Processing Pipeline¶

1) Input collection¶

tests/build.py parses CLI args, resolves inputs, and loads file content strings.

For URLs (Google Drive), it downloads to temp first, then reads local content.

2) XML parsing¶

DrawioStrategy calls read_drawio (tricc_oo/parsers/xml.py), which parses the full XML payload via lxml.

The parser expects draw.io XML (<mxfile ...>). Non-XML or HTML content fails early.

3) Page/activity creation¶

create_activity in tricc_oo/converters/xml_to_tricc.py:

• resolves root node (start or activity_start)
• creates activity object
• loads edges and nodes
• enriches nodes (hints/help/media)
• assigns process/start-page metadata

4) Experimental page control¶

When root status is experimental, detailed activity processing is skipped/limited by design.

This allows draft pages to stay in the model without entering full downstream generation.

5) Graph linking¶

DrawioStrategy.linking_nodes wires page graph semantics:

• normal next/prev links
• goto traversal and activity instances
• link_out to link_in resolution
• loop warnings and edge validation

6) Output strategy execution¶

After project graph creation, selected output strategy runs (-O):

• XLSForm variants
• CHT variants
• OpenMRS / FHIR / HTML / DHIS2 (availability varies by maturity)