chore: Support Python 3.10 and bump pandas 1.4 and pyarrow 6 (apache#…

…21002)

* Bump pandas 1.4 and pyarrow 6

* Use engine="pyarrow" for pd.read_csv()

* Refactoring

* Refactoring

* Refactoring

* Use bytes in pd.read_json()

* Fix test_contribution

* Fix pandas issue when 'arrays' are empty but 'names' contain values

* fix: ValueError: For argument "ascending" expected type bool, received type NoneType.

* Remove engine="pyarrow" and convert bytes to string

* make copy of selected df to fix regression

* Simplify pd.read_json() and pd.read_csv() for example data

Co-authored-by: Ville Brofeldt <[email protected]>

CONTRIBUTING.md

@@ -420,7 +420,7 @@ Commits to `master` trigger a rebuild and redeploy of the documentation site. Su
 Make sure your machine meets the [OS dependencies](https://superset.apache.org/docs/installation/installing-superset-from-scratch#os-dependencies) before following these steps.
 You also need to install MySQL or [MariaDB](https://mariadb.com/downloads).
 
-Ensure that you are using Python version 3.8 or 3.9, then proceed with:
+Ensure that you are using Python version 3.8, 3.9 or 3.10, then proceed with:
 
 ```bash
 # Create a virtual environment and activate it (recommended)

Makefile

@@ -15,8 +15,8 @@
 # limitations under the License.
 #
 
-# Python version installed; we need 3.8-3.9
-PYTHON=`command -v python3.9 || command -v python3.8`
+# Python version installed; we need 3.8-3.10
+PYTHON=`command -v python3.10 || command -v python3.9 || command -v python3.8`
 
 .PHONY: install superset venv pre-commit
 
@@ -70,7 +70,7 @@ update-js:
 
 venv:
 	# Create a virtual environment and activate it (recommended)
-	if ! [ -x "${PYTHON}" ]; then echo "You need Python 3.8 or 3.9 installed"; exit 1; fi
+	if ! [ -x "${PYTHON}" ]; then echo "You need Python 3.8, 3.9 or 3.10 installed"; exit 1; fi
 	test -d venv || ${PYTHON} -m venv venv # setup a python3 virtualenv
 	. venv/bin/activate
 

UPDATING.md

@@ -26,6 +26,7 @@ assists people when migrating to a new version.
 
 - [20606](https://github.com/apache/superset/pull/20606): When user clicks on chart title or "Edit chart" button in Dashboard page, Explore opens in the same tab. Clicking while holding cmd/ctrl opens Explore in a new tab. To bring back the old behaviour (always opening Explore in a new tab), flip feature flag `DASHBOARD_EDIT_CHART_IN_NEW_TAB` to `True`.
 - [20799](https://github.com/apache/superset/pull/20799): Presto and Trino engine will now display tracking URL for running queries in SQL Lab. If for some reason you don't want to show the tracking URL (for example, when your data warehouse hasn't enable access for to Presto or Trino UI), update `TRACKING_URL_TRANSFORMER` in `config.py` to return `None`.
+- [21002](https://github.com/apache/superset/pull/21002): Support Python 3.10 and bump pandas 1.4 and pyarrow 6.
 
 ### Breaking Changes
 

requirements/base.txt

@@ -185,7 +185,7 @@ packaging==21.3
     # via
     #   bleach
     #   deprecation
-pandas==1.3.4
+pandas==1.4.3
     # via apache-superset
 parsedatetime==2.6
     # via apache-superset
@@ -197,7 +197,7 @@ prison==0.2.1
     # via flask-appbuilder
 prompt-toolkit==3.0.28
     # via click-repl
-pyarrow==5.0.0
+pyarrow==6.0.1
     # via apache-superset
 pycparser==2.20
     # via cffi

setup.py

@@ -100,15 +100,15 @@ def get_git_sha() -> str:
         "markdown>=3.0",
         "msgpack>=1.0.0, <1.1",
         "numpy==1.22.1",
-        "pandas>=1.3.0, <1.4",
+        "pandas>=1.4.3, <1.5",
         "parsedatetime",
         "pgsanity",
         "polyline",
         "pyparsing>=3.0.6, <4",
         "python-dateutil",
         "python-dotenv",
         "python-geohash",
-        "pyarrow>=5.0.0, <6.0",
+        "pyarrow>=6.0.1, <7",
         "pyyaml>=5.4",
         "PyJWT>=2.4.0, <3.0",
         "redis",
@@ -183,5 +183,6 @@ def get_git_sha() -> str:
     classifiers=[
         "Programming Language :: Python :: 3.8",
         "Programming Language :: Python :: 3.9",
+        "Programming Language :: Python :: 3.10",
     ],
 )

superset/examples/bart_lines.py

@@ -23,7 +23,7 @@
 from superset import db
 
 from ..utils.database import get_example_database
-from .helpers import get_example_data, get_table_connector_registry
+from .helpers import get_example_url, get_table_connector_registry
 
 
 def load_bart_lines(only_metadata: bool = False, force: bool = False) -> None:
@@ -34,8 +34,8 @@ def load_bart_lines(only_metadata: bool = False, force: bool = False) -> None:
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        content = get_example_data("bart-lines.json.gz")
-        df = pd.read_json(content, encoding="latin-1")
+        url = get_example_url("bart-lines.json.gz")
+        df = pd.read_json(url, encoding="latin-1", compression="gzip")
         df["path_json"] = df.path.map(json.dumps)
         df["polyline"] = df.path.map(polyline.encode)
         del df["path"]

superset/examples/birth_names.py

@@ -33,7 +33,7 @@
 
 from ..utils.database import get_example_database
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_slice_json,
     get_table_connector_registry,
     merge_slice,
@@ -66,7 +66,8 @@ def gen_filter(
 
 
 def load_data(tbl_name: str, database: Database, sample: bool = False) -> None:
-    pdf = pd.read_json(get_example_data("birth_names2.json.gz"))
+    url = get_example_url("birth_names2.json.gz")
+    pdf = pd.read_json(url, compression="gzip")
     # TODO(bkyryliuk): move load examples data into the pytest fixture
     if database.backend == "presto":
         pdf.ds = pd.to_datetime(pdf.ds, unit="ms")

superset/examples/country_map.py

@@ -27,7 +27,7 @@
 from superset.utils.core import DatasourceType
 
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_slice_json,
     get_table_connector_registry,
     merge_slice,
@@ -44,10 +44,8 @@ def load_country_map_data(only_metadata: bool = False, force: bool = False) -> N
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        csv_bytes = get_example_data(
-            "birth_france_data_for_country_map.csv", is_gzip=False, make_bytes=True
-        )
-        data = pd.read_csv(csv_bytes, encoding="utf-8")
+        url = get_example_url("birth_france_data_for_country_map.csv")
+        data = pd.read_csv(url, encoding="utf-8")
         data["dttm"] = datetime.datetime.now().date()
         data.to_sql(
             tbl_name,

superset/examples/energy.py

@@ -28,7 +28,7 @@
 from superset.utils.core import DatasourceType
 
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_table_connector_registry,
     merge_slice,
     misc_dash_slices,
@@ -46,8 +46,8 @@ def load_energy(
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("energy.json.gz")
-        pdf = pd.read_json(data)
+        url = get_example_url("energy.json.gz")
+        pdf = pd.read_json(url, compression="gzip")
         pdf = pdf.head(100) if sample else pdf
         pdf.to_sql(
             tbl_name,

superset/examples/flights.py

@@ -20,7 +20,7 @@
 import superset.utils.database as database_utils
 from superset import db
 
-from .helpers import get_example_data, get_table_connector_registry
+from .helpers import get_example_url, get_table_connector_registry
 
 
 def load_flights(only_metadata: bool = False, force: bool = False) -> None:
@@ -32,12 +32,12 @@ def load_flights(only_metadata: bool = False, force: bool = False) -> None:
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("flight_data.csv.gz", make_bytes=True)
-        pdf = pd.read_csv(data, encoding="latin-1")
+        flight_data_url = get_example_url("flight_data.csv.gz")
+        pdf = pd.read_csv(flight_data_url, encoding="latin-1", compression="gzip")
 
         # Loading airports info to join and get lat/long
-        airports_bytes = get_example_data("airports.csv.gz", make_bytes=True)
-        airports = pd.read_csv(airports_bytes, encoding="latin-1")
+        airports_url = get_example_url("airports.csv.gz")
+        airports = pd.read_csv(airports_url, encoding="latin-1", compression="gzip")
         airports = airports.set_index("IATA_CODE")
 
         pdf[  # pylint: disable=unsupported-assignment-operation,useless-suppression

superset/examples/helpers.py

@@ -17,10 +17,7 @@
 """Loads datasets, dashboards and slices in a new superset instance"""
 import json
 import os
-import zlib
-from io import BytesIO
 from typing import Any, Dict, List, Set
-from urllib import request
 
 from superset import app, db
 from superset.connectors.sqla.models import SqlaTable
@@ -73,14 +70,5 @@ def get_slice_json(defaults: Dict[Any, Any], **kwargs: Any) -> str:
     return json.dumps(defaults_copy, indent=4, sort_keys=True)
 
 
-def get_example_data(
-    filepath: str, is_gzip: bool = True, make_bytes: bool = False
-) -> BytesIO:
-    content = request.urlopen(  # pylint: disable=consider-using-with
-        f"{BASE_URL}{filepath}?raw=true"
-    ).read()
-    if is_gzip:
-        content = zlib.decompress(content, zlib.MAX_WBITS | 16)
-    if make_bytes:
-        content = BytesIO(content)
-    return content
+def get_example_url(filepath: str) -> str:
+    return f"{BASE_URL}{filepath}?raw=true"

superset/examples/long_lat.py

@@ -27,7 +27,7 @@
 from superset.utils.core import DatasourceType
 
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_slice_json,
     get_table_connector_registry,
     merge_slice,
@@ -44,8 +44,8 @@ def load_long_lat_data(only_metadata: bool = False, force: bool = False) -> None
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("san_francisco.csv.gz", make_bytes=True)
-        pdf = pd.read_csv(data, encoding="utf-8")
+        url = get_example_url("san_francisco.csv.gz")
+        pdf = pd.read_csv(url, encoding="utf-8", compression="gzip")
         start = datetime.datetime.now().replace(
             hour=0, minute=0, second=0, microsecond=0
         )

superset/examples/multiformat_time_series.py

@@ -25,7 +25,7 @@
 
 from ..utils.database import get_example_database
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_slice_json,
     get_table_connector_registry,
     merge_slice,
@@ -44,8 +44,8 @@ def load_multiformat_time_series(  # pylint: disable=too-many-locals
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("multiformat_time_series.json.gz")
-        pdf = pd.read_json(data)
+        url = get_example_url("multiformat_time_series.json.gz")
+        pdf = pd.read_json(url, compression="gzip")
         # TODO(bkyryliuk): move load examples data into the pytest fixture
         if database.backend == "presto":
             pdf.ds = pd.to_datetime(pdf.ds, unit="s")

superset/examples/paris.py

@@ -22,7 +22,7 @@
 import superset.utils.database as database_utils
 from superset import db
 
-from .helpers import get_example_data, get_table_connector_registry
+from .helpers import get_example_url, get_table_connector_registry
 
 
 def load_paris_iris_geojson(only_metadata: bool = False, force: bool = False) -> None:
@@ -33,8 +33,8 @@ def load_paris_iris_geojson(only_metadata: bool = False, force: bool = False) ->
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("paris_iris.json.gz")
-        df = pd.read_json(data)
+        url = get_example_url("paris_iris.json.gz")
+        df = pd.read_json(url, compression="gzip")
         df["features"] = df.features.map(json.dumps)
 
         df.to_sql(

superset/examples/random_time_series.py

@@ -24,7 +24,7 @@
 from superset.utils.core import DatasourceType
 
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_slice_json,
     get_table_connector_registry,
     merge_slice,
@@ -42,8 +42,8 @@ def load_random_time_series_data(
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("random_time_series.json.gz")
-        pdf = pd.read_json(data)
+        url = get_example_url("random_time_series.json.gz")
+        pdf = pd.read_json(url, compression="gzip")
         if database.backend == "presto":
             pdf.ds = pd.to_datetime(pdf.ds, unit="s")
             pdf.ds = pdf.ds.dt.strftime("%Y-%m-%d %H:%M%:%S")

superset/examples/sf_population_polygons.py

@@ -22,7 +22,7 @@
 import superset.utils.database as database_utils
 from superset import db
 
-from .helpers import get_example_data, get_table_connector_registry
+from .helpers import get_example_url, get_table_connector_registry
 
 
 def load_sf_population_polygons(
@@ -35,8 +35,8 @@ def load_sf_population_polygons(
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("sf_population.json.gz")
-        df = pd.read_json(data)
+        url = get_example_url("sf_population.json.gz")
+        df = pd.read_json(url, compression="gzip")
         df["contour"] = df.contour.map(json.dumps)
 
         df.to_sql(

superset/examples/world_bank.py

@@ -33,7 +33,7 @@
 
 from ..connectors.base.models import BaseDatasource
 from .helpers import (
-    get_example_data,
+    get_example_url,
     get_examples_folder,
     get_slice_json,
     get_table_connector_registry,
@@ -56,8 +56,8 @@ def load_world_bank_health_n_pop(  # pylint: disable=too-many-locals, too-many-s
     table_exists = database.has_table_by_name(tbl_name)
 
     if not only_metadata and (not table_exists or force):
-        data = get_example_data("countries.json.gz")
-        pdf = pd.read_json(data)
+        url = get_example_url("countries.json.gz")
+        pdf = pd.read_json(url, compression="gzip")
         pdf.columns = [col.replace(".", "_") for col in pdf.columns]
         if database.backend == "presto":
             pdf.year = pd.to_datetime(pdf.year)

superset/result_set.py

@@ -161,6 +161,9 @@ def __init__(  # pylint: disable=too-many-locals
                         except Exception as ex:  # pylint: disable=broad-except
                             logger.exception(ex)
 
+        if not pa_data:
+            column_names = []
+
         self.table = pa.Table.from_arrays(pa_data, names=column_names)
         self._type_dict: Dict[str, Any] = {}
         try:

superset/utils/pandas_postprocessing/contribution.py

@@ -49,6 +49,9 @@ def contribution(
     """
     contribution_df = df.copy()
     numeric_df = contribution_df.select_dtypes(include=["number", Decimal])
+    # TODO: copy needed due to following regression in 1.4, remove if not needed:
+    # https://github.com/pandas-dev/pandas/issues/48090
+    numeric_df = numeric_df.copy()
     numeric_df.fillna(0, inplace=True)
     # verify column selections
     if columns:

superset/viz.py

@@ -2172,14 +2172,14 @@ def get_data(self, df: pd.DataFrame) -> VizData:
             if df is not None and not df.empty:
                 if metric:
                     df = df.sort_values(
-                        utils.get_metric_name(metric), ascending=flt.get("asc")
+                        utils.get_metric_name(metric), ascending=flt.get("asc", False)
                     )
                     data[col] = [
                         {"id": row[0], "text": row[0], "metric": row[1]}
                         for row in df.itertuples(index=False)
                     ]
                 else:
-                    df = df.sort_values(col, ascending=flt.get("asc"))
+                    df = df.sort_values(col, ascending=flt.get("asc", False))
                     data[col] = [
                         {"id": row[0], "text": row[0]}
                         for row in df.itertuples(index=False)