zero323 commented on a change in pull request #29180:
URL: https://github.com/apache/spark/pull/29180#discussion_r495487007
##########
File path: dev/lint-python
##########
@@ -122,6 +123,31 @@ function pycodestyle_test {
fi
}
+function mypy_test {
+ local MYPY_REPORT=
+ local MYPY_STATUS=
+
+ if ! hash "$MYPY_BUILD" 2> /dev/null; then
+ echo "The mypy command was not found."
+ echo "mypy checks failed."
+ exit 1
+ fi
Review comment:
I think we should make this optional for now, i.e.
```bash
if ! hash "$MYPY_BUILD" 2> /dev/null; then
echo "The $MYPY_BUILD command was not found. Skipping MyPy test
check build for now."
echo
return
fi
```
##########
File path: python/pyspark/serializers.pyi
##########
@@ -0,0 +1,128 @@
+#
Review comment:
This has been already ported.
##########
File path: dev/lint-python
##########
@@ -122,6 +123,31 @@ function pycodestyle_test {
fi
}
+function mypy_test {
+ local MYPY_REPORT=
+ local MYPY_STATUS=
+
+ if ! hash "$MYPY_BUILD" 2> /dev/null; then
+ echo "The mypy command was not found."
+ echo "mypy checks failed."
+ exit 1
+ fi
+
+ echo "starting $MYPY_BUILD test..."
+ MYPY_REPORT=$( ($MYPY_BUILD --ignore-missing-imports --config-file
python/mypy.ini python/) 2>&1)
Review comment:
I wouldn't use `--ignore-missing-imports` here. It is an atomic option.
Instead I would use either specific excludes (as we have right now in
mypy.ini) or targeted `type: ignore` (as in some test modules and such).
##########
File path: python/pyspark/accumulators.pyi
##########
@@ -0,0 +1,76 @@
+#
Review comment:
This has been already ported.
##########
File path: python/pyspark/broadcast.pyi
##########
@@ -0,0 +1,51 @@
+#
Review comment:
This has been already ported.
##########
File path: python/mypy.ini
##########
@@ -0,0 +1,20 @@
+[mypy]
+python_version = 3.6
+
+[mypy-pyspark.cloudpickle.*]
+ignore_errors = True
+
+[mypy-pyspark.streaming.tests.*]
+ignore_errors = True
+
+[mypy-pyspark.sql.tests.*]
+ignore_errors = True
+
+[mypy-pyspark.testing.*]
+ignore_errors = True
+
+[mypy-pyspark.tests.*]
+ignore_errors = True
Review comment:
I am strongly against excluding tests and testing here.
These already type check, with only a bunch of `ignores` and it is an
additional coverage.
##########
File path: dev/lint-python
##########
@@ -246,6 +272,7 @@ PYTHON_SOURCE="$(find . -name "*.py")"
compile_python_test "$PYTHON_SOURCE"
pycodestyle_test "$PYTHON_SOURCE"
flake8_test
+mypy_test
Review comment:
Personally I'd consider having type checks in a separate script, as we
might want to extend things over time (I plan to add tests on examples, same as
in
[`pyspark-stubs`](https://github.com/zero323/pyspark-stubs/blob/d317c27083549b6727542d825e642ad91f275f9d/.travis.yml#L18)
and thinking about migrating original data driven tests).
But it is something we can discuss later.
##########
File path: python/pyspark/mllib/linalg/__init__.py
##########
@@ -469,7 +469,7 @@ def __getattr__(self, item):
def __neg__(self):
return DenseVector(-self.array)
- def _delegate(op):
+ def _delegate(op: str): # type: ignore
Review comment:
I believe we can skip this one.
##########
File path: python/pyspark/ml/base.py
##########
@@ -309,7 +309,8 @@ def setPredictionCol(self, value):
"""
return self._set(predictionCol=value)
- @abstractproperty
+ @property # type: ignore
Review comment:
These ignores are not strictly required now, but I think it is OK to
keep them, as there is not real harm here and they're likely to prove required
in the future.
##########
File path: python/pyspark/sql/session.py
##########
@@ -76,7 +76,7 @@ class Builder(object):
"""
_lock = RLock()
- _options = {}
+ _options = {} # type: ignore
Review comment:
I believe we can skip this one.
##########
File path: python/pyspark/sql/types.pyi
##########
@@ -0,0 +1,238 @@
+#
Review comment:
This has been already ported.
##########
File path: python/setup.py
##########
@@ -28,7 +28,8 @@
print("Failed to load PySpark version file for packaging. You must be in
Spark's python dir.",
file=sys.stderr)
sys.exit(-1)
-VERSION = __version__ # noqa
+
+VERSION = __version__ # type: ignore # noqa
Review comment:
I don't think it is required
##########
File path: python/pyspark/util.pyi
##########
@@ -0,0 +1,33 @@
+#
Review comment:
This has been already ported, though I am not sure if we're going to
consider it a part of the API.
##########
File path: python/pyspark/storagelevel.py
##########
@@ -19,6 +19,14 @@
class StorageLevel(object):
+ DISK_ONLY: 'StorageLevel'
Review comment:
This is already covered with stubs and I think it is better to avoid
annotations in `.py` unless they're strictly needed.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
