rag-pdf-highlighter 0.1.0

Highlight text chunks in PDFs – a FastAPI microservice for RAG pipelines

RAG PDF Highlighter

Highlight text chunks inside PDF documents — built for RAG pipelines.

Given a PDF URL and a list of text chunks (with page numbers), this service locates each chunk in the PDF and returns an annotated copy with highlights applied.

Features

• 3-tier text matching — exact → sentence → collapsed-whitespace fallback
• Async PDF download — non-blocking I/O via httpx
• Stateless — temp files are cleaned up after every request
• Docker-ready — single-command container deployment
• Library-friendly — core logic raises plain Python exceptions, no FastAPI dependency required

Installation

pip install callisto-pdf-highlighter

Or install from source:

git clone https://github.com/RipeSeed/calli-pdf-highlighter-.git
cd calli-pdf-highlighter-
pip install -e ".[dev]"

Quick Start

As a microservice

uvicorn rag_pdf_highlighter.main:app --host 0.0.0.0 --port 8000

Then send a POST request:

curl -X POST http://localhost:8000/highlight \
  -H "Content-Type: application/json" \
  -d '{
    "pdf_url": "https://example.com/report.pdf",
    "documents": [
      {
        "page_content": "Text to highlight in the PDF",
        "metadata": {"page": 0}
      }
    ]
  }' \
  --output highlighted.pdf

With Docker

docker build -t rag-pdf-highlighter .
docker run -p 8000:8000 rag-pdf-highlighter

As a library

from langchain_core.documents import Document
from rag_pdf_highlighter.utils.pdf_helpers import highlight_chunks_in_pdf

documents = [
    Document(page_content="Text to find", metadata={"page": 0}),
]

output_path = highlight_chunks_in_pdf(
    pdf_path="./report.pdf",
    documents=documents,
)
print(f"Highlighted PDF saved to: {output_path}")

API Reference

GET /

Health check endpoint.

POST /highlight

Field	Type	Description
pdf_url	string	URL of the PDF to highlight
documents	array	List of {page_content, metadata} objects
documents[].page_content	string	The text chunk to locate and highlight
documents[].metadata.page	int	Zero-indexed page number

Returns: The highlighted PDF file (application/pdf).

Development

# Install with dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Lint
ruff check src/ test/

License

MIT