~asc/meshing/bastion : changes

~asc/meshing/bastion » Changes from revision 20

From Revision 20 to 1

Rev	Summary	Authors	Date
20	Final fix. Final fix. Licensing formally implemented. Please update your branches and any derivative works. Additional tidying and restructure.	Adam Candy	10 years ago
19	Final fix (hopefully). Final fix (hopefully). Licensing formally implemented. Please update your branches and any derivative works. Additional tidying and restructure.	Adam Candy	10 years ago
18	Fix to previous licencing and restructuring commit. Fix to previous licencing and restructuring commit. Licensing formally implemented. Please update your branches and any derivative works. Additional tidying and restructure.	Adam Candy	10 years ago
17	Licensing formally implemented. Please update your branches... Licensing formally implemented. Please update your branches and any derivative works. Additional tidying and restructure.	Adam Candy	10 years ago
16	Fixed the previous bad merge, and made changes to ensure the... Fixed the previous bad merge, and made changes to ensure the tests still pass. This splits the large single file into modules and brings in additional documentation. Note that there are some copied and renamed functions present (e.g. the gmsh_geo_define_). These are unused and need to be hunted down and removed.	Adam Candy	11 years ago
15	Updated the test 'valid' files. All tests pass. Updated the test 'valid' files. All tests pass.	Adam Candy	11 years ago
14	It is now possible to generate the Southern Ocean simulation... It is now possible to generate the Southern Ocean simulation with this version, which should (awaiting confirmation) run without the failures we've seen recently. Note the changes to the boundary IDs - the open and contour boundaries are now set as prescribed by the boundary class object - and are fixed, where they were previously variable depending on what the Gmsh command 'newll' returned.	Adam Candy	11 years ago
13	Updated the test suite engine so that it's easy to build all... Updated the test suite engine so that it's easy to build all the examples without testing them (make build), and also possible to update the valid files which contain The Truth (make updatevalid). There is now less repetition in the Makefile too, and each case is indentified by just a single line - the script call.	Adam Candy	11 years ago
12	New example - the mesh currently being used for the Southern... New example - the mesh currently being used for the Southern Ocean simulations. This contains the main land mass of Antarctica and an ocean domain extending to -30 degrees South.	Adam Candy	11 years ago
11	- Consistent use of dx_default. - Consistent use of dx_default. - The Antarctica main example contour can now be meshed (uses dx=10 in the parallel creation now). - Renamed the variables associated with the inclusion of the ice shelf ocean cavities to be more intuitive.	Adam Candy	11 years ago
10	Modified the dataset path to ease out headaches. Modified the dataset path to ease out headaches.	Adam Candy	11 years ago
9	Modified the test engine to allow for a provided dataset loc... Modified the test engine to allow for a provided dataset location. Note this necessitated a change in the testing, so that the first header line is allowed to differ. This line contains the arguments provided to the script.	Adam Candy	11 years ago
8	Added a test suite for the meshing code, containing five cas... Added a test suite for the meshing code, containing five cases. Run with 'make' in the root folder.	Adam Candy	11 years ago
7	Quick commit, to bring in changes made by the three of us wh... Quick commit, to bring in changes made by the three of us while attempting to generate a mesh for the Southern Ocean.	Adam Candy	11 years ago
6	option to smooth boundaries. There is definitely room for im... option to smooth boundaries. There is definitely room for improvements here, as it's quite slow at the moment.	Simon Mouradian	11 years ago
5	We now have the option to include or exclude the under ice s... We now have the option to include or exclude the under ice shelf region from the mesh.	Simon Mouradian	11 years ago
4	ability to choose bounding latitude. Previous implementation... ability to choose bounding latitude. Previous implementation bounded the domain at -50 degrees latitude	Simon Mouradian	11 years ago
3	Update to the code to generate a gmsh field containing a bat... Update to the code to generate a gmsh field containing a bathymetry metric imported from netCDF.	Adam Candy	12 years ago
2	Intial commit of code to generate a gmsh field containing a ... Intial commit of code to generate a gmsh field containing a bathymetry metric.	Adam Candy	12 years ago
1	Latest version of a new meshing approach for realistic domai... Latest version of a new meshing approach for realistic domains. This code generates Gmsh .geo files from RTopo NetCDF files. It is currently coded to use the mask to generate a boundary along the coastlines and grounding line* (which is not positioned at a constant depth). The code is easily extended to boundaries along depth contours and to work with other NetCDF sources. The new approach uses Python contouring routines (as opposed to the GSHHS Gmsh plugin, or the GMT approach Changze's UROP used). It would be useful to test this new code on a wide range of regions - particularly to check the contouring routines are behaving satisfactorily. It works very well in the Antarctic region (including the region inside ice shelf cavities). The dependencies are all Python modules (e.g. GMT is not required). * which involves a neat trick with modulo arithmetic to keep calculation time down significantly. Supported features: - Define boundaries from NetCDF (currently restricted to the RTopo mask - trivial to extend) - Regions defined by arbitrary functions of longitude and latitude (given on the command line) - Simpler definition of regions by boxes (such as 'longmin:longmax,latmin:longmax') - Define included paths by Gmsh ID number (useful to include or exclude specific islands/land masses) - Deals with multiple open boundaries - Islands and boundaries split by the global boundary are treated - Automatically closes boundaries with parallels and meridians (with a prescribed length step) - Applies different boundary IDs on open and closed boundaries - Exclude smaller islands - restricted by a given minimum area - Option to extend domain in latitude on open boundaries (e.g. for sponge regions, or large open regions in the open ocean) - Command line used is saved in the Gmsh .geo file for reference - Option to generate the ACC average track line (and then to refine the mesh to this) TODO: 1. Extend input options to extract boundaries from other NetCDF data (i.e. give field name and condition for contour) 2. Calculate area in the right projection 3. Length step for drawing parallels and meridians should be projected to the correct space (currently in radians (well degrees on the command line)). 4. Make it possible to generate 3 distinct regions (ocean, ice-covered ocean, land) instead of the current 2 (ocean, land). This is necessary for applying boundary conditions on the top surface, for example. 5. Look into using the Python Polygon library (http://pypi.python.org/pypi/Polygon/3.0.3) for more flexible definition of regions. This might also help in achieving point 4.	Adam Candy	12 years ago