Overview
The most common workflow: define a document as JSON, generate clean PDF/PNG, then produce realistic photo variations.
Step 1: Define Your Document
Create a JSON file with header, items, and optional observations:
{
"header": {
"doc_type": "guia_despacho",
"doc_number": "00098765",
"date": "20/04/2026",
"emitter_name": "ACME LOGISTICS S.A.",
"emitter_rut": "76.543.210-K",
"emitter_address": "Av. Industrial 500, Santiago",
"receiver_name": "HOTEL PACIFIC S.A.",
"receiver_rut": "96.123.456-7",
"receiver_address": "Costanera Norte 2100, Vina del Mar",
"oc_number": "4500002345",
"vehicle_plate": "HJKL-42",
"driver_name": "Carlos Mendez"
},
"items": [
{
"pos": 1,
"code": "AL-3001",
"description": "SALMON FRESCO FILETE (kg)",
"qty": 25,
"unit": "KG",
"unit_price": 14500,
"total": 362500
},
{
"pos": 2,
"code": "AL-3050",
"description": "CAMARON ECUATORIANO 16/20 (kg)",
"qty": 10,
"unit": "KG",
"unit_price": 18900,
"total": 189000
}
],
"observations": "Salmon: 25kg en 5 bandejas de 5kg cada una."
}
Step 2: Generate the Document
penquify pdf --doc-json doc.json --output output/
import asyncio
import json
from penquify.models import Document, DocHeader, DocItem
from penquify.generators.pdf import generate_document_files
with open("doc.json") as f:
data = json.load(f)
doc = Document(
header=DocHeader(**data["header"]),
items=[DocItem(**it) for it in data["items"]],
observations=data.get("observations", ""),
)
files = asyncio.run(generate_document_files(doc, "output/"))
print(files)
# {"html": "output/guia_despacho_00098765.html",
# "png": "output/guia_despacho_00098765.png",
# "pdf": "output/guia_despacho_00098765.pdf"}
curl -X POST http://localhost:8000/generate/document \
-H "Content-Type: application/json" \
-d @doc.json
Step 3: Generate Photo Variations
penquify photos --image output/guia_despacho_00098765.png \
--presets full_picture folded_skewed blurry coffee_stain
from penquify.generators.photo import generate_dataset
results = asyncio.run(generate_dataset(
"output/guia_despacho_00098765.png",
output_dir="output/photos",
preset_names=["full_picture", "folded_skewed", "blurry", "coffee_stain"],
doc_description="guia 00098765, OC 4500002345, 2 items seafood",
))
for r in results:
print(f"{r['name']}: {'OK' if r['ok'] else 'FAIL'} -> {r['path']}")
The doc_description parameter helps the image generator preserve key fields. Include the document number, reference numbers, and any critical values.
Full Pipeline (Document + Photos in One Call)
penquify demo --output output/ --presets full_picture blurry
curl -X POST http://localhost:8000/generate/dataset \
-H "Content-Type: application/json" \
-d @doc.json