metaindex/CHANGELOG.md

Changelog

This file contains the changes made between released versions.

The format is based on Keep a changelog and the versioning tries to follow Semantic Versioning.

2.3.0

Added

• Very basic commandline interface. Try metaindex cli

Changed

• Use ~/.local/state/metaindex/ instead of ~/.cache/metaindex/ for the database
• Don't keep the fulltext in the CacheEntrys
• Fix umask of the metaindex database

Fixed

• YAML sidecar file with a date would cause errors

2.2.2

Fixed

• Fixed issues with last_modified and the new unified timestamp format

2.2.1

Changed

• date field in sidecar files will be split into date and time upon loading, if possible, and saved as a merged date field upon saving

Fixed

• Overriding values by setting them to none/null in the sidecar file was not actually doing anything
• Language hinting for OCR was too demanding; now also short ISO 639 codes work

2.2.0

Changed

• Use PyPDF instead of pdfminer for PDF indexing
• OCRFacility.run can be given a language to work with to prevent guessing
• Collection and side-car files are processed first now, instead of last to allow defining languages
• The optional callback, that can be used to inspect the status of the indexer, sees the un-reduced CacheEntry. Makes it more useful for testing rule indexers and such.
• Timestamp-like occurrences (like filetags.date) have been split into .date and .time. You can search each respectively with date: or time:

Added

• OCRFacility can be queried for supported languages
• Another level of verbosity when indexing: -vvvvv, prints also all found metadata

Fixed

• Server could crash when read-only connections were made and the cache directory doesn’t exist yet
• Some code duplication removed for writing YAML and JSON sidecar files

2.1.1

Changed

• Server will return results even if the match score is very low

Fixed

• The rule based indexer will be more resilient against non-matching find directives

2.1.0

Added

• Configuration option for server, index-new-tags

2.0.0

Changed

• Drastic refactoring, back-end changed to xapian

1.5.0

Changed

• Updated the ABC indexer to match the specs of version 2.1 better

1.4.0

Changed

• Cache.cleanup accepts a list of paths to limit the clean-up to
• metaindex index -m expects a list of paths to clean up.

Added

• Support for metadata from GPX files
• Support for metadata from OpenDocument files (odt, ods, etc.)
• Support for office open XML files (docx, etc.)

1.3.0

Changed

• CacheEntry.__contains__ allows to check for presence of a tag with a particular value
• When logging to a file, metaindex will now rotate the log file at a 4 KB size limit
• resolve_sidecar_for can ignore existing collection sidecars

1.2.0

Added

• Indexer for HTML files (metadata)
• Indexer for ComicBook files using ComicInfo.xml (cbz, cbt, cbr if python-unrar is installed)
• File size humanizer
• Support special value * in synonym definitions and ignore-tags to expand the existing definition

Fixed

• The JSON sidecar writer could break when encountering unexpected metadata types (like datetime)

Changed

• Changed default synonym 'tags' to 'tag'

1.1.1

Fixed

• Audio indexer did not expect that mutagen can return lists of values per key

1.1.0

Changed

• Creating a Configuration instance will start with the configuration defaults
• Some properties moved from CacheBase to Configuration
• The exiv/image indexer expands values that are lists or dictionaries

Added

• final directive for rule based indexer

Fixed

• Rule based indexer was a bit broken
• Epub indexer follows the standard to determine the root OPF file(s)
• Addons were not able to use user-installed system packages

1.0.1

Fixed

• The SQL backend stored the humanized value instead of the raw value

1.0.0

Changed

• Dropped multidict dependency
• Introduction of shared.CacheEntry (instead of a named tuple)
• IndexerBase API changed

Added

• Support for humanizers
• ABC notation metadata indexing

0.8.0

Added

• bulk_rename operation in MemoryCache.
• Implemented forget in all caches

Fixed

• When querying all sidecar files for a file, metaindex would return the file itself, when asking this for a sidecar file (odd scenario, still bad)

0.7.1

Fixed

• find_all_sidecar_files would not return all sidecar files (ignoring those that are not writable)
• General.ignore-dirs, as documented, is now actually considered

Changed

• Added a few more files and directories to ignore-dirs and ignore-files
• Fixed the unit tests on file collection

0.7.0

Added

• Function to insert a whole set of new files in the MemoryCache

0.6.0

Added

• Function to iterate through all sidecar files of a file
• Function to rename entries in the database
• stores.as_collection to ease the pain when writing collection metadata sidecars

Changed

• The filename is no longer stored as a separate metadata tag in the database

0.5.1

Fixed

• Collection metadata files could be missed in some cases

0.5.0

Added

• ThreadedCache and MemoryCache both support the keys function, too

Fixed

• MemoryCache's refresh function would forget the refreshed files
• Various issues where MemoryCache's API would not be quite the same as Cache's API.

0.4.0

Added

• MemoryCache for fast queries in multi-threaded applications
• Query.matches to run a query against metadata and see if the query matches

0.3.0

Added

• ThreadedCache for use of the cache in multi-threaded applications

Changed

• Major refactoring of the cache.refresh routine
• Sidecar files are now handled via indexers (running last in order)
• Logging is no longer done through the default logger

0.2.0

Added

• Indexer that applies rules to the fulltext of a document to extract meta data
• OCR support
• YAML style metadata files are supported
• Indexer addons

Changed

• API for Indexer changed slightly
• stores.get, stores.get_for_collection accept byte streams, too

Removed

• add, remove, and edit commandline parameters are gone. Use metaindexmanager for these functions instead

0.1.0

• Initial release