Added an optimization to use multithreading in gdalwarp (#69)

* Added an optimization to use multithreading in gdalwarp

* Updated changelog and service version

Author: jackiryan

CHANGELOG.md

@@ -4,6 +4,11 @@ HyBIG follows semantic versioning. All notable changes to this project will be
 documented in this file. The format is based on [Keep a
 Changelog](http://keepachangelog.com/en/1.0.0/).
 
+## [2.8.2] - 2026-07-30
+
+* Added multithreading to gdalwarp call to improve performance.
+
+
 ## [2.8.1] - 2026-07-28
 
 * [GITC-9248](https://bugs.earthdata.nasa.gov/browse/GITC-9248): Reduced number of output tiles from HyBIG service through a number of optimizations.

docker/service_version.txt

@@ -1 +1 @@
-2.8.1
+2.8.2

hybig/browse.py

@@ -1,5 +1,6 @@
 """Module containing core functionality for browse image generation."""
 
+import os
 import re
 from itertools import zip_longest
 from logging import Logger, getLogger
@@ -37,6 +38,32 @@
 
 DST_NODATA = NODATA_IDX
 
+# Upper bound on gdalwarp threads, aka rasterio.warp.reproject
+MAX_WARP_THREADS = 8
+
+
+def warp_thread_count() -> int:
+    """Return the thread count to use for GDAL warp/reproject.
+
+    gdalwarp parallelizes the reprojection across output chunks, which is
+    the dominant cost for large jobs but it has minimal impact on memory
+    to multithread.
+
+    The count is the number of CPUs available to the process (uses
+    container CPU limits on Linux based on scheduler affinity), capped at
+    MAX_WARP_THREADS. Operators may override it with the HYBIG_NUM_THREADS
+    environment variable.
+    """
+    # sched_getaffinity checks container cpuset limits but is Linux-only, so
+    # fall back to the reported CPU count on other platforms (local dev)
+    sched_getaffinity = getattr(os, 'sched_getaffinity', None)
+    if sched_getaffinity is not None:
+        available = len(sched_getaffinity(0))
+    else:
+        available = os.cpu_count() or 1
+
+    return max(1, min(available, MAX_WARP_THREADS))
+
 
 def create_browse(
     source_tiff: str,
@@ -806,6 +833,7 @@ def write_georaster_as_browse(
         dst_crs=grid_parameters['crs'],
         dst_nodata=int(dst_nodata),
         resampling=Resampling.nearest,
+        num_threads=warp_thread_count(),
     )
 
     # Skip tiles that turned out empty once mapped onto the actual footprint.

tests/test_service/test_adapter.py

@@ -19,6 +19,7 @@
 from harmony_service.exceptions import HyBIGServiceError
 from hybig.browse import (
     convert_multiband_to_raster,
+    warp_thread_count,
 )
 from tests.utilities import Granule, create_stac
 
@@ -249,6 +250,7 @@ def move_tif(*args, **kwargs):
         )
         self.assertEqual(actual_call.kwargs['dst_crs'], expected_params['crs'])
         self.assertEqual(actual_call.kwargs['resampling'], Resampling.nearest)
+        self.assertEqual(actual_call.kwargs['num_threads'], warp_thread_count())
 
         # Ensure the browse image and ESRI world file were staged as expected:
         mock_stage.assert_has_calls(
@@ -453,6 +455,7 @@ def fake_reproject(*args, **kwargs):
             actual_call.kwargs['dst_nodata'], expected_params['dst_nodata']
         )
         self.assertEqual(actual_call.kwargs['resampling'], Resampling.nearest)
+        self.assertEqual(actual_call.kwargs['num_threads'], warp_thread_count())
 
         # Ensure the browse image and ESRI world file were staged as expected:
         mock_stage.assert_has_calls(

tests/unit/test_browse.py

@@ -19,6 +19,7 @@
 from rasterio.warp import Resampling
 
 from hybig.browse import (
+    MAX_WARP_THREADS,
     convert_multiband_to_raster,
     convert_singleband_to_raster,
     create_browse,
@@ -31,6 +32,7 @@
     reprojected_output_is_empty,
     validate_file_crs,
     validate_file_type,
+    warp_thread_count,
     write_georaster_as_browse,
 )
 from hybig.color_utility import (
@@ -241,6 +243,8 @@ def test_create_browse_imagery_with_mocks(self, rasterio_open_mock, reproject_mo
         self.assertEqual(actual_call.kwargs['dst_crs'], CRS.from_string('EPSG:4326'))
         self.assertEqual(actual_call.kwargs['dst_nodata'], 0)  # TRANSPARENT
         self.assertEqual(actual_call.kwargs['resampling'], Resampling.nearest)
+        # Reprojection runs multithreaded to speed up large jobs.
+        self.assertEqual(actual_call.kwargs['num_threads'], warp_thread_count())
 
         self.assertEqual(
             (self.tmp_dir / 'input_file_path.jpg').resolve(), actual_image.resolve()
@@ -442,6 +446,12 @@ def test_process_tile_returns_false_for_all_nan_window(self, rasterio_open_mock)
             f'Skipping all-NaN tile: {self.tmp_dir / "output.png"}'
         )
 
+    def test_warp_thread_count_is_bounded(self):
+        """Thread count is between 1 and MAX_WARP_THREADS."""
+        count = warp_thread_count()
+        self.assertGreaterEqual(count, 1)
+        self.assertLessEqual(count, MAX_WARP_THREADS)
+
     def test_reprojected_output_is_empty(self):
         """Test reprojected_output_is_empty across output types."""
         # Paletted single-band: every cell holds the fill index -> empty.