Rr2Annotate

A .NET command-line tool that extracts annotations from PDFs reviewed in RailReader2 and produces structured Markdown documents. Designed for feeding annotated documents to AI for summarisation or explanation.

Features

• Extracts all annotation types: highlights, text notes, rectangles, and freehand drawings
• Groups annotations under document headings from the PDF outline
• Arranges content in reading order
• Summary table at the top with annotation counts per section
• Highlights appear bold within their surrounding text context (fuzzy whitespace matching)
• Text notes are rendered as blockquotes with the nearby document text
• Deduplicates block text when multiple annotations overlap the same paragraph
• Cleans PDF text extraction artifacts (soft hyphens, control characters)
• Optional cropped screenshots for rectangle and freehand annotations
• Page range filtering to export only specific pages
• Colour filtering to export only annotations of specific colours

Prerequisites

• .NET 9+
• RailReader2 CLI installed and accessible

Setup

Build the project:

dotnet build

Configure the path to your RailReader2 CLI on first run (or any time with --configure):

dotnet run --project Rr2Annotate/ -- --configure

This stores your CLI command in ~/.config/rr2annotate/settings.json. You can point it to a wrapper script, a direct binary path, or any command that invokes the RailReader2 CLI.

Usage

rr2annotate <pdf> [options]

Options:
  -o <path>       Output markdown file (default: <pdf-stem>-annotations.md)
  --pages <range> Only include annotations from these pages (e.g. "1,3,5-10")
  --color <hex>   Filter by annotation colour (e.g. "#FF0000" or "ff0000,ffcc00")
  --images        Include cropped screenshots for rect/freehand annotations
  --configure     Set or update the path to the RailReader2 CLI
  -h, --help      Show this help

Text-only export

dotnet run --project Rr2Annotate/ -- document.pdf -o notes.md

With images

dotnet run --project Rr2Annotate/ -- document.pdf -o notes.md --images

This creates notes.md and a notes-images/ directory with cropped screenshots of rectangle and freehand annotations.

Specific pages only

dotnet run --project Rr2Annotate/ -- document.pdf -o notes.md --pages "1,3,5-10"

Filter by colour

# Only red annotations
dotnet run --project Rr2Annotate/ -- document.pdf --color "#FF0000"

# Multiple colours
dotnet run --project Rr2Annotate/ -- document.pdf --color "ff0000,ffcc00"

Run tests

dotnet test

Install as a global tool

dotnet pack Rr2Annotate/
dotnet tool install --global --add-source Rr2Annotate/nupkg Rr2Annotate

Then use directly:

rr2annotate document.pdf -o notes.md --images

License

MIT