Define a transcript's layout by marking up a sample

[benfoley]

Transcript parsing (ca-data-prep) hard-codes one document convention: a Speakers: block of CODE: name [alternate] (demographic) #id lines, upper-case section markers, and [turn] CODE: text rows. A collection whose transcripts follow a different layout has no way to describe it short of editing process.js.

Wanted: a plugin that lets a person define the layout by example —

• take a transcript as pasted text or an uploaded file;
• mark regions of the document as header metadata, speaker info and main;
• mark up individual rows: in speaker rows, which characters are the code / name / alternate name / affiliation / id; in content rows, the turn number / speaker / text;
• generate regular expressions from that markup, usable to parse other documents in the same format;
• save them as standing configuration under _config/.