[load_examples] download data at runtime (#7314)

* [load_examples] download data at runtime

When running `superset load_examples` to load example data sets,
Superset used to load from the local package. This created a few issues
notably around licensing (what are these datasets licensed as?) and
around package size.

For now, I moved the data sets here:
https://github.com/apache-superset/examples-data

Altered the logic to download the data from where it is stored.

* flakes

superset/data/unicode_test_data.py

@@ -16,7 +16,6 @@
 # under the License.
 import datetime
 import json
-import os
 import random
 
 import pandas as pd
@@ -27,7 +26,7 @@ from superset.utils import core as utils
 from .helpers import (
     config,
     Dash,
-    DATA_FOLDER,
+    get_example_data,
     get_slice_json,
     merge_slice,
     Slice,
@@ -38,8 +37,9 @@ from .helpers import (
 
 def load_unicode_test_data():
     """Loading unicode test dataset from a csv file in the repo"""
-    df = pd.read_csv(os.path.join(DATA_FOLDER, 'unicode_utf8_unixnl_test.csv'),
-                     encoding='utf-8')
+    data = get_example_data(
+        'unicode_utf8_unixnl_test.csv', is_gzip=False, make_bytes=True)
+    df = pd.read_csv(data, encoding='utf-8')
     # generate date/numeric data
     df['dttm'] = datetime.datetime.now().date()
     df['value'] = [random.randint(1, 100) for _ in range(len(df))]