diff --git a/DG_Approximation.py b/DG_Approximation.py index 10fdf13c604e2d1306470fbac11af31601b21285..48ea3e508b4f40b07094179339da7d3002100f04 100644 --- a/DG_Approximation.py +++ b/DG_Approximation.py @@ -16,19 +16,28 @@ 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() TODO: Contemplate keeping Theoretical threshold over all degrees (only p) -TODO: Contemplate reasoning behind Adjusted-Outer-Fence method + -> Done (same options as Boxplot) +TODO: Discuss whether Threshold detector is needed +TODO: Contemplate reasoning behind Adjusted-Outer-Fence method -> Done TODO: Discuss why indexed_coeffs are largest absolute value instead of - mean for Boxplot TC + mean for Boxplot TC -> Done (give option) TODO: Discuss why overlapping cells are only introduced for one adjacent - fold (right neighbour, except for last fold) + fold (right neighbour, except for last fold) -> Done TODO: Discuss whether quartiles do correspond to Thea's calculation + -> Done (do not, change) TODO: Discuss that quartiles do not take overlapping cells into account + -> Done (change) +TODO: Contemplate using pre-implemented quartile functions -> Done (do) TODO: Contemplate restricting number of overlapping cells to half of fold - length + length -> Done (Restrict to 1 for now) +TODO: Contemplate whether overlapping cells should be evaluated in + neighboring folds -> Done (no) TODO: Discuss if folds are valid for all possible num_grid_cells/fold_len -TODO: Contemplate combining get_cells() and _get_cells() + -> Done (enforce) +TODO: Contemplate combining get_cells() and _get_cells() -> Done (do) Urgent: +TODO: Enforce mesh with 2^n cells TODO: Change heaviside random to uniform(-100, 100) TODO: Adjust Heaviside to have non-symmetric values (left- and right_value) TODO: Rename 'adjustment' to 'shift' @@ -76,6 +85,7 @@ 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: Add ThresholdDetector TODO: Double-check everything! (also with pylint, pytype, pydoc, pycodestyle, pydocstyle) TODO: Check whether documentation style is correct diff --git a/Troubled_Cell_Detector.py b/Troubled_Cell_Detector.py index ce4cecfde141ced7ef274284d05b66abca9c291c..08df2524312cf61132af7d76d280f0517ccc7e96 100644 --- a/Troubled_Cell_Detector.py +++ b/Troubled_Cell_Detector.py @@ -6,9 +6,13 @@ TODO: Vectorize _get_cells() in Boxplot method -> Done TODO: Restructure Boxplot method -> Done TODO: Introduce lower/upper extreme outliers in Boxplot (each cell is also checked for neighboring domains if existing) -> Done +TODO: Give option to choose from multiwavelet degrees (first, last or + highest magnitude) +TODO: Change method of calculating quartiles +TODO: Include overlapping cells in quartile calculation (if needed) TODO: Determine max_value for Theoretical only over highest degree TODO: Check if indexing in wavelets is correct -TODO: Add ThresholdDetector +TODO: Combine get_cells() and _get_cells() TODO: Add TC condition to only flag cell if left-adjacent one is flagged as well (remove this condition) TODO: Check coarse_projection calculation for indexing errors