Workflow:
1. Load template JSON (defines section structure)
2. Paste multiple text-format forms into textarea
3. Parser splits by title mark, collects tokens per section key
4. Merged output in text or JSON format
Key changes:
- splitIntoInstances(): split at each title-mark line boundary
- parseInstance(): match section headers by mark+name, tokenize value lines
- mergedToText(): walk template tree, output tokens with per_line chunking
- mergedToJSON(): reconstruct ROOT/SECTION/VALUE tree with merged tokens
- Template view: compact inline chip display of section structure
- Instance counter badge shows how many forms were detected
- JSON export: each token becomes a separate VALUE node (data_type:string)
- per_line attribute on SECTION controls line wrapping in text output
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
gm
5a86d52813