First release

Centers detection:

• Centers now detected based on an affinity matrix, computed as product between the local similarity matrix, and the local flows per node
• Centers of a certain degree can be found now, based on indirect flows of that same degree. The indirect flows represent the implicit flows from one node to another, based on the direct flows. The indirect flows of degree n are defined as the indirect flows of the one of degree n-1.
• Faster and clearer raw assignment stabilization. The method now only takes care of creating one stabilized assignment, with no regards to which node is the center of that assignment. Graph terminology has been used more rigorously, for high code clarity.
• Loading bars for long processes in the algorithm

Refactoring:

• Moved non-I/O utilities out of FlowsHandlers, either to FlowsFeatureExtractor or removed from the code
• Added tests for FlowsFeatureExtractor and CentersDetector
• Removed FlowsAssignmentOptimizer, which wasn't used anymore

Maintainability

• Removed INSEE-specific acronyms from the codebase
• Added a CI-pipeline for tests and automatic documentation

Accessibility

• Added a README with proper introduction to the codebase and description on how to use it
• Added ALGORITHM.md to detail how the algorithm works, separately from the README
• CITATION.cff and license added added
• Examples on how to use the codebase with either local or custom data
• Added illustrations of the algorithm's results for the documentation