v.2.11.7

Breaking changes:

• Renamed --size option to --thumbnail-size
• Renamed --quality option to --thumbnail-quality
• --thumbnail-size option no longer accepts negative numbers to disable thumbnails (User --thumbnail-count 0 instead)
• --content-size option no longer accepts negative numbers to disable parsing file contents (Use --content-size 0 instead)

New features:

• Add --thumbnail-count option. When set to a value >= 2, thumbnails for a video preview slideshow are generated. (A reasonable number of images, around 10-30, is recommended). When set to 0, thumbnails are disabled entirely. (#232)
• Added configurable delay in ms between frames of the video preview created with the --thumbnails-count option
• Added #/file page. Lets users or third party applications link to a specific document in the sist2 interface. (#250)
• Added --mem-throttle scan option to throttle the file parser threads when the total amount of memory consumed exceeds a certain amount, in MB. (#245, thanks @yatli !)
• Reworked incremental scan (#238, thanks @yatli !)
  • Option to use --incremental-index to only send the incremental difference between the last scan to Elasticsearch (much faster indexing!)
  • Build an index of deleted documents since the last scan, send DELETE commands to Elasticsearch to remove such documents
• "Update the Media Types tree in real time" option is disabled by default
• Add option to use a date picker instead of a slider to filter file modification dates (#239)
• Fix segmentation fault when a very long user script is specified (#234)
• Replace some error messages with a FATAL log string, return non-zero error code on error (#235)

Bug fixes:

• Fix some CSS visual glitches (#248)
• Fix checkbox in index picker not working (#247)
• Fix infinite reloading problem with mime tree (#237)
• Fix lightbox caption for top-level file names (#254)

v2.11.6

Breaking changes:

• --ocr option is renamed to --ocr-lang.

New features:

• Add Simplified Chinese UI translations zh-CN (#180)
• Add --lang argument for the web module. Used to set a default UI language (en by default)
• Allow specifying multiple OCR languages with + separator (#228)
• Add --ocr-ebooks and --ocr-images options. Add option to OCR images. (#217)
• Add option to update the Media Type tab in real time based on the current query (#213)
• Add Media Type break down table in the Details pane. Add button to copy TSV (#213)
• Add option to scan files from a list using the --list-file rather than recursive directory scan (#124)

Fixes:

• Fix segmentation fault when scanning some RAW files (#219)
• Update Tesseract OCR data paths (#212)
• Fix race condition in incremental scan (#223)
• Statically link libasan for debug build (#222)
• Fix memory leak in ebook.c when OCR is enabled
• Fix visual glitch when hovering GIF files
• Fix configuration option of slideshow slide duration
• Localize the add/delete tag buttons
• Fix index picker CSS in dark mode (#218)

2.11.5

Bug fixes:

• Fix visual glitch in console logs when indexing

v2.11.4

Bug fixes:

• Read version from ES before indexing, handle legacy versions, add notice & debug info (#203)
• Don't show resolution badge on narrow images (#201)
• Indicate when there are no results (#204)

New features:

• Print progress bar in index module (#144)
• Shift click & select all/none in index picker (#206)
• Add natural sort by file name (#207)

Other:

• Update to libmobi 0.9
• Update libmupdf 1.19.0-rc2

v2.11.3

New features:

• Add --checksums argument to calculate checksums. When enabled, duplicate files are hidden in the search results (this behaviour can be toggled in the Configuration page)
• Rewind buffer support for files in archives with unknown file extension (#189)
• Add basic support for JSON/NDJSON files (#187)

Bug fixes & enhancements:

• Rework index picker (#199)
• Don't render thumbnail component in document info modal when there is no thumbnail (#194)
• Set alternate list item color for sub-documents in list mode (#190)
• Fix sidecar files (#193)
• Better error handling in store_write (#195)
• Fix query error related to max_analyzed_offset (#198)

v2.11.2

• Add basic support for WordPerfect files (#159)
• Include libasan5 in Docker images (#185)
• Parse page numbers from .docx files (#179)
• Add v prefix in web UI navbar (v2.11.2 instead of 2.11.2)

v2.11.1

• Fix bugs related to the signal handler (#177)
• Fix UI bug with tags/mime tree (#183)

v2.11.0

Big release!

Re-wrote the web UI from scratch:

• Search arguments are updated in the URL bar in real-time, useful for sharing/bookmarking complicated queries
• Infinite scrolling in Grid mode is more seamless
• Image/video viewer with more features (fullscreen, thumbnail gallery, keyboard bindings, slideshow)
• Option to tweak the number of results per page
• Option to sort by 'random'
• Localized! (currently only English and French available) (See #180 for information on how to help add more languages)
• Add --tagline argument to customize the tagline in the navigation bar
• Switching between 'Grid' and 'List' display modes no longer reloads the search results
• Renamed "Mime Types" to "Media types" in the UI
• Document contents are loaded asynchronously when opening the document info modal
• "Theme" button moved to the configuration page. Changing the theme no longer requires a page reload and does not create a cookie.
• Add debugging information in the configuration page.
• Reworked List mode styling
• Reworked tags:
  • Simplified tag color picker with a small palette
  • Tags are now indexed with path_hierarchy analyzer. Searching for a.b will also match all of a.b.*
  • Tag colors are saved in the tag.#RRGGBB format instead of tag#RRGGBB
  • Added option to query tags with "OR" operator or "AND" operator
  • Removed the "any" checkbox in tags & mimetype tabs; selecting no element implies "any"

Other:

• Reorganized the index encoding on disk: document information is now saved as a single compressed newline-delimited JSON file, index files are now much smaller! (#170)
• Decimal GPS tags are saved as floating point number (#171)
• Fixed some valgrind warnings for invalid reads
• Add Cache-Control header in /t/* endpoint to explicitly tell browser to cache thumbnails
• Add --dev argument for web module to serve html/js files from disk (used for development)
• Ebook thumbnails now scale with both --size and --quality option

v2.10.3

• Add some error handling in walk_directory_tree() (#147)
• Return 404 when thumbnail is not found (#167)
• Add --fast-epub argument (#169), workaround for (#129 #168)
• Faster thread pool destruction (#138)

2.10.2

• Fix --rewrite-url bug in the web module (#165)
• Fix --auth bug in the web module