methodology-extractor

Provide the set of papers and a target method type, and the extractor parses each methods section, identifies protocol steps, and reports which steps are common across studies versus unique to individual papers.

The bundled script scans paper text for a Methods or Materials and Methods heading, captures lines until the Results, Discussion, or References section, and then identifies numbered or bulleted protocol steps.

No external packages are required. The script uses only the Python standard library (argparse, json, collections), so it runs in a default Python environment without installing dependencies.

Extraction relies on simple heading and step-pattern matching, so papers with nonstandard section titles, unstructured prose methods, or missing headings may yield incomplete results. Verify extracted steps against the source papers before reuse.

The script accepts a JSON file mapping paper IDs to their methods text via the --papers argument, and can write results to an output JSON file. A --demo flag runs a built-in three-paper example.