Use CSR format consistently

[ole]

use sparse matrix (CSR) format in "build_interpolation_matrix_A"

[duncan]

This change request is in responce to memeory issues. Here's an example of the memory errors we are getting: F:\inundation\sandpits\tdhu\inundation\pyvolution>test_data_manager.py WARNING: psyco (speedup) could not import, you may want to consider installing i t WARNING: psyco (speedup) could not import, you may want to consider installing i t Reading from karratha_100m.sww Reading quantity depth Reducing quantity depth Creating grid Building mesh Building interpolation matrix Number of datapoints: 12061352 Number of basis functions: 65227 Doing 0 of 12061352 Doing 1206136 of 12061352 Doing 2412272 of 12061352 Doing 3618408 of 12061352 Doing 4824544 of 12061352 E ====================================================================== ERROR: Test that sww information can be converted correctly to asc/prj

---

Traceback (most recent call last):

File "F:\inundation\sandpits\tdhu\inundation\pyvolution\test_data_manager.py",

line 2204, in testz_sww2ers_real

verbose = True)

File "R:\inundation\sandpits\tdhu\inundation\pyvolution\data_manager.py", line

2873, in sww2ers

verbose = verbose)

File "R:\inundation\sandpits\tdhu\inundation\pyvolution\least_squares.py", lin

e 351, in init

precrop = precrop)

File "R:\inundation\sandpits\tdhu\inundation\pyvolution\least_squares.py", lin

e 382, in build_coefficient_matrix_B

precrop = precrop)

File "R:\inundation\sandpits\tdhu\inundation\pyvolution\least_squares.py", lin

e 551, in build_interpolation_matrix_A

self.A[i,j] = sigmas[j]

File "R:\inundation\sandpits\tdhu\inundation\pyvolution\sparse.py", line 66, i

n setitem

self.Data[key] = float(x)

MemoryError?

---

Ran 1 test in 16129.731s

FAILED (errors=1)

[anonymous]

CSR isn't a good format to use to build a sparse matrix, so it shouldn't be used in the building phase. So it will not reduce the max memory used. Therefore let's not implement it to save memory.

[duncan]

The last comment was me!