Fix gauge id parsing in gauge CLI and validate fields

src/pyclaw/gauges.py

@@ -326,6 +326,13 @@ def compare_gauges(paths, gauge_id, fields='all'):
         if fields.lower() == 'all':
             fields = range(gauges[0].q.shape[0])
 
+    num_fields = min(gauge.q.shape[0] for gauge in gauges)
+    bad_fields = [n for n in fields if not 0 <= n < num_fields]
+    if bad_fields:
+        raise ValueError("Fields %s out of range, gauge %s has %s fields "
+                         "available in both paths."
+                         % (bad_fields, gauge_id, num_fields))
+
     fig = plt.figure()
     fig.suptitle("Gauge %s" % gauges[0].id)
     for (i, n) in enumerate(fields):
@@ -488,6 +495,28 @@ def check_old_gauge_data(path, gauge_id, new_gauge_path="./regression_data"):
 
     return fig
 
+
+def find_gauge_ids(path):
+    r"""Return the sorted gauge ids of the gauge files found in `path`
+
+    Gauge files are named `gaugeNNNNN.txt` where the id is zero padded to at
+    least 5 digits, so ids with more digits, e.g. `gauge1234567.txt`, are
+    valid.  Files matching `gauge*.txt` without an integer id are ignored.
+
+    :Input:
+     - *path* (path) - Directory to search for gauge files.
+
+    :Output:
+     - (list) Sorted list of integer gauge ids.
+    """
+    gauge_ids = []
+    for gauge_path in Path(path).glob("gauge*.txt"):
+        id_str = gauge_path.stem[len("gauge"):]
+        if id_str.isdigit():
+            gauge_ids.append(int(id_str))
+    return sorted(gauge_ids)
+
+
 if __name__ == "__main__":
     import matplotlib.pyplot as plt
     parser = argparse.ArgumentParser()
@@ -503,8 +532,7 @@ def check_old_gauge_data(path, gauge_id, new_gauge_path="./regression_data"):
     if args.id:
         gauge_ids = args.id
     else:
-        gauge_ids = [int(str(path.name[5:10]))
-                        for path in args.path[0].glob("gauge*.txt")]
+        gauge_ids = find_gauge_ids(args.path[0])
     if args.fields:
         fields = args.fields
     else:

src/pyclaw/meson.build

@@ -42,6 +42,7 @@ python_sources = {
   ],
   'tests': [
     '__init__.py',
+    'test_gauges.py',
     'test_io.py',
   ],
 }

src/pyclaw/tests/test_gauges.py

@@ -0,0 +1,42 @@
+"""Tests for the gauge utilities in clawpack.pyclaw.gauges"""
+
+import numpy as np
+import pytest
+
+from clawpack.pyclaw import gauges
+
+
+def _write_gauge(path, gauge_id, num_eqn=2):
+    gauge = gauges.GaugeSolution()
+    gauge.id = gauge_id
+    gauge.location = (0.0, 0.0)
+    gauge.level = np.ones(3, dtype=int)
+    gauge.t = np.linspace(0.0, 1.0, 3)
+    gauge.q = np.arange(num_eqn * 3, dtype=float).reshape((num_eqn, 3))
+    gauge.write(str(path))
+
+
+def test_find_gauge_ids(tmp_path):
+    # 5-digit padded ids, an id longer than 5 digits, and non-gauge files
+    for gauge_id in (12, 1, 1234567):
+        _write_gauge(tmp_path, gauge_id)
+    (tmp_path / "gauge_notes.txt").write_text("not a gauge\n")
+    (tmp_path / "gauges.data").write_text("not a gauge\n")
+
+    assert gauges.find_gauge_ids(tmp_path) == [1, 12, 1234567]
+
+    # Long ids must round trip through the reader with the found id
+    gauge = gauges.GaugeSolution(gauge_id=1234567, path=str(tmp_path))
+    assert gauge.id == 1234567
+
+
+def test_compare_gauges_bad_fields(tmp_path):
+    pytest.importorskip("matplotlib")
+    for name, num_eqn in (("a", 2), ("b", 3)):
+        (tmp_path / name).mkdir()
+        _write_gauge(tmp_path / name, 1, num_eqn=num_eqn)
+    paths = [str(tmp_path / "a"), str(tmp_path / "b")]
+
+    # Field 2 exists only in the second path
+    with pytest.raises(ValueError, match="out of range"):
+        gauges.compare_gauges(paths, 1, fields=[0, 2])