umputun / tg-spam / 29293948129

84%

push

github

feat: prohibited languages spam detector (#427)

Add a spam detector that blocks messages written in prohibited Unicode scripts.
Configure a blocklist of scripts (chinese=Han, russian=Cyrillic, etc.); a message
with enough letters of a banned script is flagged as spam. It is a hard block that
bypasses the LLM, runs for non-approved users (everyone in paranoid mode), and is
disabled by default. Wired through the detector, CLI/env flags, DB-backed settings,
the web settings UI, and an e2e test.

The prohibited-language check scores only the user's authored text, not quoted or
reply-to content, so replying to a foreign-script message cannot trigger a ban.
Cross-field settings validation is shared via config.Settings.Validate() across
startup, save-config, and the web save handler.

148 of 156 new or added lines in 9 files covered. (94.87%)

8826 of 10554 relevant lines covered (83.63%)

236.63 hits per line