docslight 0.1.2

Lightweight ComPDF document parsing and extraction SDK

DocSlight

Lightweight Python SDK & CLI for document parsing and structured extraction

What is DocSlight?

A lightweight Python library that turns PDFs, images, and Office documents into clean Markdown or structured JSON — with one line of code. Works with ComPDF Cloud (recommended) or fully offline with local parsers.

from docslight import DocSlight

client = DocSlight(api_key="your-api-key")
result = client.parse("invoice.pdf")
print(result.to_markdown())

Quick Start

pip install docslight

Parse any document:

docslight parse invoice.pdf --output invoice.md
docslight parse invoice.pdf --format zip --output invoice.zip

Extract specific fields:

docslight extract invoice.pdf --fields invoice_number,total_amount

Launch the local Web UI workbench:

pip install "docslight[web]"
docslight web
# Open http://127.0.0.1:8000

# Or run the same Web UI directly as a module
python -m docslight.web_app --host 0.0.0.0 --port 8000 --debug

Features

• Dual mode — ComPDF Cloud for production-grade results, or local CPU parsing for offline evaluation
• Parse → Markdown — Convert PDF, DOCX, PPTX, XLSX, and images (PNG, JPG, TIFF, BMP, WebP) to clean Markdown
• Extract → JSON — Pull structured data by field list, JSON Schema, or structured template (key-value + table extraction)
• CLI first — Full-featured command-line interface, script-friendly
• Web UI — Local Flask workbench with drag-and-drop, live preview with bbox highlights, and a Fields Builder UI
• Batch processing — parse_batch() / extract_batch() for multiple files
• Local LLM extraction — Ollama or any OpenAI-compatible provider for offline extraction
• Document types — Classify and route documents by type for cloud extraction
• Error-safe — Typed result objects, structured error hierarchy, no credential leaks

Install

Scenario	Command
Core SDK & CLI	pip install docslight
+ Local parsing (OCR, Office)	pip install "docslight[local]"
+ Local LLM extraction	pip install "docslight[local,local-llm]"
+ Web UI workbench	pip install "docslight[web]"

Local CPU parsing is experimental. Validate accuracy and latency on your own documents before production use.

SDK Usage

Cloud — Parse

from docslight import DocSlight

client = DocSlight(mode="cloud", api_key="your-api-key")

result = client.parse("invoice.pdf")
print(result.to_markdown())   # Clean markdown
print(result.to_json())       # Full result with pages + metadata

Cloud — Extract

result = client.extract(
    "invoice.pdf",
    fields=["invoice_number", "invoice_date", "total_amount"],
)
print(result.to_json())

With a JSON Schema:

schema = {
    "type": "object",
    "properties": {
        "invoice_number": {"type": "string"},
        "total_amount": {"type": "number"},
    },
    "required": ["invoice_number"],
}
result = client.extract("invoice.pdf", schema=schema)

With document type classification:

result = client.extract(
    "invoice.pdf",
    fields=["invoice_number"],
    document_types=["invoice"],
)

Local — Parse (Offline)

client = DocSlight(mode="local")
result = client.parse("invoice.pdf")
print(result.to_markdown())

Local — Extract with Ollama

client = DocSlight(
    mode="local",
    local_llm={"provider": "ollama", "model": "llama3.1"},
)
result = client.extract(
    "invoice.pdf",
    fields=["invoice_number", "invoice_date"],
)

Local — Extract with OpenAI-Compatible API

client = DocSlight(
    mode="local",
    local_llm={
        "provider": "openai-compatible",
        "base_url": "https://your-endpoint/v1",
        "model": "your-model",
        "api_key": "your-api-key",
        "extra_body": {"enable_thinking": False},  # e.g., DashScope qwen3
    },
)

Batch Processing

results = client.parse_batch(["doc1.pdf", "doc2.pdf", "doc3.pdf"])
for r in results:[release.ps1](scripts/release.ps1)
    print(r.to_markdown()[:200])

CLI Usage

# Parse
docslight parse invoice.pdf --mode cloud -o invoice.zip
docslight parse invoice.pdf --mode cloud --format zip -o invoice.zip
docslight parse invoice.pdf --mode local -o invoice.zip

# Extract
docslight extract invoice.pdf --mode cloud --fields invoice_number,total_amount
docslight extract invoice.pdf --mode local --fields invoice_number --local-llm-provider ollama --local-llm-model llama3.1
docslight extract "D:\pdf\invoice\1.pdf" --mode local --fields invoice_number --local-llm-provider ollama

# Extract with schema
docslight extract invoice.pdf --schema schema.json

# Web UI
docslight web --host 127.0.0.1 --port 8000

Web UI Workbench

DocSlight Workbench is a local Flask app for visual document processing.

pip install "docslight[web]"
docslight web
python -m docslight.web_app

• Parse & Extract tabs — Switch between parsing and extraction workflows
• Drag-and-drop upload — PDF, images, DOCX, PPTX, XLSX
• Live preview — PDF page rendering with bbox highlight overlays
• Fields Builder — Structured UI for building key-value and table extraction templates
• Download results — One-click download of Markdown or JSON output

Environment Variables

Variable	Description
DOCSLIGHT_API_KEY	API key for cloud mode
DOCSLIGHT_MODE	Processing mode: cloud or local (default: cloud)

Supported Inputs

Mode	Formats
Cloud	PDF, images (PNG/JPG/TIFF/BMP/WebP), DOCX, PPTX, XLSX, and more via ComPDF Cloud API
Local	PDF, images (PNG/JPG/TIFF/BMP/WebP), DOCX, PPTX, XLSX

Legacy Office formats (.doc, .ppt, .xls) must be converted to DOCX/PPTX/XLSX for local processing.

Development

pip install -e ".[dev]"
ruff check .
mypy docslight
pytest
python -m build

License

MIT License. See LICENSE.