---

title: "External preflight imports" description: "Import third-party preflight reports (Enfocus PitStop, callas pdfToolbox, Adobe Acrobat Pro Preflight) so their findings join LintPDF's output — useful for vendor-neutral consolidated reporting." group: "Reference" order: 9

External preflight imports

LintPDF can ingest preflight reports produced by other vendors and fold their findings into the same Job shape that native analyzers emit. Useful when you already run PitStop or pdfToolbox upstream and want a single consolidated audit trail without re-running the file through every checker.

Supported formats

Five format tokens map 1:1 to parsers under src/lintpdf/imports/:

Token	Vendor / source	Parser
pitstop_xml	Enfocus PitStop / PitStop Server reports	pitstop.py
callas_json	callas pdfToolbox JSON export	callas.py
callas_xml	callas pdfToolbox XML export	callas.py
acrobat_xml	Adobe Acrobat Pro Preflight XML export	acrobat.py
lintpdf_json	LintPDF native re-import (round-trip / replay)	lintpdf_native.py

The parser-protocol contract lives in base.py: each parser exposes a format class attribute (one of the tokens above) and a parse(payload: bytes) -> ImportedReport method.

Auto-detection

If you don't set external_format explicitly on a submission, the engine sniffs the payload via detect.py:

• JSON payloads: routed to lintpdf_json if the document carries schema_version + findings, or to callas_json if it has hits / results plus pdfToolbox-shaped metadata.
• XML payloads: dispatched by root element — <Preflight> / <AcrobatReport> → acrobat_xml, <EnfocusReport> / <PitStopReport> → pitstop_xml, <preflight_report> / <callasreport> → callas_xml. Heuristic fallbacks check for signature child elements (<Hit>, <ResultItem>, <Problem>).

When detection fails the parser raises ParserError and the API responds 422 carrying the field path that broke. Per the engine contract: parsers must fail cleanly — never swallow a parse error and emit zero findings, since the caller would assume their report was clean.

Adding a new format

Per the engine's CLAUDE.md contract, adding a built-in format is exactly three changes:

1. New parser under src/lintpdf/imports/<vendor>.py, registered in src/lintpdf/imports/detect.py.
2. New round-trippable sample fixture committed alongside the parser's tests.
3. New row in the external_format enum appendix and the supported-formats table in the public docs.

Vendor-specific implementations stay isolated — the engine's analyzer tier never touches the parser modules; everything funnels through the unified ImportedReport shape.