istupakov/pyquasar

Finite and boundary element methods on Python

pyquasar

Finite and boundary element methods on Python

pyquasar's Issues

Add documentation for public classes & methods

As for now project lacks of any documentation and can be confusing for the users. To enhance clarity for users, we need to add comprehensive documentation to the project.

This will provide clear instructions and explanations, reducing confusion and improving user experience.

Refactoring load_mesh function into a Mesh class with support for multiple providers and global numeration

Issue Description

Currently, codebase includes a load_mesh function responsible for loading meshes, but it lacks extensibility and flexibility. To improve maintainability and add support for multiple mesh provider options (such as GMSH, NETGEN, etc.), as well as to make the default numeration global and move numeration localization into the FETI class because of FEMProblem, we need to refactor this function into a Mesh class.

Proposed Changes

• Create a Mesh class and use it inside PyQuasar classes. #2
• Separate numeration localization method into another method. #2
• Implement class methods within Mesh for each mesh provider option (e.g., from_gmsh, from_netgen, etc.).

Proposed API example

mesh = Mesh.load(
    # The path to the geometry file
    "dipole2d.geo", # Or "dipole2d.msh"
    # How many times to refine the mesh by half (0 means no refinement)
    refine_k=3,
    # How many partitions to make (0 means no partitions)
    num_part=0
)
domains = [FemDomain(domain) for domain in mesh]
problem = FemProblem(domains)

or

mesh = Mesh.load(
    # The path to the geometry file
    "dipole2d.geo", # Or "dipole2d.msh"
    # How many times to refine the mesh by half (0 means no refinement)
    refine_k=3,
    # How many partitions to make (0 means no partitions)
    num_part=0
)
mesh.localize_numaration()
domains = [FemDomain(domain) for domain in mesh]
problem = FetiProblem(domains)

Expected Benefits

• Improved code organization and maintainability by encapsulating mesh loading functionality into a dedicated class.
• Enhanced flexibility to support multiple mesh provider options without modifying existing code extensively.
• Global default numeration allows for consistent behavior (FEMProblem) across the application.
• Localization of numeration within the FETI class ensures separation of concerns and better modularization.

Additional Considerations

• Write comprehensive unit tests to validate the functionality of the Mesh class and its methods.
• Document the usage of the Mesh class and any changes in numeration handling for developers' reference.

Add type hints

Type hints enhance code readability and maintainability by specifying the types of variables, function parameters, and return values. Adding type hints to the project will improve code quality and facilitate better understanding for developers working on the project.

Restructure Attributes and Methods for Improved Encapsulation and Readability

To enhance code clarity and maintainability, we need to refactor certain attributes and methods in the classes. This involves making some attributes non-public to encapsulate their behavior and prevent direct access from outside the class. Additionally, we should convert most attributes into properties to provide clear documentation and enforce read-only access where appropriate. By implementing these changes, we can improve code readability, prevent unintended modifications, and promote best practices in our codebase.

Create a Mesh class with separated `localize_numeration` method and use it inside PyQuasar classes.

Discussed in #1.