dist-s1-enumerator 1.0.4__tar.gz → 1.0.5__tar.gz

{dist_s1_enumerator-1.0.4 → dist_s1_enumerator-1.0.5}/CHANGELOG.md

@@ -6,6 +6,12 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [PEP 440](https://www.python.org/dev/peps/pep-0440/)
 and uses [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.0.5] - 2025-09-29
+
+### Fixed
+* CMR metadata does not have correctly migrated urls from ASF datapool to ASF cumulus - see https://github.com/opera-adt/dist-s1/issues/158.
+
+
 ## [1.0.4] - 2025-09-29
 
 ### Added
@@ -14,7 +20,7 @@ and uses [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 * We now ensure that single polarization data is excluded from baselines and not used in the creation of the post-image set.
 
 ### Fixed
-* Bug in enumerating 1 product.
+* Bug in enumerating 1 product - we did not ensure spatial bursts were consistent between pre-/post-image sets.
 
 
 ## [1.0.3] - 2025-09-09

{dist_s1_enumerator-1.0.4/src/dist_s1_enumerator.egg-info → dist_s1_enumerator-1.0.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dist-s1-enumerator
-Version: 1.0.4
+Version: 1.0.5
 Summary: Enumeration and ops library for the OPERA DIST-S1 project
 Author-email: "Richard West, Charlie Marshak, Talib Oliver-Cabrera, and Jungkyo Jung" <charlie.z.marshak@jpl.nasa.gov>
 License: Apache-2.0

{dist_s1_enumerator-1.0.4 → dist_s1_enumerator-1.0.5}/src/dist_s1_enumerator/asf.py

@@ -12,6 +12,27 @@ from dist_s1_enumerator.mgrs_burst_data import get_burst_ids_in_mgrs_tiles, get_
 from dist_s1_enumerator.tabular_models import reorder_columns, rtc_s1_resp_schema, rtc_s1_schema
 
 
+def convert_asf_url_to_cumulus(url: str) -> str:
+    asf_base = 'https://datapool.asf.alaska.edu/RTC/OPERA-S1/'
+    cumulus_base = 'https://cumulus.asf.earthdatacloud.nasa.gov/OPERA/OPERA_L2_RTC-S1/'
+
+    if not (url.startswith(cumulus_base) or url.startswith(asf_base)):
+        warn(f'URL {url} is not a valid ASF datapool or cumulus earthdatacloud URL.')
+        return url
+
+    if not url.startswith(asf_base):
+        return url
+
+    filename = url.split('/')[-1]
+    granule_pol_parts = filename.rsplit('_', 1)
+    if len(granule_pol_parts) != 2:
+        raise ValueError(f'Could not extract granule name from filename: {filename}')
+
+    granule_name = granule_pol_parts[0]
+    new_url = f'{cumulus_base}{granule_name}/{filename}'
+    return new_url
+
+
 def format_polarization(pol: list | str) -> str:
     if isinstance(pol, list):
         if ('VV' in pol) and len(pol) == 2:
@@ -167,6 +188,8 @@ def get_rtc_s1_ts_metadata_by_burst_ids(
 
     df_rtc['url_copol'] = url_copol
     df_rtc['url_crosspol'] = url_crosspol
+    df_rtc['url_copol'] = df_rtc['url_copol'].map(convert_asf_url_to_cumulus)
+    df_rtc['url_crosspol'] = df_rtc['url_crosspol'].map(convert_asf_url_to_cumulus)
     df_rtc = df_rtc.drop(columns=['all_urls'])
 
     # Ensure the data is sorted by jpl_burst_id and acq_dt

{dist_s1_enumerator-1.0.4 → dist_s1_enumerator-1.0.5/src/dist_s1_enumerator.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dist-s1-enumerator
-Version: 1.0.4
+Version: 1.0.5
 Summary: Enumeration and ops library for the OPERA DIST-S1 project
 Author-email: "Richard West, Charlie Marshak, Talib Oliver-Cabrera, and Jungkyo Jung" <charlie.z.marshak@jpl.nasa.gov>
 License: Apache-2.0

{dist_s1_enumerator-1.0.4 → dist_s1_enumerator-1.0.5}/tests/test_asf.py

@@ -1,6 +1,6 @@
 import pytest
 
-from dist_s1_enumerator.asf import append_pass_data, get_rtc_s1_ts_metadata_by_burst_ids
+from dist_s1_enumerator.asf import append_pass_data, convert_asf_url_to_cumulus, get_rtc_s1_ts_metadata_by_burst_ids
 
 
 @pytest.mark.integration
@@ -35,3 +35,24 @@ def test_appending_mgrs_tiles() -> None:
 
     df_rtc_formatted_no_rows = append_pass_data(df_rtc_resp, ['22NFF'])
     assert df_rtc_formatted_no_rows.empty
+
+
+@pytest.mark.parametrize('pol_token', ['VV', 'VH', 'HH', 'HV'])
+def test_convert_asf_url_to_cumulus_from_datapool(pol_token: str) -> None:
+    """Test converting ASF datapool URL to cumulus earthdatacloud URL."""
+    asf_url = f'https://datapool.asf.alaska.edu/RTC/OPERA-S1/OPERA_L2_RTC-S1_T001-000189-IW2_20211028T180924Z_20250703T015334Z_S1A_30_v1.0_{pol_token}.tif'
+    expected_cumulus_url = f'https://cumulus.asf.earthdatacloud.nasa.gov/OPERA/OPERA_L2_RTC-S1/OPERA_L2_RTC-S1_T001-000189-IW2_20211028T180924Z_20250703T015334Z_S1A_30_v1.0/OPERA_L2_RTC-S1_T001-000189-IW2_20211028T180924Z_20250703T015334Z_S1A_30_v1.0_{pol_token}.tif'
+
+    result = convert_asf_url_to_cumulus(asf_url)
+
+    assert result == expected_cumulus_url
+
+
+@pytest.mark.parametrize('pol_token', ['VV', 'VH', 'HH', 'HV'])
+def test_convert_asf_url_to_cumulus_already_cumulus(pol_token: str) -> None:
+    """Test that cumulus URLs are returned unchanged."""
+    cumulus_url = f'https://cumulus.asf.earthdatacloud.nasa.gov/OPERA/OPERA_L2_RTC-S1/OPERA_L2_RTC-S1_T001-000189-IW2_20211028T180924Z_20250703T015334Z_S1A_30_v1.0/OPERA_L2_RTC-S1_T001-000189-IW2_20211028T180924Z_20250703T015334Z_S1A_30_v1.0_{pol_token}.tif'
+
+    result = convert_asf_url_to_cumulus(cumulus_url)
+
+    assert result == cumulus_url