Adding helper to parse a Bigtable column family protobuf.

dhermes · dhermes · commit e361f79da022 · 2016-02-15T14:20:56.000-08:00
diff --git a/gcloud/bigtable/row.py b/gcloud/bigtable/row.py
@@ -19,6 +19,7 @@
 
 import six
 
+from gcloud._helpers import _datetime_from_microseconds
 from gcloud._helpers import _microseconds_from_datetime
 from gcloud._helpers import _to_bytes
 from gcloud.bigtable._generated import bigtable_data_pb2 as data_pb2
@@ -1192,3 +1193,41 @@ def to_pb(self):
             condition_kwargs['false_filter'] = self.false_filter.to_pb()
         condition = data_pb2.RowFilter.Condition(**condition_kwargs)
         return data_pb2.RowFilter(condition=condition)
+
+
+def _parse_family_pb(family_pb):
+    """Parses a Family protobuf into a dictionary.
+
+    :type family_pb: :class:`._generated.bigtable_data_pb2.Family`
+    :param family_pb: A protobuf
+
+    :rtype: tuple
+    :returns: A string and dictionary. The string is the name of the
+              column family and the dictionary has column names (within the
+              family) as keys and cell lists as values. Each cell is
+              represented with a two-tuple with the value (in bytes) and the
+              timestamp for the cell. For example:
+
+              .. code:: python
+
+                  {
+                      b'col-name1': [
+                          (b'cell-val', datetime.datetime(...)),
+                          (b'cell-val-newer', datetime.datetime(...)),
+                      ],
+                      b'col-name2': [
+                          (b'altcol-cell-val', datetime.datetime(...)),
+                      ],
+                  }
+    """
+    result = {}
+    for column in family_pb.columns:
+        result[column.qualifier] = cells = []
+        for cell in column.cells:
+            val_pair = (
+                cell.value,
+                _datetime_from_microseconds(cell.timestamp_micros),
+            )
+            cells.append(val_pair)
+
+    return family_pb.name, result
diff --git a/gcloud/bigtable/test_row.py b/gcloud/bigtable/test_row.py
@@ -1535,6 +1535,65 @@ def test_to_pb_false_only(self):
         self.assertEqual(filter_pb, expected_pb)
 
 
+class Test__parse_family_pb(unittest2.TestCase):
+
+    def _callFUT(self, family_pb):
+        from gcloud.bigtable.row import _parse_family_pb
+        return _parse_family_pb(family_pb)
+
+    def test_it(self):
+        from gcloud._helpers import _datetime_from_microseconds
+        from gcloud.bigtable._generated import bigtable_data_pb2 as data_pb2
+
+        col_fam1 = u'col-fam-id'
+        col_name1 = b'col-name1'
+        col_name2 = b'col-name2'
+        cell_val1 = b'cell-val'
+        cell_val2 = b'cell-val-newer'
+        cell_val3 = b'altcol-cell-val'
+
+        microseconds = 5554441037
+        timestamp = _datetime_from_microseconds(microseconds)
+        expected_dict = {
+            col_name1: [
+                (cell_val1, timestamp),
+                (cell_val2, timestamp),
+            ],
+            col_name2: [
+                (cell_val3, timestamp),
+            ],
+        }
+        expected_output = (col_fam1, expected_dict)
+        sample_input = data_pb2.Family(
+            name=col_fam1,
+            columns=[
+                data_pb2.Column(
+                    qualifier=col_name1,
+                    cells=[
+                        data_pb2.Cell(
+                            value=cell_val1,
+                            timestamp_micros=microseconds,
+                        ),
+                        data_pb2.Cell(
+                            value=cell_val2,
+                            timestamp_micros=microseconds,
+                        ),
+                    ],
+                ),
+                data_pb2.Column(
+                    qualifier=col_name2,
+                    cells=[
+                        data_pb2.Cell(
+                            value=cell_val3,
+                            timestamp_micros=microseconds,
+                        ),
+                    ],
+                ),
+            ],
+        )
+        self.assertEqual(expected_output, self._callFUT(sample_input))
+
+
 class _Client(object):
 
     data_stub = None
