Module jukugo 

Jukugo (熟語) table — JP compound words with reading + freq.

Generated by tools/nihongo/build_jukugo_rs.py from the supplemental TSV. DO NOT edit by hand — re-run the script. Corrections go to the TSV.

Codepoint-identity-with-SC constraint REMOVED — Inputx’s JP plugin is professional-grade now, so shinjitai-only compounds (経済/愛情/電話/…) are first-class.

Structs

JukugoEntry

Constants

JUKUGO_TABLE

Functions

lookup_by_reading

lookup_by_reading_prefix

Prefix-prediction lookup: jukugo whose reading STARTS WITH romaji but is longer (the user is mid-typing toward it — shinjuk → 新宿/しんじゅく). Yields (kanji, freq, reading_len_bytes) so the composite layer can score by proximity = typed_len / reading_len. Excludes exact matches (== reading), which lookup_by_reading already covers. Linear scan over JUKUGO_TABLE (~27k; ~27µs measured, same as exact lookup).