jbrowser.services.adfilter

Generated from the source code of JBrowser 1.5.4.

jbrowser/services/adfilter.py · 370 lines

Adblock Plus / EasyList filter engine: URL-pattern blocking, exceptions and element hiding.

The downloaded lists (see privacy.BLOCKLIST_SOURCES) are split by the updater into:

• blocklist.txt: plain ||domain^ rules and hosts-file entries. Fast set lookups in jbrowser.services.privacy.Blocklist.
• filters.txt: everything else this engine understands. That is URL patterns such as /ads/banner*.js$third-party, exceptions (@@), and element-hiding rules (##.ad-slot, example.com##.promo), which hide the empty boxes and "Advertisement" frames that domain blocking alone leaves behind.

Matching uses the same trick as Adblock Plus: every filter is filed under one "token" (a word it must contain), so a request only checks the handful of filters whose token appears in its URL. Unsupported syntax (procedural cosmetics, scriptlets, regex filters) is skipped.

Constants#

NetworkFilter#

class NetworkFiltersource

NetworkFilter(pattern: str, types: frozenset, third: bool | None, include: tuple, exclude: tuple, important: bool)

NetworkFilter.matches#

matches(url: str, first: str, rtype: str, third: bool) -> boolsource

FilterEngine#

class FilterEnginesource

Parsed filters.txt: network filters indexed by token, plus element-hiding rules.

FilterEngine.from_text#

class method from_text(text: str) -> 'FilterEngine'source

FilterEngine.load#

class method load(path: Path) -> 'FilterEngine | None'source

FilterEngine.site_allowed#

site_allowed(first: str) -> boolsource

The list itself exempts this site from blocking (@@||site^$document).

FilterEngine.blocking_filter#

blocking_filter(url: str, first: str, rtype: str, third: bool) -> NetworkFilter | Nonesource

FilterEngine.excepted#

excepted(url: str, first: str, rtype: str, third: bool) -> boolsource

FilterEngine.cosmetic_data#

cosmetic_data() -> dictsource

Element-hiding rules in the compact form the page script uses (see engine/js.py cosmetic_js): the script picks the rules for its own site as the page starts. That is the only way they apply to the very first page of a site, because a script added while a navigation is already under way only takes effect from the next one.

FilterEngine.css_for#

css_for(host: str) -> strsource

The style sheet the page script builds for host (for tests and diagnostics).

Functions#

split_list#

split_list(text: str) -> tuple[set[str], list[str]]source

Split a downloaded list into plain domains (for blocklist.txt) and the lines the FilterEngine handles (for filters.txt). Hosts files only produce domains.