3.0.1

Added

• Add CLI flag to customize ZIM Name (#340)

Fixed

• Add missing flags in offliner definition (#339)

3.0.0

Breaking

• Remove any optimization logic + use of S3 cache, dropping --optimization-cache and --use-any-optimized-version flags (#300)
• Move to another default mirror + add CLI flag to select mirror to use, dropping --rdf-url flag and adding --mirror_url CLI flag (#301)
• Reengineer the scraper (#312)
  • Source list of books from CSV at https://gutenberg.pglaf.org/cache/epub/feeds/pg_catalog.csv.gz for quicker startup
  • Stop downloading big RDF archive and download only needed individual RDF files (saves lot of time when only few books are requested ; looks like negligible penalty when many books are needed)
  • Get rid of SQLite database used to persist data across runs, too much maintenance effort and impact on the filesystem for limited benefit
  • Get rid of the option to generate one ZIM per language ; scraper now always produce one single ZIM
  • Data is now directly transferred to the ZIM, without touching the filesystem
  • Get rid of the "steps" approach, not used (anymore?) in production and difficult to maintain
  • Many CLI options removed: --use-any-optimized-version, --zim, --download, --parse, --prepare, -m/--one-language-one-zim, --dlc, -d/--dl-folder, -e/--static-folder
• Drop support for Gutenberg bookshelves and add support for LibraryOfCongress (#265)

Added

• Add standard --output option to control where ZIM files are written (#314)
• Add ability to generate smaller selection (#184)
• Log failing book ID in case of fatal error (#333)

Changed

• Finalize implementation of scraper progress (#289)
• Move to another default mirror + add CLI flag to select mirror to use (#301)
• Split build logic in Dockerfile to separate dependencies layer from scraper code layer (#302)
• Prepare structure for TranslateWiki (#312)
• Remove « .html » extension (#166)
• Configure libzim verbosity based on --debug flag (#326)

Fixed

• Stop ignoring HTML illustrations containing cover in their name (#270)
• Fix JS/JSON files generation (#297, #298)
• Fix navigation to bookshelves with special characters (#305)
• Bookshelves with special characters cannot be opened (#306)
• Fix internationalization of the "Copyrighted" license label (#253)
• Properly compute (including sorting) ZIM Language items + allow to override with --zim-languages (#323)
• Fix support for ZIMs without full-text index (#326)

2.2.0

Added

• Add support for --debug flag to output debug logs
• Add support for -L long_description flags
• Add request timeout for util.py (#197)
• Add Booklanguage DB to support multi-languages books (#218)
• Add RTL support to UI (#248)
• Add language filter to combobox for requested languages (#249)

Changed

• Simplify Gutenberg scraping (no more rsync, no more fallback URLs / filenames) (#97)
• Prefer EPUB 3 to EPUB (#235)
• Do not force the presence of PDF format for all books (#160)
• Replace usage of os.path and path.py with pathlib.Path (#195)
• Finalize ZIM metadata title translations and multilingual detection (#229)
• Replaced magic number with named constant and clarified comment regarding book ID URL rules (#196)
• Replace print and pp calls with logger (#192)
• Update to Python3.13
• Update python-scraperlib to 5.1.1 and dependencies (#188)
• Rename Book DB table fields (#199)
• Update multi-resolution favicons (#165)

Fixed

• Fix regression on missing HTML content (#219)
• Simplify the logger name (used gutenberg2zim instead of gutenberg2zim.constants) (#206)
• Add retry logic on book downloads (#254)
• Fix UI and navigation glitches on bookshelves (#262)
• Remove dependencies on binaries + buggy pngquant (#257)

2.1.1

Added

• Publisher ZIM metadata can now be customized at CLI (#210)

Changed

• Publisher ZIM metadata default value is changed to openZIM intead of Kiwix (#210)

Fixed

• Do not fail if temporary directory already exists (#207)
• Typo in Scraper ZIM metadata (#212)
• Adapt to hatchling v1.19.0 which mandates packages setting (#211)

2.1.0

Changed

• Fixed regression with broken filters on on multiple-languages ZIM (#175)
• Fixed Name metadata that was incorrectly including period (#177)
• Fixed Language metadata (and filename) for multilang ZIMs (#174)
• Using zimscraperlib 2.1.0
• Using localized Title and Description metadata (#148)
• Fixed regression with epub files stored as application/zip (#181)
• Adopt Python bootstrap conventions, especially migration to hatch instead of setuptools and Github CI Workflows adaptations (#190)
• Removed inline Javascript in HTML files (#145)

Fixed

• Support single quotes in author names (#162)
• Migrated to another Gutenberg server (#187)
• Removed useless file languages_06_2018 (#180)

Removed

• Removed Datatables JS code from repository, fetch online now (#116)
• Dropped Python 2 support (#191)

2.0.0

Added

• Porgress report using --stats-filename

Changed

• Updated dependencies, including zimscraperlib (2.0)
• Now creating no-namespace ZIM with Illustration
• Fixed/reduced sqlite timeouts
• Better handling of rsync'd list of URLs
• RDF files are not extracted to disk anymore (faster on selections)
• Remove all Urls from DB before processing rsync'd ones
• Fixed --concurrency short flag (now -c)
• Docker image now uses python3.11
• DB don't use a separate Format table anymore

Removed

• Dependency to zimwriterfs binary.
• -r/--rdf-folder flag: rdf not extracted to disk anymore
• --export: HTML files not written to disk first anymore
• --dev: idem
• Binaries from docker images: jpegoptim, pngquant, gifsicle, zip, curl, p7zip