• More robust FIND_LINKS_REGEX expression by @danishashko (#130)
• Maintenance: code hardening and modernization, more tests (#124, #125, #127, #131, #132)
• Breaking: UrlStore() parameter trailing renamed to trailing_slash
• Drop support for Python 3.8 and 3.9
• filter_links() and load_store() are now exported at the top level

• UrlStore.get_download_urls(): timelimit removed, fix type hints (#119, 19c580e)
• extract_links(): deprecate base_url parameter (#121)
• setup: simplify workflow (#118)

• UrlStore compression: make bz2 and zlib optional, update pickle protocol (#113)
• extract_links(): review and document, add deprecation warning for base_url argument (#115)
• maintenance: add __all__ to init.py and lint code (#116)

• parsing: validate netloc with port number by @naz-theori in #104
• cleaning: fix handling of apostrophes (#107)
• maintenance: deprecate Python 3.6 & 3.7, add pyproject.toml setup file (#59, #105)

• more compact UrlStore: use bytes instead of str for URL paths (#88)
• UrlStore maintenance: deprecate timelimit argument (#101)
• maintenance: simplify code (#103)
• support for Python 3.13

• replace langcodes by babel and use its information on locales (#89, #92)
• simplified and faster code: domain extraction, cleaning, filters and UrlStore (#90, #93, #94, #95)
• UrlStore: better url batches, replace timelimit parameter by time_limit (#91)
• maintenance: update readme and convert it to markdown (#97)

• license change from GPLv3+ to Apache 2.0 (#81)
• UrlStore: write() method and load_store() function added (#83)
• add parameter trailing_slash to keep of discard slashes at the end of URLs (#52)
• maintenance: fix whitespace in clean_url() (#77), simplify code (#79)

• IRI to URI normalization: encode path, query and fragments (#58, #60)
• normalization: strip common trackers (#65)
• new function is_valid_url() (#63)
• hardening of domain filter (#64)

Full Changelog: v0.9.4...v0.9.5

• new UrlStore functions: add_from_html() (#42), discard() (#44), get_unvisited_domains
• CLI: removed --samplesize, use --sample with an integer instead (#54)
• added plausibility filter for domains/hosts (#48)
• speedups and more efficient processing (#47, #49, #50)
• fixed handling of relative URLs with @feltcat in #46
• fixed bugs and ensured compatibility (#41, #43, #51, #56)
• official support for Python 3.12

Full Changelog: v0.9.3...v0.9.4

• more efficient URL parsing (#33)
• refined link extraction and link filters (#30, #36)
• more efficient normalization (#32)
• more efficient sampling strategy (#31, #35)
• added meta function to clear LRU caches (#34)
• added parallel option in command-line interface (#37, #39)
• added get_unvisited_domains() method to UrlStore (#40)

Full Changelog: v0.9.2...v0.9.3