Fixes for different formats for gate ID metadata attributes

cmac/cmac_ppi_quicklooks.py

@@ -16,6 +16,7 @@
     generate_radar_name, generate_radar_time_begin)
 
 from .config import get_plot_values, get_field_names
+from .gate_id import get_gate_id_categories, gate_id_has_category
 
 plt.switch_backend('agg')
 
@@ -190,13 +191,12 @@ def _range(key, default):
 
     # Four panel plot of gate_id, velocity_texture, reflectivity, and
     # cross_correlation_ratio.
-    cat_dict = {}
+    gate_id_field = radar.fields['gate_id']
+    cat_dict = get_gate_id_categories(gate_id_field)
     print('##')
     print('## Keys for each gate id are as follows:')
-    for i, pair_str in enumerate(radar.fields['gate_id']['notes'].split(',')):
-        pair_str = pair_str.split(':')[1].strip()
-        print('##   ', str(pair_str))
-        cat_dict.update({pair_str: i})
+    for label in sorted(cat_dict, key=cat_dict.get):
+        print('##   ', str(label))
     sorted_cats = sorted(cat_dict.items(), key=operator.itemgetter(1))
     cat_colors = dict(cat_colors_cfg)
     lab_colors = [cat_colors[kitty[0]] for kitty in sorted_cats]
@@ -218,7 +218,8 @@ def _range(key, default):
                         colors='k')
 
     cbax = ax[0, 0]
-    if 'ground_clutter' in radar.fields.keys() or 'terrain_blockage' in radar.fields['gate_id']['notes']:
+    if ('ground_clutter' in radar.fields.keys()
+            or gate_id_has_category(gate_id_field, 'terrain_blockage')):
         tick_locs = np.linspace(
             0, len(sorted_cats) - 1, len(sorted_cats)) + 0.5
     else:

cmac/cmac_rhi_quicklooks.py

@@ -15,6 +15,7 @@
     generate_radar_name, generate_radar_time_begin)
 
 from .config import get_plot_values, get_field_names
+from .gate_id import get_gate_id_categories, gate_id_has_category
 
 plt.switch_backend('agg')
 
@@ -106,13 +107,12 @@ def _range(key, default):
 
     # Four panel plot of gate_id, velocity_texture, reflectivity, and
     # cross_correlation_ratio.
-    cat_dict = {}
+    gate_id_field = radar.fields['gate_id']
+    cat_dict = get_gate_id_categories(gate_id_field)
     print('##')
     print('## Keys for each gate id are as follows:')
-    for i, pair_str in enumerate(radar.fields['gate_id']['notes'].split(',')):
-        pair_str = pair_str.split(':')[1].strip()
-        print('##   ', str(pair_str))
-        cat_dict.update({pair_str: i})
+    for label in sorted(cat_dict, key=cat_dict.get):
+        print('##   ', str(label))
     sorted_cats = sorted(cat_dict.items(), key=operator.itemgetter(1))
 
     cat_colors = dict(cat_colors_cfg)
@@ -126,7 +126,8 @@ def _range(key, default):
                      cmap=cmap, vmin=0, vmax=6)
 
     cbax = ax[0, 0]
-    if 'ground_clutter' in radar.fields.keys() or 'terrain_blockage' in radar.fields['gate_id']['notes']:
+    if ('ground_clutter' in radar.fields.keys()
+            or gate_id_has_category(gate_id_field, 'terrain_blockage')):
         tick_locs = np.linspace(
             0, len(sorted_cats) - 1, len(sorted_cats)) + 0.5
     else:

cmac/gate_id.py

@@ -0,0 +1,91 @@
+"""
+Helpers for interpreting the category metadata attached to a ``gate_id``
+(hydrometeor ID) radar field.
+
+CMAC's own gate id fields document their categories with a ``notes``
+attribute, which in practice shows up in a few different shapes:
+
+- ``"0: multi_trip, 1: rain, 2: snow"`` -- comma separated ``"index: label"``
+  pairs, colon separated.
+- ``"0 multi_trip, 1 rain, 2 snow"`` -- comma separated ``"index label"``
+  pairs, whitespace separated.
+- ``"multi_trip rain snow melting no_scatter clutter terrain_blockage"`` --
+  a plain, unindexed list of labels in order, with no indices or commas at
+  all.
+
+Fields that follow the CF conventions instead (or radar objects re-read
+from a file that converted ``notes`` on save) may document the same
+information with a ``flag_meanings`` attribute and a parallel
+``flag_values`` attribute (the matching integer codes). ``flag_meanings``
+is generally comma separated, though a plain whitespace separated string
+is also accepted.
+"""
+
+import re
+
+_PAIR_SEP_RE = re.compile(r'[:\s]+')
+
+
+def _label_from_pair(pair_str):
+    """Extract the category label from a single ``"index: label"`` pair,
+    where the index/label separator is a colon, whitespace, or both."""
+    parts = _PAIR_SEP_RE.split(pair_str.strip(), maxsplit=1)
+    return parts[-1].strip()
+
+
+def _split_list(text):
+    """Split a comma or whitespace separated list of labels into its
+    individual, stripped entries."""
+    if ',' in text:
+        parts = text.split(',')
+    else:
+        parts = text.split()
+    return [part.strip() for part in parts if part.strip()]
+
+
+def _labels_from_notes(notes):
+    """Return the ordered list of category labels encoded in a ``notes``
+    attribute, handling both indexed ``"index: label"``/``"index label"``
+    pairs and a plain, unindexed list of labels."""
+    pieces = [p.strip() for p in notes.split(',') if p.strip()]
+    if len(pieces) > 1 or (pieces and ':' in pieces[0]):
+        return [_label_from_pair(piece) for piece in pieces]
+    return _split_list(notes)
+
+
+def get_gate_id_categories(gate_id_field):
+    """
+    Return a dict mapping each gate id category label to its integer code.
+
+    Parameters
+    ----------
+    gate_id_field : dict
+        A Py-ART field dictionary, e.g. ``radar.fields['gate_id']``.
+
+    """
+    if 'notes' in gate_id_field:
+        labels = _labels_from_notes(gate_id_field['notes'])
+        return {label: i for i, label in enumerate(labels)}
+
+    if 'flag_meanings' in gate_id_field and 'flag_values' in gate_id_field:
+        labels = _split_list(gate_id_field['flag_meanings'])
+        values = gate_id_field['flag_values']
+        return {label: int(value) for label, value in zip(labels, values)}
+
+    raise KeyError(
+        "The 'gate_id' field must have either a 'notes' attribute or "
+        "'flag_values'/'flag_meanings' attributes describing its "
+        "categories.")
+
+
+def gate_id_has_category(gate_id_field, category):
+    """
+    Return True if ``category`` is one of the documented categories of a
+    ``gate_id`` field, whether documented via ``notes`` or via
+    ``flag_meanings``.
+    """
+    if 'notes' in gate_id_field:
+        return category in _labels_from_notes(gate_id_field['notes'])
+    if 'flag_meanings' in gate_id_field:
+        return category in _split_list(gate_id_field['flag_meanings'])
+    return False