Skip to main content

Module scan

Module scan 

Source
Expand description

xil scan β€” pre-flight check of a production script. Port of XILP000_script_scanner.py.

StructsΒ§

Args πŸ”’
Scan πŸ”’

ConstantsΒ§

ALLOWED_TAGS πŸ”’
Native paralinguistic cues Chatterbox Turbo renders. Anything else is silently stripped at generation time, which is what makes a near-miss worth flagging. Mirrors ALLOWED_TAGS in chatterbox_turbo_worker.py.
NEAR_MISS_CUTOFF πŸ”’
Similarity floor for suggesting a Turbo cue. Genuine mistakes score
TAG_ALIASES πŸ”’
Variants difflib scores too low because the wording diverges.

StaticsΒ§

INLINE_TAG πŸ”’

FunctionsΒ§

all_scripts πŸ”’
glob(dir/**/*.md, recursive=True) or glob(dir/*.md).
arr πŸ”’
backfill_cast πŸ”’
build_scan πŸ”’
The full scan result, in the key order the JSON output uses.
command
The clap definition behind --help, for man pages.
execute πŸ”’
format_report πŸ”’
harvest_cast πŸ”’
inner_text πŸ”’
line[1:-1].strip() β€” the bracket interior, as the scanner slices it.
is_all_caps_candidate πŸ”’
A bare ALL-CAPS line worth classifying.
line_list πŸ”’
load_and_normalize πŸ”’
Read and apply the parser’s two-pass normalization.
pad πŸ”’
Pad to width by character count, as Python’s :<N does.
resolve_speakers_path πŸ”’
--speakers, else configs/{slug}/speakers.json when --show names one.
run
scan_ambience_coverage πŸ”’
Every looping AMBIENCE needs a stop marker.
scan_direction_texts πŸ”’
Audit direction texts against an existing SFX config.
scan_paralinguistic_tags πŸ”’
Inline [tags] that look like misspelled Turbo cues.
scan_script πŸ”’
Classify every ALL-CAPS candidate into section, speaker or unknown.
scan_span_pairing πŸ”’
Every <marker> ENGAGES must have a matching DISENGAGES.