CODES – Code Frequency Table

Status: Current

Purpose

Tabulates the frequency and distribution of coding annotations found on %cod: dependent tiers, organized by speaker. This is useful for analyzing hand-coded behavioral or discourse annotations attached to transcripts.

Codes on %cod: tiers typically use colon-separated hierarchical structure (e.g., AC:DI:PP), but this implementation treats each whitespace-delimited token as a single code string without parsing the internal hierarchy.

The CLAN manual does not appear to contain a standalone CODES section. The closest direct evidence is the CHAINS section, which documents coding tiers and the codes.ord file used to order $-prefixed codes.

Usage

chatter clan codes file.cha
chatter clan codes --speaker CHI file.cha
chatter clan codes --format json file.cha

Options

Option	Description
`--speaker <CODE>`	Include speaker
`--max-depth <N>`	Maximum depth of code parsing (0 = all levels)
`--format <FMT>`	Output format: text, json, csv, clan

Output

Per-speaker frequency tables listing each code and its count, plus a per-speaker total and a grand total across all speakers.

Differences from CLAN

Manual coverage gap: the CLAN manual has no standalone CODES section, so this chapter is based on implemented behavior, golden tests, and nearby coding-tier documentation such as CHAINS.
Code extraction: Codes are extracted from a clan-local semantic %cod layer built from the parsed AST, not from raw line text.
Selector handling: %cod extraction preserves optional selectors like <w4> or <w4-5> as item scope rather than counting them as standalone codes.
Hierarchy handling: Each whitespace-delimited token is treated as a single code string; colon-separated hierarchy is preserved but not parsed into sublevels.
Output formats: Supports text, JSON, and CSV formats (CLAN produces text only).
Deterministic ordering: BTreeMap ordering ensures deterministic output across runs.
Golden test parity: Verified against CLAN C binary output.