Mango core data structures and functions of the mango 3D image processing package.
| Dds | Mango distributed array (DistributedDataStruct) python API. |
| mtype(name) | Image type for mango Dds distributed array objects. |
| DdsNonHaloSubDomain | Non-halo-sub-region data associated with MPI domain decomposition of a Dds object. |
| DdsHaloSubDomain | Halo-sub-region data associated with MPI domain decomposition of a Dds object. |
| DdsMpiInfo | MPI data associated with sub-domain of a Dds object. |
| DdsMetaData | Meta-data associated with a Dds object. |
| empty(shape[, dtype, mtype, halo, mpidims, ...]) | Creates an uninitialised Dds object of specified global size. |
| empty_like(dds[, shape, dtype, mtype, halo, ...]) | Create a new Dds object of same type/shape/MPI-layout as a specified Dds object. |
| zeros(shape[, dtype, mtype, halo, mpidims, ...]) | Creates an zero-initialised Dds object of specified global size. |
| zeros_like(dds[, shape, dtype, mtype, halo, ...]) | Create a new Dds object of same type/shape/MPI-layout as a specified Dds object. |
| ones(shape[, dtype, mtype, halo, mpidims, ...]) | Creates an one-initialised Dds object of specified global size. |
| ones_like(dds[, shape, dtype, mtype, halo, ...]) | Create a new Dds object of same type/shape/MPI-layout as a specified Dds object. |
| copy(dds[, shape, dtype, mtype, halo, ...]) | Copy-create a Dds object from an existing Dds object. |
These reduction operations use MPI all-reduce operations to calculate reductions over all non-halo parts of the sub-domains.
| minmax(input) | Returns Dds element-wise (min, max) pair. |
| sum(input[, dtype]) | Returns sum of all non-masked Dds elements. |
| sum2(input[, dtype]) | Returns sum of all non-masked Dds elements-squared. |
| count_equal_to(( (object)input, ...) | Count the number of non-masked voxels which have intensity equal to val. |
| count_not_equal_to(( (object)input, ...) | Count the number of non-masked voxels which have intensity not-equal to val. |
| count_less_than(( (object)input, ...) | Count the number of non-masked voxels which have intensity strictly less than val. |
| count_masked(( (object)input) -> object :) | Count the number of voxels which have the mask value. |
| count_non_masked(( (object)input) -> object :) | Count the number of voxels which do not have the mask value. |
| unique(input) | Returns sorted list of unique values in the specified input Dds array. |
| itemfreq(input) | Like scipy.stats.itemfreq(), returns a frequency count for the unique element values of the input array. |
| have_same_subd_decomp(dds0, dds1) | Returns True if pairs of non-halo-sub-domains on all processes have the same (global) non-halo-sub-domain origin index and same non-halo-sub-domain shape. |
| isDds(obj) | Returns whether specified obj is an instance of a mango.Dds. |
| setLoggingVerbosityLevel(...) | Sets the mango C++ filter logging level. |
| getDdsMtypeList() | Returns the list of possible mango.mtype objects for the mango.Dds mtype attribute. |
| gather(( (object)input [, ...) | Gathers MPI distributed sub-domain arrays to one (all) MPI process (processes). |
| copy_masked(src, dst[, srcval, srcmask, dstmask]) | Copies masked elements from the src array into the dst array. |
| copy_non_masked(src, dst[, srcval, srcmask]) | Copies non-masked elements from the src array into the dst array. |
| map_element_values(input, ufunc[, mtype, ...]) | Maps each element value if input array to new value calculated by ufunc. |
If True, support for MPI parallelism has been compiled in the python shared-object modules.
If True, development/prototype API is available in various modules.
If True, the mango.recon module API is available.
If True, only the mango.core module API and the mango.recon module API are available.
If True, the mango.image.registration module API is available.
If True, all distributed array element accesses are range-checked.
If True, mango has support for the 16 bit floating point types mtype("float16") and numpy.float16 .