fix_514 (#515)

d-chambers · web-flow · commit 72c7878edfd5 · 2025-08-22T15:41:44.000+01:00
diff --git a/dascore/io/prodml/__init__.py b/dascore/io/prodml/__init__.py
@@ -1,8 +1,6 @@
 """
 Support for prodML format.
 
-This is used by Silixa's iDAS as an HDF5 format and perhaps other interrogators.
-
 More info about ProdML can be found here:
 https://www.energistics.org/prodml-developers-users
 """
diff --git a/dascore/io/prodml/utils.py b/dascore/io/prodml/utils.py
@@ -2,6 +2,8 @@
 
 from __future__ import annotations
 
+import numpy as np
+
 import dascore as dc
 from dascore.constants import VALID_DATA_TYPES
 from dascore.core.coordmanager import get_coord_manager
@@ -60,13 +62,29 @@ def get_distance_units(attrs):
 
 def _get_time_coord(node):
     """Get the time information from a Raw node."""
-    time_attrs = node["RawDataTime"].attrs
+    time_array = node["RawDataTime"]
+    array_len = len(time_array)
+    assert array_len > 0, "Missing time array in ProdML file."
+    time_attrs = time_array.attrs
     start_str = unbyte(time_attrs["PartStartTime"]).split("+")[0]
     start = dc.to_datetime64(start_str.rstrip("Z"))
     end_str = unbyte(time_attrs["PartEndTime"]).split("+")[0]
     end = dc.to_datetime64(end_str.rstrip("Z"))
-    step = (end - start) / (len(node["RawDataTime"]) - 1)
-    return get_coord(start=start, stop=end + step, step=step, units="s")
+    step = (end - start) / (array_len - 1)
+    time_coord = get_coord(start=start, stop=end + step, step=step, units="s")
+    # Sometimes the "PartEndTime" can be wrong. Check for this and try to
+    # compensate. See #414.
+    last = np.asarray(time_array[-1:]).astype("datetime64[us]")[0]
+    tc_max = np.asarray(time_coord.max()).astype("datetime64[us]")
+    diff = float(np.abs((tc_max - last) / step))
+    # Note: just in case the time array is not in microseconds as it should
+    # be, we prefer to use the iso 8601 strings in the 'PartStartTime' attrs
+    # because they are less likely to get messed up. Therefore, we only
+    # correct time coordinate from time array if the values are "close" but off.
+    if 0 < diff < 10:
+        time_array = time_array[:].astype("datetime64[us]")
+        time_coord = get_coord(data=time_array)
+    return time_coord
 
 
 def _get_data_unit_and_type(node):
@@ -108,9 +126,9 @@ def _get_prodml_attrs(fi, extras=None) -> list[dict]:
     for node in raw_nodes.values():
         info = dict(base_info)
         t_coord = _get_time_coord(node)
-        info.update(t_coord.get_attrs_dict("time"))
         info.update(_get_data_unit_and_type(node))
-        info["dims"] = _get_dims(node)
+        dims = _get_dims(node)
+        info["dims"] = dims
         if extras is not None:
             info.update(extras)
         info["coords"] = {"time": t_coord, "distance": d_coord}
diff --git a/tests/test_io/test_prodml/test_prod_ml.py b/tests/test_io/test_prodml/test_prod_ml.py
@@ -4,11 +4,13 @@
 
 import shutil
 
+import h5py
 import pandas as pd
 import pytest
 import tables
 
 import dascore as dc
+from dascore.core.coords import get_coord
 from dascore.io.core import read
 from dascore.utils.downloader import fetch
 
@@ -53,6 +55,22 @@ def issue_221_patch_path(self, tmp_path_factory):
             node._v_attrs.Dimensions = new_dims
         return new_path
 
+    @pytest.fixture(scope="class")
+    def issue_514_patch_path(self, tmp_path_factory):
+        """Make a patch with bad endtime metadata. See #412."""
+        tmp_path = tmp_path_factory.mktemp("issue_514")
+        path = dc.utils.downloader.fetch("prodml_2.0.h5")
+        new_path = shutil.copy2(path, tmp_path / "prod_2_issue_514.h5")
+        with h5py.File(new_path, "a") as fi:
+            # monkey patch dimensions to simulate issue.
+            parent_node = fi["Acquisition"]["Raw[0]"]
+            time_node = parent_node["RawDataTime"]
+            time = time_node[:].astype("datetime64[us]")
+            time_coord = get_coord(data=time)
+            new_time = str(time_coord.max() + time_coord.step * 2)
+            time_node.attrs["PartEndTime"] = new_time
+        return new_path
+
     @pytest.fixture(scope="class")
     def silixa_h5_patch(self, idas_h5_example_path):
         """Get the silixa file, return Patch."""
@@ -76,6 +94,11 @@ def test_issue_221(self, issue_221_patch_path):
         patch = dc.read(issue_221_patch_path)[0]
         assert isinstance(patch, dc.Patch)
 
+    def test_issue_514(self, issue_514_patch_path):
+        """Ensure the patch can be read despite bad attribute info."""
+        patch = dc.read(issue_514_patch_path)[0]
+        assert isinstance(patch, dc.Patch)
+
 
 class TestReadQuantXV2:
     """Tests for reading the QuantXV2 format."""
