How novonordisk-research/pepfunn works
Main Components
• Core Library (Domain Module) – The heart of the project lies in the “src/pepfunn” directory. Here you have multiple modules:
▪ library.py – for designing or generating peptide libraries.
▪ sequence.py – for peptide sequence manipulation.
▪ pairs.py – for handling matched pairs analysis.
▪ similarity.py – for calculating sequence or molecular similarity.
▪ clustering.py – for clustering algorithms applied to peptide molecules.
▪ data subfolder – containing reference/prerequisite data files (e.g., monomers.sdf, property.txt, etc.) used by the analysis modules.
• Interactive Notebooks – The “notebooks” folder contains Jupyter scripts that demo each main functionality. They can be illustrated as a user interface layer where researchers experiment with analysis workflows and generate outputs (figures, logs, etc.) which are then compared by the end-user.
• Unit Testing – The “tests” folder (and the standalone test.py) represents a testing layer verifying each module’s functionality. This shows a commitment to quality and a solid test orchestration point.
• External Dependencies – The README and third_party_licenses indicate important third-party tools such as RDKit and BioPython. Include a block showing “External Libraries/Third-Party Tools” with directional connections to the Core Library (indicating dependency).
Representing Relationships and Interactions
• Flow from Data to Processing Modules: Show that the data files (input data such as sequences, properties, and molecular information) feed into the core classes of the “pepfunn” package.
• User/Researcher Interaction: Connect the Notebooks (as a demonstration layer) to the core modules, indicating that they call functions from library, sequence, pairs, similarity, and clustering.
• Testing as an Isolated Check: Represent the tests suite as an independent block that interacts with each main module’s unit functionalities.
Highlighting Architectural Patterns and Design Principles
• Modular Design: Emphasize the separation of concerns by showing that each functionality (e.g., clustering, library generation) has its own module.
• Testing and Validation: The tests block reinforces that modularity supports direct unit testing, a best practice in modern Python packages.