Change default value of the observed= argument in groupby to True…

… to reflect the actual behaviour (#13296)

When grouping by a categorical, we default to the behaviour `observed=True`. I'm changing the value of the kwarg to reflect that:

```python
In [43]: df = pd.DataFrame({'a': [1, 2, 3, 4]})

In [44]: idx = pd.CategoricalIndex([1, 1, 2, 3], categories=[1, 2, 3, 4, 5])

In [45]: df.groupby(idx).sum()
Out[45]:
   a
1  3
2  3
3  4
4  0
5  0

In [46]: df.groupby(idx, observed=True).sum()
Out[46]:
   a
1  3
2  3
3  4

In [47]: df = cudf.DataFrame({'a': [1, 2, 3, 4]})

In [48]: idx = cudf.CategoricalIndex([1, 1, 2, 3], categories=[1, 2, 3, 4, 5])

In [49]: df.groupby(idx).sum()
Out[49]:
   a
1  3
3  4
2  3
```

Authors:
  - Ashwin Srinath (https://github.com/shwina)

Approvers:
  - Bradley Dice (https://github.com/bdice)

URL: #13296

python/cudf/cudf/core/dataframe.py

@@ -4072,7 +4072,7 @@ def groupby(
         sort=False,
         group_keys=False,
         squeeze=False,
-        observed=False,
+        observed=True,
         dropna=True,
     ):
         return super().groupby(

python/cudf/cudf/core/indexed_frame.py

@@ -3918,7 +3918,7 @@ def groupby(
         sort=False,
         group_keys=False,
         squeeze=False,
-        observed=False,
+        observed=True,
         dropna=True,
     ):
         if axis not in (0, "index"):
@@ -3929,7 +3929,7 @@ def groupby(
                 "squeeze parameter is not yet implemented"
             )
 
-        if observed is not False:
+        if not observed:
             raise NotImplementedError(
                 "observed parameter is not yet implemented"
             )

python/cudf/cudf/core/series.py

@@ -3297,7 +3297,7 @@ def groupby(
         sort=False,
         group_keys=False,
         squeeze=False,
-        observed=False,
+        observed=True,
         dropna=True,
     ):
         return super().groupby(