Skip to content

Latest commit

 

History

History
85 lines (69 loc) · 4.11 KB

testing-guide.md

File metadata and controls

85 lines (69 loc) · 4.11 KB

Testing Guide

Requirements

  • Python 3.10+
  • uv on PATH
  • NUTRIENT_API_KEY set
  • Multi-page PDF input (3+ pages recommended for comprehensive testing)
export NUTRIENT_API_KEY="nutr_sk_..."
# NOTE: Use a multi-page PDF (3+ pages required for split.py and delete-pages.py)
PDF=/path/to/your/input.pdf
OUT=/tmp/ndp-test
mkdir -p "$OUT"

Run commands from repository root.

1. Smoke tests

All scripts should print help and exit 0:

uv run nutrient-document-processing/scripts/convert.py --help
uv run nutrient-document-processing/scripts/merge.py --help
uv run nutrient-document-processing/scripts/split.py --help
uv run nutrient-document-processing/scripts/ocr.py --help
uv run nutrient-document-processing/scripts/extract-text.py --help
uv run nutrient-document-processing/scripts/extract-table.py --help
uv run nutrient-document-processing/scripts/extract-key-value-pairs.py --help
uv run nutrient-document-processing/scripts/watermark-text.py --help
uv run nutrient-document-processing/scripts/redact-ai.py --help
uv run nutrient-document-processing/scripts/rotate.py --help
uv run nutrient-document-processing/scripts/sign.py --help
uv run nutrient-document-processing/scripts/optimize.py --help
uv run nutrient-document-processing/scripts/password-protect.py --help
uv run nutrient-document-processing/scripts/add-pages.py --help
uv run nutrient-document-processing/scripts/delete-pages.py --help
uv run nutrient-document-processing/scripts/duplicate-pages.py --help

2. Single-operation happy-path checks

uv run nutrient-document-processing/scripts/convert.py --input "$PDF" --format docx --out "$OUT/convert.docx"
uv run nutrient-document-processing/scripts/merge.py --inputs "$PDF,$PDF" --out "$OUT/merge.pdf"
uv run nutrient-document-processing/scripts/split.py --input "$PDF" --ranges "0:1" --out-dir "$OUT/split" --prefix part
uv run nutrient-document-processing/scripts/ocr.py --input "$PDF" --languages english --out "$OUT/ocr.pdf"
uv run nutrient-document-processing/scripts/extract-text.py --input "$PDF" --out "$OUT/text.json"
uv run nutrient-document-processing/scripts/extract-table.py --input "$PDF" --out "$OUT/tables.json"
uv run nutrient-document-processing/scripts/extract-key-value-pairs.py --input "$PDF" --out "$OUT/kvp.json"
uv run nutrient-document-processing/scripts/watermark-text.py --input "$PDF" --text CONFIDENTIAL --out "$OUT/watermark.pdf"
uv run nutrient-document-processing/scripts/redact-ai.py --input "$PDF" --criteria "Remove all SSNs" --mode stage --out "$OUT/redact-stage.pdf"
uv run nutrient-document-processing/scripts/rotate.py --input "$PDF" --angle 90 --out "$OUT/rotate.pdf"
uv run nutrient-document-processing/scripts/sign.py --input "$PDF" --out "$OUT/sign.pdf"
uv run nutrient-document-processing/scripts/optimize.py --input "$PDF" --out "$OUT/optimize.pdf"
uv run nutrient-document-processing/scripts/password-protect.py --input "$PDF" --user-password upass --owner-password opass --out "$OUT/protected.pdf"
uv run nutrient-document-processing/scripts/add-pages.py --input "$PDF" --count 1 --out "$OUT/add-pages.pdf"
uv run nutrient-document-processing/scripts/delete-pages.py --input "$PDF" --pages 0 --out "$OUT/delete-pages.pdf"
uv run nutrient-document-processing/scripts/duplicate-pages.py --input "$PDF" --pages 1,0,1 --out "$OUT/duplicate-pages.pdf"

3. Multi-step workflow behavior

No multi-step script should exist in scripts/.

Build runtime pipelines in a temporary file using the template:

TMP_SCRIPT=/tmp/ndp-runtime-pipeline.py
cp nutrient-document-processing/assets/templates/custom-workflow-template.py "$TMP_SCRIPT"
# edit $TMP_SCRIPT for the requested pipeline
uv run "$TMP_SCRIPT" --input "$PDF" --out "$OUT/pipeline.pdf"
rm -f "$TMP_SCRIPT"

4. Pass criteria

  • All --help commands succeed.
  • Single-operation scripts produce output files (using a multi-page PDF).
  • Multi-step workflows are run from temporary scripts only.
  • No committed pipeline scripts exist in nutrient-document-processing/scripts/.

Note: Single-page PDFs will cause failures in split.py and delete-pages.py. Use a multi-page PDF (3+ pages) for complete test coverage.