syllabus-to-metadata

syllabus-to-metadata is a Python library that extracts bibliographic citation metadata from course syllabi in PDF or Word format. It uses an LLM to identify and parse citations in messy, real-world documents, outputting structured metadata as TSV for use in downstream workflows such as course reserve requests.

Quickstart

Install

pip install syllabus-metadata

Run as a command-line tool

$ syllabus-metadata <prompt_file> <syllabus_file> [--backend ollama|bedrock|openai]

Example:

$ syllabus-metadata prompts/citations.txt syllabus.pdf --backend bedrock

A sample prompt file is provided in prompts/citations.txt.

Use as a library

from syllabus_metadata import extract_citations
from syllabus_metadata.ingestion import extract_text

doc_text = extract_text("syllabus.pdf")  # also accepts .docx
prompt = open("prompts/citations.txt").read()
output = extract_citations(prompt, doc_text, backend="bedrock")
print(output)

Installing from Source

To install directly from a cloned repository:

# Ensure Python >= 3.10 is available (install via Homebrew if needed)
brew install python@3.12

# Create and activate a virtualenv
python3.12 -m venv ~/.venvs/syllabus-metadata
source ~/.venvs/syllabus-metadata/bin/activate

# Install dependencies and the package in editable mode
pip install --upgrade pip
pip install -r requirements.txt
pip install -e .

AWS Bedrock credentials

To use the bedrock backend, set the following environment variables before running the tool:

export AWS_ACCESS_KEY_ID=your_access_key_id
export AWS_SECRET_ACCESS_KEY=your_secret_access_key
export AWS_DEFAULT_REGION=us-east-1   # optional; defaults to us-east-1

Your AWS account must also have access to Amazon Bedrock and the us.anthropic.claude-sonnet-4-6 inference profile enabled in the target region.

Background and System Requirements

Python Build Requirements

• Python >= 3.10
• python-docx — Word document ingestion
• pdfplumber — PDF ingestion
• boto3 — required for the bedrock backend
• openai — required for the openai backend

System Requirements

Backend	Requirement
ollama (default)	Ollama running locally with llama3.1:latest pulled
bedrock	AWS credentials in the environment (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY) and Bedrock access enabled in your AWS account
openai	OPENAI_API_KEY set in the environment