diff --git a/DG_Approximation.py b/DG_Approximation.py index 26e3fa98cd3ac9d8720870c96a3c90b69eaa6b82..f5008ddbdc7e397a57f78b625cd9d288ae7e8012 100644 --- a/DG_Approximation.py +++ b/DG_Approximation.py @@ -14,6 +14,7 @@ TODO: Contemplate containing boundary condition in Mesh TODO: Ask whether all quadratures depend on freely chosen num_nodes TODO: Contemplate saving training data for each IC separately TODO: Contemplate removing TrainingDataGenerator class +TODO: Ask why no fine mesh is used for calculate_coarse_projection() Urgent: TODO: Investigate self-referencing in classes -> Done @@ -22,45 +23,44 @@ TODO: Refactor random_stencil() -> Done TODO: Fix bug applying wrong boundary condition when generating ANN training data -> Done TODO: Find errors in centering for ANN training -> Done -TODO: Remove stencil_length as instance variable +TODO: Remove stencil_length as instance variable -> Done +TODO: Note down TCD adjustment TODOs -> Done TODO: Adapt TCD from Soraya (Dropbox->...->TEST_troubled-cell-detector->Troubled_Cell_Detector) -TODO: Add TC condition to only flag cell if left-adjacent one is flagged as - well TODO: Move plot_approximation_results() into plotting script -TODO: Add verbose output TODO: Improve file naming (e.g. use '.' instead of '__') TODO: Check whether ghost cells are handled/set correctly TODO: Unify use of 'length' and 'len' in naming TODO: Unify use of 'initial_condition' and 'init_cond' in naming -TODO: Ensure uniform use of mesh and grid +TODO: Unify use of 'mesh' and 'grid' in naming +TODO: Check sign change in stiffness matrix to accommodate negative wave speed +TODO: Check whether 'projection' is always a ndarray +TODO: Force input_size for each ANN model to be stencil length Critical, but not urgent: -TODO: Combine ANN workflows -TODO: Build package (module?) for DG scheme -TODO: Add an environment file for Snakemake +TODO: Add verbose output +TODO: Outsource scripts into separate directory +TODO: Check whether all instance variables are sensible +TODO: Combine ANN workflows if feasible TODO: Investigate profiling for speed up TODO: Rework Theoretical TC for efficiency -TODO: Check sign change in stiffness matrix to accommodate negative wave speed -TODO: Investigate g-mesh(?) -TODO: Create g-mesh with Mesh class -TODO: Force input_size for each ANN model to be stencil length -TODO: Use full path for ANN model state TODO: Extract object initialization from DGScheme -TODO: Use cfl_number for updating, not just time - -Currently not critical: TODO: Replace loops with list comprehension if feasible TODO: Replace loops/list comprehension with vectorization if feasible -TODO: Check whether 'projection' is always a ndarray -TODO: Check whether all instance variables are sensible + +Currently not critical: +TODO: Add an environment file for Snakemake +TODO: Build package (module?) for DG scheme +TODO: Investigate g-mesh(?) +TODO: Create g-mesh with Mesh class TODO: Rename files according to standard -TODO: Outsource scripts into separate directory TODO: Allow comparison between ANN training datasets +TODO: Use full path for ANN model state TODO: Add a default model state TODO: Look into validators for variable checks Not feasible yet or doc-related: +TODO: Use cfl_number for updating, not just time (equation-related?) TODO: Adjust code to allow classes for all equations (Burger, linear advection, 1D Euler) TODO: Double-check everything! (also with pylint, pytype, pydoc, diff --git a/Troubled_Cell_Detector.py b/Troubled_Cell_Detector.py index 5adf0474380730ee7e558d8251ccbfffd648068f..1871ad7d47aea064bbfcdf99eb717c318d53d98a 100644 --- a/Troubled_Cell_Detector.py +++ b/Troubled_Cell_Detector.py @@ -2,6 +2,12 @@ """ @author: Laura C. Kühle, Soraya Terrab (sorayaterrab) +TODO: Add TC condition to only flag cell if left-adjacent one is flagged as + well +TODO: Determine max_value for Theoretical only over 0th degree (?) +TODO: Introduce Adjusted Outer Fence method in Boxplot using global_mean +TODO: Introduce overlapping cell for adjacent folds in Boxplot +TODO: Introduce lower extreme outliers in Boxplot TODO: Adjust TCs for wavelet detectors (sliding window over all cells instead of every second) TODO: Adjust Boxplot approach (adjacent cells, outer fence, etc.)