Jake/met_office_radar_data
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

more speed ups

Browse Source
This commit is contained in:
Jake Pullen
2025-11-19 07:46:07 +00:00
parent 1373432b02
commit 407baa86ad
1 changed files with 18 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/nimrod.py
+18 -18
View File
@@ -68,6 +68,7 @@ import struct
import array import array
import argparse import argparse
from typing import IO from typing import IO
import numpy as np
class Nimrod: class Nimrod:
@@ -89,7 +90,7 @@ class Nimrod:
x_pixel_size (float): Size of pixel in x-direction x_pixel_size (float): Size of pixel in x-direction
x_right (float): Right easting coordinate x_right (float): Right easting coordinate
y_bottom (float): Bottom northing coordinate y_bottom (float): Bottom northing coordinate
data (array.array): Raster data array data (np.ndarray): Raster data array
""" """
class RecordLenError(Exception): class RecordLenError(Exception):
@@ -253,11 +254,16 @@ class Nimrod:
array_size = self.ncols * self.nrows array_size = self.ncols * self.nrows
check_record_len(infile, array_size * 2, "data start") check_record_len(infile, array_size * 2, "data start")
self.data = array.array("h")
try: try:
data = infile.read(array_size * 2) # Read data as big-endian 16-bit integers
self.data.frombytes(data) # numpy.frombuffer is efficient for reading from bytes
self.data.byteswap() data_bytes = infile.read(array_size * 2)
self.data = np.frombuffer(data_bytes, dtype='>h').astype(np.int16)
# Reshape to (nrows, ncols) for easier 2D manipulation
# Note: NIMROD data is row-major (C-style), starting from top-left
self.data = self.data.reshape((self.nrows, self.ncols))
except Exception: except Exception:
infile.close() infile.close()
raise Nimrod.PayloadReadError raise Nimrod.PayloadReadError
@@ -383,16 +389,12 @@ class Nimrod:
yMinPixelId = int((self.y_top - ymax) / self.y_pixel_size + 0.5) yMinPixelId = int((self.y_top - ymax) / self.y_pixel_size + 0.5)
yMaxPixelId = int((self.y_top - ymin) / self.y_pixel_size + 0.5) yMaxPixelId = int((self.y_top - ymin) / self.y_pixel_size + 0.5)
bbox_data = [] # Use numpy slicing to extract the sub-array
for i in range(yMinPixelId, yMaxPixelId + 1): # Note: y indices correspond to rows, x indices to columns
bbox_data.extend( # Slicing is [start:end], so we need +1 for the end index
self.data[ self.data = self.data[yMinPixelId : yMaxPixelId + 1, xMinPixelId : xMaxPixelId + 1]
i * self.ncols + xMinPixelId : i * self.ncols + xMaxPixelId + 1
]
)
# Update object where necessary # Update object where necessary
self.data = bbox_data
self.x_right = self.x_left + xMaxPixelId * self.x_pixel_size self.x_right = self.x_left + xMaxPixelId * self.x_pixel_size
self.x_left += xMinPixelId * self.x_pixel_size self.x_left += xMinPixelId * self.x_pixel_size
self.ncols = xMaxPixelId - xMinPixelId + 1 self.ncols = xMaxPixelId - xMinPixelId + 1
@@ -431,11 +433,9 @@ class Nimrod:
outfile.write("cellsize %.1f\n" % self.y_pixel_size) outfile.write("cellsize %.1f\n" % self.y_pixel_size)
outfile.write("nodata_value %.1f\n" % self.hdr_element[38]) outfile.write("nodata_value %.1f\n" % self.hdr_element[38])
# Write raster data to output file # Write raster data to output file using numpy.savetxt
for i in range(self.nrows): # This is significantly faster than iterating in Python
for j in range(self.ncols - 1): np.savetxt(outfile, self.data, fmt='%d', delimiter=' ')
outfile.write("%d " % self.data[i * self.ncols + j])
outfile.write("%d\n" % self.data[i * self.ncols + self.ncols - 1])
outfile.close() outfile.close()