count-zh — Chinese character (字數) count

The framework ships ./fw count-zh (run from the course root). This skill is the entry point for "how many words / 字數" requests on Chinese-language work; the script is the implementation. Run — and extend — ./fw count-zh rather than counting 字數 by hand, with wc, or with len(); each gives the wrong number for Chinese (see below).

Why a dedicated tool

Chinese text has no spaces between words, so the usual reflexes are wrong:

• wc -w counts whitespace-delimited tokens → drastically undercounts Chinese.
• len(text) / a raw character count → overcounts (includes punctuation, English, digits, whitespace, markup).

count-zh.py counts CJK Unified Ideographs only ([一-鿿]), excluding punctuation, English letters, digits, and whitespace — which is what a 字數 limit on an assignment or journal actually means.

Usage

./fw count-zh <file>        # prints the integer 字數 to stdout

• Count a draft: ./fw count-zh <unit>/homework.qmd
• Check against a limit: run it, compare to the assignment's 字數 requirement, and report the gap.

Notes

• It counts the raw file. Markdown/LaTeX/Quarto markup keywords are English, so they add no CJK and don't inflate the count — but Chinese inside YAML front matter (e.g. a Chinese title:) does count. For a strict body-only number, point it at the prose, not a generated wrapper.
• One file at a time. For a multi-chapter thesis, sum the per-file counts.