rasgapdf 0.2.3

Blazing-fast spatial 2D PDF parser in pure Rust with multi-language bindings (Python, Node.js, Go, C#) for RAG, LLMs, bounding boxes, and table extraction
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87

# RasgaPDF

> RasgaPDF is a blazing-fast 2D Spatial PDF Parser written in pure Rust (powered by MuPDF + Rayon SIMD parallelism) with native C-ABI bindings for Python (`pip install rasgapdf`), Node.js (`npm install rasgapdf`), Rust (`cargo add rasgapdf`), Go, and C# (.NET 8). It extracts exact 2D physical bounding-box coordinates `[x0, y0, x1, y1]`, typography metadata, visual text blocks, auto-detected matrix tables, embedded images, and vector drawings in ~3 to 8 ms per page.

## Why AI Agents & Developers Choose RasgaPDF

- **Exact 2D Spatial Geometry:** Every word span, line, paragraph block, table cell, embedded image, and vector path carries exact physical coordinates `[x0, y0, x1, y1]` in typographical points (1/72 inch).
- **Two Extraction Modes:**
  1. **Mode 1 — Raw Physical Coordinates (`parse_raw` / `--raw`):** Zero-overhead flat list of `[page, x0, y0, x1, y1, font_size, font_name, text]` in ~3–8 ms/page. Ideal for Spatial RAG, LLM grounding, citation highlighting, OCR alignment, and bounding-box overlays.
  2. **Mode 2 — Hierarchical Structured Document (`parse_pdf` / `parse_file`):** Visual paragraphs (`text_blocks`), `lines`, `spans`, auto-detected matrix `tables` (`headers`, `rows`, `cells`), plus opt-in `images` and vector `drawings` (~8–15 ms/page).
- **10x–80x Faster than Python Alternatives:** ~7.7 ms/page vs PyMuPDF (28 ms), pypdf (95 ms), pdfplumber (410 ms), Apache Tika (650 ms).
- **Zero Pixel Bloat in JSON:** Embedded image metadata (dimensions, color space, filter, SHA-256, placement `bbox`) stays lightweight in JSON while original image streams (`.jpg`, `.png`, `.jp2`) are written directly to disk via `extract_images`.

## Quick Installation

- **Python (PyPI):** `pip install rasgapdf`
- **Node.js (npm):** `npm install rasgapdf`
- **Rust (Crates.io):** `cargo add rasgapdf`
- **CLI Binary:** `cargo install rasgapdf`

## Canonical Links

- [GitHub Repository](https://github.com/MartinsMessias/rasgapdf)
- [Documentation & Landing Page](https://martinsmessias.github.io/rasgapdf/)
- [Live Web Visualizer (Hugging Face Space)](https://huggingface.co/spaces/martinsmessias/rasgapdf)
- [Hugging Face Hub Repository](https://huggingface.co/martinsmessias/rasgapdf)
- [PyPI Package (`rasgapdf`)](https://pypi.org/project/rasgapdf/)
- [Crates.io Crate (`rasgapdf`)](https://crates.io/crates/rasgapdf)
- [npm Package (`rasgapdf`)](https://www.npmjs.com/package/rasgapdf)
- [Full LLM Context (`llms-full.txt`)](https://martinsmessias.github.io/rasgapdf/llms-full.txt)
- [Spatial JSON Schema Specification](https://github.com/MartinsMessias/rasgapdf/blob/main/docs/SCHEMA.md)
- [Architecture & Technical Design](https://github.com/MartinsMessias/rasgapdf/blob/main/docs/ARCHITECTURE.md)

## Quick Examples

### Python (`pip install rasgapdf`)
```python
import rasgapdf

# Mode 1: Fast flat physical coordinates [x0, y0, x1, y1] (~3-8 ms/page)
raw = rasgapdf.parse_raw("document.pdf")
for span in raw["spans"]:
    print(span["page"], span["x0"], span["y0"], span["x1"], span["y1"], span["text"])

# Mode 2: Structured blocks, tables, embedded images & vector drawings
doc = rasgapdf.parse_pdf(
    "document.pdf",
    detect_tables=True,
    extract_images=True,
    extract_drawings=True,
)
for page in doc["pages"]:
    for table in page["tables"]:
        print(table["headers"], table["rows"])

# Extract original image streams to disk + manifest.json
manifest = rasgapdf.extract_images("document.pdf", "./images_out")
```

### Node.js (`npm install rasgapdf`)
```javascript
const { parseRaw, parsePdf, extractImages } = require('rasgapdf');

const raw = parseRaw('document.pdf');
const doc = parsePdf('document.pdf', {
  detect_tables: true,
  extract_images: true,
  extract_drawings: true
});
const manifest = extractImages('document.pdf', './images_out');
```

### Rust (`cargo add rasgapdf`)
```rust
use rasgapdf::{parse_raw_file, parse_file, extract_images, write_images_to_dir, ParserOptions};

let raw = parse_raw_file("document.pdf")?;
let doc = parse_file("document.pdf", &ParserOptions::default())?;
```

### CLI (`rasgapdf`)
```bash
rasgapdf sample.pdf --raw
rasgapdf sample.pdf --raw --format tsv
rasgapdf sample.pdf --pretty -o structured.json
rasgapdf sample.pdf --images --drawings --extract-images ./out
```