code-analyzer

To extract computational methodology, the project requires a cloned GitHub repository with an accessible codebase and a scope.md file. The analysis identifies data flow, preprocessing steps, and modeling decisions to draft a scientific Methods section.

You can generate a Methods section by running an analyzer on a project folder containing scope.md and a code/ repository. It outputs notes/code-analysis.md and drafts/methods.md after validating the extracted methodology against the project scope.

You need a scope.md document and a code/ repository context containing notebooks, scripts, and data pipelines. These files provide the project scope and accessible codebase required to extract the methodology.

Yes, you can analyze notebooks, scripts, and configs to extract data preprocessing steps. The process identifies data sources, flow, and analysis steps from the repository's data pipelines to output structured documentation.

The best way to document data sources and analysis steps is extracting the methodology from notebooks and configs, then producing structured notes. These notes feed directly into a Methods draft aligned with the project scope.