sayou-document
Overview
The Universal Document Parsing Gateway for Sayou Fabric.
sayou-document is a high-fidelity parsing engine that converts diverse document formats (PDF, DOCX, PPTX, XLSX, Images) into a unified, structured Document Object Model (DOM).
Unlike simple text extractors, it preserves the semantic structure of documents—headers, tables, charts, and layout coordinates—making it ideal for RAG applications that require layout awareness.
1. Architecture & Role
The Document engine acts as a normalizer. It accepts raw file bytes and applies the optimal Parser Strategy to output a structured SayouDocument.
graph LR
File[Raw File] --> Pipeline[Document Pipeline]
subgraph Parsers
PDF[PDF Parser + OCR]
Office[Office Parser]
Img[Image Converter]
end
Pipeline -->|Type Detection| Parsers
Parsers --> DOM[Structured DOM]
1.1. Core Features
- Smart Routing: Automatically detects file types (signatures) and selects the best parser.
- Hybrid Extraction: Combines native text extraction for digital PDFs with OCR fallback for scanned images.
- Strict Schema: Outputs a standardized hierarchy (
Document>Page>Element) regardless of input format.
2. Supported Formats
sayou-document supports the following file types out-of-the-box.
| Format | Strategy Key | Description |
|---|---|---|
pdf |
Extracts text, images, and TOC using PyMuPDF. Supports OCR. |
|
| Word | docx |
Parses DOCX files, preserving heading levels and lists. |
| PowerPoint | pptx |
Extracts text frames, speaker notes, and tables from slides. |
| Excel | xlsx |
Converts sheets into table elements and extracts embedded charts. |
| Image | image |
Auto-converts JPG/PNG/TIFF to PDF, then applies OCR. |
3. Installation
pip install sayou-document
# For OCR support (requires Tesseract installed on OS)
pip install "sayou-document[ocr]"
4. Usage
The DocumentPipeline orchestrates file detection and parsing. It standardizes the input via the process method.
Case A: PDF Parsing (Standard)
Processes a PDF file to extract structured text and layout info.
import os
from sayou.document import DocumentPipeline
file_path = "quarterly_report.pdf"
with open(file_path, "rb") as f:
file_bytes = f.read()
doc = DocumentPipeline.process(
data=file_bytes,
metadata={"filename": os.path.basename(file_path)}
)
# 4. Result
print(f"File: {doc.file_name}, Pages: {doc.page_count}")
print(f"First Element: {doc.pages[0].elements[0].text}")
Case B: Office Documents (Word/Excel)
Parses Office formats while preserving table structures.
from sayou.document import DocumentPipeline
with open("salary_table.xlsx", "rb") as f:
file_bytes = f.read()
doc = DocumentPipeline.process(
data=file_bytes,
metadata={"filename": "salary_table.xlsx"}
)
# Access tables
tables = [e for p in doc.pages for e in p.elements if e.category == "table"]
print(f"Extracted {len(tables)} tables.")
Case C: Image with OCR
Automatically handles image conversion and OCR processing.
from sayou.document import DocumentPipeline
# Initialize with OCR enabled
pipeline = DocumentPipeline(config={"use_ocr": True, "ocr_lang": "eng"})
with open("scanned_receipt.png", "rb") as f:
file_bytes = f.read()
doc = pipeline.process(
data=file_bytes,
metadata={"filename": "scanned_receipt.png"}
)
print(f"OCR Result: {doc.pages[0].elements[0].text}")
5. Configuration Keys
Customize the parsing behavior via the config dictionary.
use_ocr: (bool) Enable OCR for scanned pages or images.ocr_lang: (str) Tesseract language code (default:eng+kor).extract_images: (bool) Whether to extract embedded images to disk.table_strategy: (str)fast(text-based) oraccurate(vision-based).
6. License
Apache 2.0 License © 2026 Sayouzone
7. Plugin List
| Plugin | Example | Description |
|---|---|---|
Docx Parser |
▶ | |
Excel Parser |
▶ | |
PPTX Parser |
▶ | |
PDF Parser |
▶ |
1 maintainer
Metadata
1 maintainer
Release files for sayou-document 0.5.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| sayou_document-0.5.2.tar.gz | 60.5 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| sayou_document-0.5.2-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 106.9 kB
Release files / sayou_document-0.5.2.tar.gz
| Download URL | sayou_document-0.5.2.tar.gz |
|---|---|
| Size | 60.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
6dab15d8b19843a7349f29c392918f706c0d98dfa99952e5537e79c3a7e44b6c
|
|
BLAKE2b-256 checksum How to use checksums |
d3f9486b6c261e15e0ef0e0d12671907d5842fb0016d665c9db829b1416c672d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Release files / sayou_document-0.5.2-py3-none-any.whl
| Download URL | sayou_document-0.5.2-py3-none-any.whl |
|---|---|
| Size | 46.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
607e05c99780f0699449a9f0f61667959d61ae2c94c77cf11948802e99fea37a
|
|
BLAKE2b-256 checksum How to use checksums |
f47324b6a2ecad95f8fa83599bb3bf53966c38476a9220ad6a1b92aa8e247cd8
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Security for sayou-document 0.5.2
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
