url )
f.write( '\n' )
except Exception as e:
HydrusData.Print( 'Tried to export URLs for missing file {}, but encountered this error:'.format( hash.hex() ) )
HydrusData.PrintException( e, do_wait = False )
useful_urls = []
for url in urls:
add_it = False
try:
url_class = HG.client_controller.network_engine.domain_manager.GetURLClass( url )
except HydrusExceptions.URLClassException:
continue
if url_class is None:
add_it = True
else:
if url_class.GetURLType() in ( HC.URL_TYPE_FILE, HC.URL_TYPE_POST ):
add_it = True
if add_it:
useful_urls.append( url )
if job_type == REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY:
try_redownload = False
delete_record = False
elif job_type in ( REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD ):
try_redownload = len( useful_urls ) > 0
delete_record = not try_redownload
else:
try_redownload = job_type in ( REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL ) and len( useful_urls ) > 0
delete_record = job_type in ( REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD )
do_export = file_is_invalid and ( job_type in ( REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE ) or ( job_type == REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL and try_redownload ) )
if do_export:
HydrusPaths.MakeSureDirectoryExists( error_dir )
dest_path = os.path.join( error_dir, os.path.basename( path ) )
try:
HydrusPaths.MergeFile( path, dest_path )
except Exception as e:
raise Exception( f'Could not move the damaged file "{path}" to "{dest_path}"!' ) from e
if not self._pubbed_message_about_invalid_file_export:
self._pubbed_message_about_invalid_file_export = True
message = 'During file maintenance, a file was found to be invalid. It and any known URLs have been moved to "{}".'.format( error_dir )
message += os.linesep * 2
message += 'More files may be invalid, but this message will not appear again during this boot.'
HydrusData.ShowText( message )
if try_redownload:
def qt_add_url( url ):
HG.client_controller.gui.ImportURL( url, 'missing files redownloader' )
for url in useful_urls:
HG.client_controller.CallBlockingToQt( HG.client_controller.gui, qt_add_url, url )
if delete_record:
leave_deletion_record = job_type == REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD
content_update = HydrusData.ContentUpdate( HC.CONTENT_TYPE_FILES, HC.CONTENT_UPDATE_DELETE, ( hash, ), reason = 'Record deleted during File Integrity check.' )
service_keys_to_content_updates = { CC.COMBINED_LOCAL_FILE_SERVICE_KEY : [ content_update ] }
self._controller.WriteSynchronous( 'content_updates', service_keys_to_content_updates )
if not leave_deletion_record:
content_update = HydrusData.ContentUpdate( HC.CONTENT_TYPE_FILES, HC.CONTENT_UPDATE_CLEAR_DELETE_RECORD, ( hash, ) )
service_keys_to_content_updates = { CC.COMBINED_LOCAL_FILE_SERVICE_KEY : [ content_update ] }
self._controller.WriteSynchronous( 'content_updates', service_keys_to_content_updates )
if not self._pubbed_message_about_bad_file_record_delete:
self._pubbed_message_about_bad_file_record_delete = True
if leave_deletion_record:
m = 'Its file record has been deleted from the database, leaving a deletion record (so it cannot be easily reimported).'
else:
m = 'Its file record has been removed from the database without leaving a deletion record (so it can be easily reimported).'
message = 'During file maintenance, a file was found to be missing or invalid. {} Its file hash and any known URLs have been written to "{}".'.format( m, error_dir )
message += os.linesep * 2
message += 'This may happen to more files in the near future, but this message will not appear again during this boot.'
HydrusData.ShowText( message )
return file_was_bad
def _CheckSimilarFilesMembership( self, media_result ):
mime = media_result.GetMime()
return mime in HC.FILES_THAT_HAVE_PERCEPTUAL_HASH
def _ClearJobs( self, hashes, job_type ):
if len( hashes ) > 0:
cleared_jobs = [ ( hash, job_type, None ) for hash in hashes ]
self._controller.WriteSynchronous( 'file_maintenance_clear_jobs', cleared_jobs )
def _DeleteNeighbourDupes( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
self._controller.client_files_manager.DeleteNeighbourDupes( hash, mime )
def _FixFilePermissions( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
HydrusPaths.TryToGiveFileNicePermissionBits( path )
except HydrusExceptions.FileMissingException:
return None
def _HasEXIF( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.FILES_THAT_CAN_HAVE_EXIF:
return False
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
try:
raw_pil_image = HydrusImageOpening.RawOpenPILImage( path )
has_exif = HydrusImageMetadata.HasEXIF( path )
except:
has_exif = False
additional_data = has_exif
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _HasHumanReadableEmbeddedMetadata( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.FILES_THAT_CAN_HAVE_HUMAN_READABLE_EMBEDDED_METADATA:
return False
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
has_human_readable_embedded_metadata = HasHumanReadableEmbeddedMetadata( path, mime )
additional_data = has_human_readable_embedded_metadata
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _HasICCProfile( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.FILES_THAT_CAN_HAVE_ICC_PROFILE:
return False
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
if mime == HC.APPLICATION_PSD:
try:
has_icc_profile = HydrusPSDHandling.PSDHasICCProfile( path )
except:
return None
else:
try:
raw_pil_image = HydrusImageOpening.RawOpenPILImage( path )
except:
return None
has_icc_profile = HydrusImageMetadata.HasICCProfile( raw_pil_image )
additional_data = has_icc_profile
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _HasTransparency( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.MIMES_THAT_WE_CAN_CHECK_FOR_TRANSPARENCY:
return False
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
has_transparency = HasTransparency( path, mime, duration = media_result.GetDurationMS(), num_frames = media_result.GetNumFrames(), resolution = media_result.GetResolution() )
additional_data = has_transparency
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _RegenFileMetadata( self, media_result ):
hash = media_result.GetHash()
original_mime = media_result.GetMime()
try:
path = self._controller.client_files_manager.GetFilePath( hash, original_mime )
( size, mime, width, height, duration, num_frames, has_audio, num_words ) = HydrusFileHandling.GetFileInfo( path, ok_to_look_for_hydrus_updates = True )
additional_data = ( size, mime, width, height, duration, num_frames, has_audio, num_words )
if mime != original_mime and not media_result.GetFileInfoManager().FiletypeIsForced():
if not HydrusPaths.PathIsFree( path ):
time.sleep( 0.5 )
needed_to_dupe_the_file = self._controller.client_files_manager.ChangeFileExt( hash, original_mime, mime )
if needed_to_dupe_the_file:
self._controller.WriteSynchronous( 'file_maintenance_add_jobs_hashes', { hash }, REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES, HydrusTime.GetNow() + 3600 )
return additional_data
except HydrusExceptions.UnsupportedFileException:
self._CheckFileIntegrity( media_result, REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD )
return None
except HydrusExceptions.FileMissingException:
return None
def _RegenFileModifiedTimestamp( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime in HC.HYDRUS_UPDATE_FILES:
return None
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
file_modified_timestamp = HydrusFileHandling.GetFileModifiedTimestamp( path )
additional_data = file_modified_timestamp
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _RegenFileOtherHashes( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime in HC.HYDRUS_UPDATE_FILES:
return None
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
( md5, sha1, sha512 ) = HydrusFileHandling.GetExtraHashesFromPath( path )
additional_data = ( md5, sha1, sha512 )
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _RegenFileThumbnailForce( self, media_result ):
good_to_go = self._CanRegenThumbForMediaResult( media_result )
if not good_to_go:
return
try:
return self._controller.client_files_manager.RegenerateThumbnail( media_result )
except HydrusExceptions.FileMissingException:
pass
def _RegenFileThumbnailRefit( self, media_result ):
good_to_go = self._CanRegenThumbForMediaResult( media_result )
if not good_to_go:
return False
try:
was_regenerated = self._controller.client_files_manager.RegenerateThumbnailIfWrongSize( media_result )
return was_regenerated
except HydrusExceptions.FileMissingException:
return False
def _RegenPixelHash( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.FILES_THAT_CAN_HAVE_PIXEL_HASH:
return None
duration = media_result.GetDurationMS()
if duration is not None:
return None
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
try:
pixel_hash = HydrusImageHandling.GetImagePixelHash( path, mime )
except:
return None
additional_data = pixel_hash
return additional_data
except HydrusExceptions.FileMissingException:
return None
def _RegenBlurhash( self, media ):
if media.GetMime() not in HC.MIMES_WITH_THUMBNAILS:
return None
try:
thumbnail_path = self._controller.client_files_manager.GetThumbnailPath( media )
except HydrusExceptions.FileMissingException as e:
return None
try:
thumbnail_mime = HydrusFileHandling.GetThumbnailMime( thumbnail_path )
numpy_image = ClientImageHandling.GenerateNumPyImage( thumbnail_path, thumbnail_mime )
return HydrusBlurhash.GetBlurhashFromNumPy( numpy_image )
except:
return None
def _RegenSimilarFilesMetadata( self, media_result ):
hash = media_result.GetHash()
mime = media_result.GetMime()
if mime not in HC.FILES_THAT_HAVE_PERCEPTUAL_HASH:
self._controller.WriteSynchronous( 'file_maintenance_add_jobs_hashes', { hash }, REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP )
return None
try:
path = self._controller.client_files_manager.GetFilePath( hash, mime )
except HydrusExceptions.FileMissingException:
return None
perceptual_hashes = ClientImageHandling.GenerateShapePerceptualHashes( path, mime )
return perceptual_hashes
def _ReInitialiseWorkRules( self ):
file_maintenance_idle_throttle_files = self._controller.new_options.GetInteger( 'file_maintenance_idle_throttle_files' )
file_maintenance_idle_throttle_time_delta = self._controller.new_options.GetInteger( 'file_maintenance_idle_throttle_time_delta' )
self._idle_work_rules = HydrusNetworking.BandwidthRules()
self._idle_work_rules.AddRule( HC.BANDWIDTH_TYPE_REQUESTS, file_maintenance_idle_throttle_time_delta, file_maintenance_idle_throttle_files * NORMALISED_BIG_JOB_WEIGHT )
file_maintenance_active_throttle_files = self._controller.new_options.GetInteger( 'file_maintenance_active_throttle_files' )
file_maintenance_active_throttle_time_delta = self._controller.new_options.GetInteger( 'file_maintenance_active_throttle_time_delta' )
self._active_work_rules = HydrusNetworking.BandwidthRules()
self._active_work_rules.AddRule( HC.BANDWIDTH_TYPE_REQUESTS, file_maintenance_active_throttle_time_delta, file_maintenance_active_throttle_files * NORMALISED_BIG_JOB_WEIGHT )
def _RunJob( self, media_results_to_job_types, job_status, job_done_hook = None ):
if self._serious_error_encountered:
return
cleared_jobs = []
try:
big_pauser = HydrusThreading.BigJobPauser( wait_time = 0.8 )
last_time_jobs_were_cleared = HydrusTime.GetNow()
for ( media_result, job_types ) in media_results_to_job_types.items():
big_pauser.Pause()
hash = media_result.GetHash()
if job_status.IsCancelled() or self._shutdown:
return
for job_type in job_types:
if HG.file_report_mode:
HydrusData.ShowText( 'file maintenance: {} for {}'.format( regen_file_enum_to_str_lookup[ job_type ], hash.hex() ) )
if job_done_hook is not None:
job_done_hook()
clear_job = True
additional_data = None
try:
if job_type == REGENERATE_FILE_DATA_JOB_FILE_METADATA:
additional_data = self._RegenFileMetadata( media_result )
# media_result has just changed
break
elif job_type == REGENERATE_FILE_DATA_JOB_FILE_MODIFIED_TIMESTAMP:
additional_data = self._RegenFileModifiedTimestamp( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_OTHER_HASHES:
additional_data = self._RegenFileOtherHashes( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FILE_HAS_TRANSPARENCY:
additional_data = self._HasTransparency( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FILE_HAS_EXIF:
additional_data = self._HasEXIF( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FILE_HAS_HUMAN_READABLE_EMBEDDED_METADATA:
additional_data = self._HasHumanReadableEmbeddedMetadata( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FILE_HAS_ICC_PROFILE:
additional_data = self._HasICCProfile( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_PIXEL_HASH:
additional_data = self._RegenPixelHash( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FORCE_THUMBNAIL:
additional_data = self._RegenFileThumbnailForce( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_REFIT_THUMBNAIL:
was_regenerated = self._RegenFileThumbnailRefit( media_result )
additional_data = was_regenerated
if was_regenerated:
num_thumb_refits = job_status.GetIfHasVariable( 'num_thumb_refits' )
if num_thumb_refits is None:
num_thumb_refits = 0
num_thumb_refits += 1
job_status.SetVariable( 'num_thumb_refits', num_thumb_refits )
elif job_type == REGENERATE_FILE_DATA_JOB_DELETE_NEIGHBOUR_DUPES:
self._DeleteNeighbourDupes( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_CHECK_SIMILAR_FILES_MEMBERSHIP:
additional_data = self._CheckSimilarFilesMembership( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_SIMILAR_FILES_METADATA:
additional_data = self._RegenSimilarFilesMetadata( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_FIX_PERMISSIONS:
self._FixFilePermissions( media_result )
elif job_type == REGENERATE_FILE_DATA_JOB_BLURHASH:
additional_data = self._RegenBlurhash( media_result )
elif job_type in (
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_DELETE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_TRY_URL_ELSE_REMOVE_RECORD,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_DATA_SILENT_DELETE,
REGENERATE_FILE_DATA_JOB_FILE_INTEGRITY_PRESENCE_LOG_ONLY
):
file_was_bad = self._CheckFileIntegrity( media_result, job_type )
if file_was_bad:
num_bad_files = job_status.GetIfHasVariable( 'num_bad_files' )
if num_bad_files is None:
num_bad_files = 0
num_bad_files += 1
job_status.SetVariable( 'num_bad_files', num_bad_files + 1 )
except HydrusExceptions.ShutdownException:
# no worries
clear_job = False
return
except IOError as e:
HydrusData.PrintException( e )
job_status = ClientThreading.JobStatus()
message = 'Hey, while performing file maintenance task "{}" on file {}, the client ran into an I/O Error! This could be just some media library moaning about a weird (probably truncated) file, but it could also be a significant hard drive problem. Look at the error yourself. If it looks serious, you should shut the client down and check your hard drive health immediately. Just to be safe, no more file maintenance jobs will be run this boot, and a full traceback has been written to the log.'.format( regen_file_enum_to_str_lookup[ job_type ], hash.hex() )
message += os.linesep * 2
message += str( e )
job_status.SetStatusText( message )
job_status.SetFiles( [ hash ], 'I/O error file' )
HG.client_controller.pub( 'message', job_status )
self._serious_error_encountered = True
self._shutdown = True
return
except Exception as e:
HydrusData.PrintException( e )
job_status = ClientThreading.JobStatus()
message = 'There was an unexpected problem performing maintenance task "{}" on file {}! The job will not be reattempted. A full traceback of this error should be written to the log.'.format( regen_file_enum_to_str_lookup[ job_type ], hash.hex() )
message += os.linesep * 2
message += str( e )
job_status.SetStatusText( message )
job_status.SetFiles( [ hash ], 'failed file' )
HG.client_controller.pub( 'message', job_status )
finally:
self._work_tracker.ReportRequestUsed( num_requests = regen_file_enum_to_job_weight_lookup[ job_type ] )
if clear_job:
cleared_jobs.append( ( hash, job_type, additional_data ) )
if HydrusTime.TimeHasPassed( last_time_jobs_were_cleared + 10 ) or len( cleared_jobs ) > 256:
self._controller.WriteSynchronous( 'file_maintenance_clear_jobs', cleared_jobs )
last_time_jobs_were_cleared = HydrusTime.GetNow()
cleared_jobs = []
finally:
if len( cleared_jobs ) > 0:
self._controller.Write( 'file_maintenance_clear_jobs', cleared_jobs )
def CancelJobs( self, job_type ):
with self._lock:
self._controller.WriteSynchronous( 'file_maintenance_cancel_jobs', job_type )
self._reset_background_event.set()
def ClearJobs( self, hashes, job_type ):
with self._lock:
self._ClearJobs( hashes, job_type )
self._reset_background_event.set()
def ForceMaintenance( self, mandated_job_types = None ):
if self._serious_error_encountered:
return
job_status = ClientThreading.JobStatus( cancellable = True )
job_types_to_counts = HG.client_controller.Read( 'file_maintenance_get_job_counts' )
# in a dict so the hook has scope to alter it
vr_status = {}
vr_status[ 'num_jobs_done' ] = 0
total_num_jobs_to_do = sum( ( count_due for ( key, ( count_due, count_not_due ) ) in job_types_to_counts.items() if mandated_job_types is None or key in mandated_job_types ) )
def job_done_hook():
vr_status[ 'num_jobs_done' ] += 1
num_jobs_done = vr_status[ 'num_jobs_done' ]
status_text = '{}'.format( HydrusData.ConvertValueRangeToPrettyString( num_jobs_done, total_num_jobs_to_do ) )
job_status.SetStatusText( status_text )
job_status.SetVariable( 'popup_gauge_1', ( num_jobs_done, total_num_jobs_to_do ) )
add_extra_comments_to_job_status( job_status )
job_status.SetStatusTitle( 'file maintenance' )
message_pubbed = False
work_done = False
# tell mainloop to step out a sec
self._reset_background_event.set()
with self._maintenance_lock:
try:
while True:
hashes_to_job_types = self._controller.Read( 'file_maintenance_get_jobs', mandated_job_types )
if len( hashes_to_job_types ) == 0:
break
work_done = True
if not message_pubbed:
self._controller.pub( 'message', job_status )
message_pubbed = True
if job_status.IsCancelled():
return
hashes = set( hashes_to_job_types.keys() )
media_results = self._controller.Read( 'media_results', hashes )
hashes_to_media_results = { media_result.GetHash() : media_result for media_result in media_results }
media_results_to_job_types = { hashes_to_media_results[ hash ] : job_types for ( hash, job_types ) in hashes_to_job_types.items() }
with self._lock:
self._RunJob( media_results_to_job_types, job_status, job_done_hook = job_done_hook )
time.sleep( 0.0001 )
finally:
job_status.SetStatusText( 'done!' )
job_status.DeleteVariable( 'popup_gauge_1' )
job_status.FinishAndDismiss( 5 )
if not work_done:
HydrusData.ShowText( 'No file maintenance due!' )
self._controller.pub( 'notify_files_maintenance_done' )
def MainLoopBackgroundWork( self ):
def check_shutdown():
if HydrusThreading.IsThreadShuttingDown() or self._shutdown or self._serious_error_encountered:
raise HydrusExceptions.ShutdownException()
def wait_on_maintenance():
while True:
check_shutdown()
if self._AbleToDoBackgroundMaintenance() or self._reset_background_event.is_set():
break
time.sleep( 1 )
def should_reset():
if self._reset_background_event.is_set():
self._reset_background_event.clear()
return True
else:
return False
try:
time_to_start = HydrusTime.GetNow() + 15
while not HydrusTime.TimeHasPassed( time_to_start ):
check_shutdown()
time.sleep( 1 )
while True:
check_shutdown()
did_work = False
with self._maintenance_lock:
hashes_to_job_types = self._controller.Read( 'file_maintenance_get_jobs' )
if len( hashes_to_job_types ) > 0:
did_work = True
job_status = ClientThreading.JobStatus()
i = 0
try:
hashes = set( hashes_to_job_types.keys() )
media_results = self._controller.Read( 'media_results', hashes )
hashes_to_media_results = { media_result.GetHash() : media_result for media_result in media_results }
media_results_to_job_types = { hashes_to_media_results[ hash ] : job_types for ( hash, job_types ) in hashes_to_job_types.items() }
for ( media_result, job_types ) in media_results_to_job_types.items():
wait_on_maintenance()
if should_reset():
break
with self._lock:
self._RunJob( { media_result : job_types }, job_status )
time.sleep( 0.0001 )
i += 1
if i % 100 == 0:
self._controller.pub( 'notify_files_maintenance_done' )
check_shutdown()
finally:
self._controller.pub( 'notify_files_maintenance_done' )
if not did_work:
self._wake_background_event.wait( 600 )
self._wake_background_event.clear()
# a small delay here is helpful for the forcemaintenance guy to have a chance to step in on reset
time.sleep( 1 )
except HydrusExceptions.ShutdownException:
pass
def NotifyNewOptions( self ):
with self._lock:
self._ReInitialiseWorkRules()
def RunJobImmediately( self, media_results, job_type, pub_job_status = True ):
if self._serious_error_encountered and pub_job_status:
HydrusData.ShowText( 'Sorry, the file maintenance system has encountered a serious error and will perform no more jobs this boot. Please shut down and check your hard drive health immediately.' )
return
job_status = ClientThreading.JobStatus( cancellable = True )
total_num_jobs_to_do = len( media_results )
# in a dict so the hook has scope to alter it
vr_status = {}
vr_status[ 'num_jobs_done' ] = 0
def job_done_hook():
vr_status[ 'num_jobs_done' ] += 1
num_jobs_done = vr_status[ 'num_jobs_done' ]
status_text = '{} - {}'.format( HydrusData.ConvertValueRangeToPrettyString( num_jobs_done, total_num_jobs_to_do ), regen_file_enum_to_str_lookup[ job_type ] )
job_status.SetStatusText( status_text )
job_status.SetVariable( 'popup_gauge_1', ( num_jobs_done, total_num_jobs_to_do ) )
add_extra_comments_to_job_status( job_status )
job_status.SetStatusTitle( 'regenerating file data' )
if pub_job_status:
self._controller.pub( 'message', job_status )
self._reset_background_event.set()
try:
media_results_to_job_types = { media_result : ( job_type, ) for media_result in media_results }
with self._lock:
self._RunJob( media_results_to_job_types, job_status, job_done_hook = job_done_hook )
finally:
job_status.SetStatusText( 'done!' )
job_status.DeleteVariable( 'popup_gauge_1' )
job_status.FinishAndDismiss( 5 )
self._controller.pub( 'notify_files_maintenance_done' )
def ScheduleJob( self, hashes, job_type, time_can_start = 0 ):
with self._lock:
self._controller.Write( 'file_maintenance_add_jobs_hashes', hashes, job_type, time_can_start )
self._wake_background_event.set()
def ScheduleJobHashIds( self, hash_ids, job_type, time_can_start = 0 ):
with self._lock:
self._controller.Write( 'file_maintenance_add_jobs', hash_ids, job_type, time_can_start )
self._wake_background_event.set()
def Shutdown( self ):
self._shutdown = True
self._wake_background_event.set()
def Start( self ):
self._controller.CallToThreadLongRunning( self.MainLoopBackgroundWork )
def Wake( self ):
self._wake_background_event.set()
Reference in New Issue View Git Blame Copy Permalink