A broken PDF should be one row in a report

Description and use

A single corrupted PDF should not halt a large batch compression workflow. PDF Batch Resume processes each PDF in isolation, logs errors, and resumes the queue without discarding successful outputs.

Problem

A single corrupted PDF should not halt a large batch compression workflow.

Outcome

The tool successfully compressed a chart PDF, reported an error for a damaged PDF, and resumed processing after replacing the damaged file, achieving a clean exit code 0.

Workflow

• Launch supervisor script
• Supervisor spawns a worker subprocess for each PDF with a 120‑second timeout
• Worker processes PDF using pypdf and Pillow
• If timeout or error occurs, worker exits with nonzero code
• Supervisor records error and moves to next file
• Worker writes journal entry via temporary JSON file
• If output exists, compare source hash, options, and output hash; report conflict if mismatched
• On subsequent runs, resume from last unprocessed or conflicted file
• After replacing damaged PDF, re-run to complete batch

Evidence

Each PDF gets a fresh worker process.

Source

The default timeout is 120 seconds per PDF.

Source

If an output exists but does not match that evidence, the tool reports a conflict.

Source

The first two runs returned exit code 2: two complete, one needing review.

Source