[medium] fix: [import_mod/ocr] guard against zero-page PDFs

[elhoim]

BLUF — A zero-page PDF makes the OCR import module die with UnboundLocalError instead of a MISP error.

• Problem — The PDF-stitching path in misp_modules/modules/import_mod/ocr.py logs a summary line referencing the loop variable p after the page loop, so a PDF reporting zero pages never binds p and the module dies with UnboundLocalError: local variable 'p' referenced before assignment.
• Fix — Checks for pages == 0 immediately after the count is computed and returns a misperrors error before the loop and the later reference.
• Effect — Submitting an empty or malformed PDF to the OCR import module produces a clear MISP error message instead of an unhandled traceback.

The defect

In misp_modules/modules/import_mod/ocr.py, the PDF-stitching handler loops over pages and then logs a summary line that references the loop variable p outside the loop:

pages = len(pdf.sequence)
log.debug("PDF with {} page(s) detected".format(pages))
# Create new image object where the height will be the number of pages. ...
img = WImage(width=pdf.width, height=pdf.height * pages)
# Cycle through pages and stitch it together to one big file
for p in range(pages):
    log.debug("Stitching page {}".format(p + 1))
    image = img.composite(pdf.sequence[p], top=pdf.height * p, left=0)
# Create a png blob
image = img.make_blob("png")
log.debug("Final image size is {}x{}".format(pdf.width, pdf.height * (p + 1)))

When pages == 0 (a zero-page PDF), the for loop body never executes, so p is never bound. The subsequent log.debug(...".format(pdf.width, pdf.height * (p + 1))...) line then raises UnboundLocalError: local variable 'p' referenced before assignment.

Impact

An analyst who submits a malformed or empty PDF (0 pages) to the OCR import module gets an unhandled UnboundLocalError traceback instead of a clean, actionable MISP error message. This is confusing in the UI/API response and provides no guidance on what went wrong with the input.

The fix

Added an explicit pages == 0 check immediately after pages is computed, returning a clean misperrors["error"] before the loop (and before the later p reference) is ever reached.

Verification

• flake8 misp_modules/modules/import_mod/ocr.py — clean, no output.
• Full module test suite: 161 passed, 4 skipped, 5 subtests passed in 32.94s.

Found during a review of the repository; other findings are being submitted as separate PRs.

🤖 Generated with Claude Code