Draft: Restructure feflowlib

Aim is a full integration to the meshclass. A goal for the API is defined in #62. Further (old) planning can be found in the wiki.

Furhter ToDos:

• rewrite cli
• think of write_mesh_of_combined_properties
• rewrite examples (delete links)
• think of new function for boundary conditions (maybe also function for writing) or introduce BC-class
• introduce meshlib.Mesh class as type for ogs_specific meshes?
• introduce consistent SI-Unit conversion for all possible parameter
• think of renaming parameter in feflow-language to ogs-language
• refactor templates, so no code duplications occur
• allow time steps, initial time and end time to be arguments for all processes in templates
• fix permeability for HT process
• refactor add material sections for processes
• check if media is written to prj-file for generic template
• introduce generic template for processes that are not defined
• write test for generic template with a FEFLOW model of a process that is not fully supported
• allow SSD as process in feflowModel
• rename feflowModel->FeflowModel
• bug fixes Hailong
• remove -i -o from cli and use same path as input, if no output is specified
• edit log: info level only for cli, else warn-level
• allow for heterogenous material properties apart from permeability...(CT-missing, also combination of permeabilities to KF in the bulkmesh)
• ... and write corresponding tests for it
• adjust test for large model
• if MaterialIDs in user-data overwrite them in bulk mesh
• rework the user-guide, especially the mermaid figure
• add option to directly run ogs from cli after conversion?
• add fixed_out_times and other possible input arguments to the FeflowModel-prj/run_OGS-methods
• add error tolerance as argument
• also use initial and final time step from feflow file
• maybe use linear time increment as default
• use initial values from FEFLOW model
• tidy private functions and __init__.py