Bugfix/1841/maintain empty series names (#1983)

#### Reference Issues/PRs
Fixes #1841 

#### What does this implement or fix?
Before this change, if a `Series` had an empty-string as a name, this
would be roundtripped as a `None`.
This introduces a `has_name` bool to the normalization metadata
protobuf, as a backwards-compatible way of effectively making the `name`
field optional.

The behaviour (which has been verified) can be summarised as follows:
```
Writer version | Series name | Protobuf name field | Protobuf has_name field | Series name read by <=5.0.0 | Series name read by this branch
---------------|-------------|---------------------|-------------------------|-----------------------------|--------------------------------
       <=5.0.0 |     "hello" |             "hello" |             Not present |                     "hello" |                         "hello"
       <=5.0.0 |          "" |                  "" |             Not present |                        None |                            None
       <=5.0.0 |        None |                  "" |             Not present |                        None |                            None
   This branch |     "hello" |             "hello" |                    True |                     "hello" |                         "hello"
   This branch |          "" |                  "" |                    True |                        None |                              ""
   This branch |        None |                  "" |                   False |                        None |                            None
```

cpp/proto/arcticc/pb2/descriptors.proto

@@ -163,6 +163,8 @@ message NormalizationMetadata {
 
         // Pandas series support arbitrary objects but restricting to a string seems reasonable
         string name = 1;
+        // Backwards-compatible way of treating name as optional
+        bool has_name = 9;
 
         oneof index_type {
             PandasIndex index = 2;
@@ -190,6 +192,7 @@ message NormalizationMetadata {
 
         map<string, ColumnName> col_names = 7;
         PandasIndex columns = 8;
+        // Note that 9 is taken above by has_name, so the next added field should be 10
     }
 
     message PandasDataFrame {

python/arcticdb/version_store/_normalization.py

@@ -608,8 +608,10 @@ def normalize(self, item, string_max_len=None, dynamic_strings=False, coerce_col
             empty_types=empty_types
         )
         norm.series.CopyFrom(norm.df)
-        if item.name:
+        if item.name is not None:
             norm.series.common.name = _column_name_to_strings(item.name)
+            norm.series.common.has_name = True
+        # else protobuf bools default to False
 
         return NormalizedInput(item=df, metadata=norm)
 
@@ -620,9 +622,12 @@ def denormalize(self, item, norm_meta):
 
         series = pd.Series() if df.columns.empty else df.iloc[:, 0]
 
-        if norm_meta.common.name:
+        if len(norm_meta.common.name) or norm_meta.common.has_name:
             series.name = norm_meta.common.name
         else:
+            # Either the Series was written with a new client that understands the has_name field, and it was None, or
+            # the Series was written by an older client as either an empty string or None, we cannot tell, so maintain
+            # behaviour as it was before the has_name field was added
             series.name = None
 
         return series

python/tests/unit/arcticdb/version_store/test_normalization.py

@@ -495,6 +495,25 @@ def test_columns_names_timeframe(lmdb_version_store, sym):
     assert tf == vit.data
 
 
+@pytest.mark.parametrize("name", (None, "", "non_empty"))
+def test_roundtrip_series_name(lmdb_version_store_v1, name):
+    lib = lmdb_version_store_v1
+    sym = "test_roundtrip_series_name"
+    series = pd.Series(np.arange(1), name=name)
+    lib.write(sym, series)
+    assert_series_equal(series, lib.read(sym).data)
+
+
+@pytest.mark.parametrize("name", (None, "", "non_empty"))
+def test_roundtrip_index_name(lmdb_version_store_v1, name):
+    lib = lmdb_version_store_v1
+    sym = "test_roundtrip_index_name"
+    df = pd.DataFrame({"col": [0]}, index=[pd.Timestamp(0)])
+    df.index.name = name
+    lib.write(sym, df)
+    assert_frame_equal(df, lib.read(sym).data)
+
+
 def test_columns_names_series(lmdb_version_store, sym):
     dr = pd.date_range("2020-01-01", "2020-12-31", name="date")
     date_series = pd.Series(dr, index=dr)