fetch.py: address tz-related error during pd.concat

See issue #11.

This explicit index construction will create a
tz-aware DatetimeIndex even if the data is empty:


>>> pd.DatetimeIndex(data=[], tz="UTC")
DatetimeIndex([], dtype='datetime64[ns, UTC]', freq=None)

fetch.py

@@ -234,21 +234,22 @@ def clones_or_views_to_df(items, metric):
     series_count_total = []
     series_count_unique = []
     series_timestamps = []
+
     for sample in items:
-        # GitHub API docs say
-        # "Timestamps are aligned to UTC"
-        # `sample.timestamp` is a naive datetime object. Make it tz-aware.
-        ts_aware = pytz.timezone("UTC").localize(sample.timestamp)
-        series_timestamps.append(ts_aware)
+        # GitHub API docs say "Timestamps are aligned to UTC".
+        # `sample.timestamp` is a tz-naive datetime object.
+        series_timestamps.append(sample.timestamp)
         series_count_total.append(int(sample.count))
         series_count_unique.append(int(sample.uniques))
 
+    # Attach timezone information to `pd.DatetimeIndex` (make this index
+    # tz-aware, leave actual numbers intact).
     df = pd.DataFrame(
         data={
             f"{metric}_total": series_count_total,
             f"{metric}_unique": series_count_unique,
         },
-        index=series_timestamps,
+        index=pd.DatetimeIndex(data=series_timestamps, tz="UTC"),
     )
     df.index.name = "time_iso8601"