damccorm commented on code in PR #27430:
URL: https://github.com/apache/beam/pull/27430#discussion_r1258901267
##########
sdks/python/apache_beam/examples/ml_transform/ml_transform_basic.py:
##########
@@ -61,49 +61,75 @@ def parse_args():
return parser.parse_known_args()
-def run(args):
- data = [
- dict(x=["Let's", "go", "to", "the", "park"]),
- dict(x=["I", "enjoy", "going", "to", "the", "park"]),
- dict(x=["I", "enjoy", "reading", "books"]),
- dict(x=["Beam", "can", "be", "fun"]),
- dict(x=["The", "weather", "is", "really", "nice", "today"]),
- dict(x=["I", "love", "to", "go", "to", "the", "park"]),
- dict(x=["I", "love", "to", "read", "books"]),
- dict(x=["I", "love", "to", "program"]),
- ]
-
+def preprocess_data_for_ml_training(train_data, artifact_mode, args):
with beam.Pipeline() as p:
- input_data = p | beam.Create(data)
-
- # arfifacts produce mode.
- input_data |= (
- 'MLTransform' >> MLTransform(
+ input_data = (p | "CreateData" >> beam.Create(train_data))
+ transformed_data_pcoll = (
+ input_data
+ | 'MLTransform' >> MLTransform(
artifact_location=args.artifact_location,
- artifact_mode=ArtifactMode.PRODUCE,
+ artifact_mode=artifact_mode,
).with_transform(ComputeAndApplyVocabulary(
columns=['x'])).with_transform(TFIDF(columns=['x'])))
- # _ = input_data | beam.Map(logging.info)
+ _ = transformed_data_pcoll | beam.Map(logging.info)
+ return transformed_data_pcoll
Review Comment:
Same question in other function
##########
sdks/python/apache_beam/examples/ml_transform/ml_transform_basic.py:
##########
@@ -61,49 +61,75 @@ def parse_args():
return parser.parse_known_args()
-def run(args):
- data = [
- dict(x=["Let's", "go", "to", "the", "park"]),
- dict(x=["I", "enjoy", "going", "to", "the", "park"]),
- dict(x=["I", "enjoy", "reading", "books"]),
- dict(x=["Beam", "can", "be", "fun"]),
- dict(x=["The", "weather", "is", "really", "nice", "today"]),
- dict(x=["I", "love", "to", "go", "to", "the", "park"]),
- dict(x=["I", "love", "to", "read", "books"]),
- dict(x=["I", "love", "to", "program"]),
- ]
-
+def preprocess_data_for_ml_training(train_data, artifact_mode, args):
with beam.Pipeline() as p:
- input_data = p | beam.Create(data)
-
- # arfifacts produce mode.
- input_data |= (
- 'MLTransform' >> MLTransform(
+ input_data = (p | "CreateData" >> beam.Create(train_data))
+ transformed_data_pcoll = (
+ input_data
+ | 'MLTransform' >> MLTransform(
artifact_location=args.artifact_location,
- artifact_mode=ArtifactMode.PRODUCE,
+ artifact_mode=artifact_mode,
).with_transform(ComputeAndApplyVocabulary(
columns=['x'])).with_transform(TFIDF(columns=['x'])))
- # _ = input_data | beam.Map(logging.info)
+ _ = transformed_data_pcoll | beam.Map(logging.info)
+ return transformed_data_pcoll
Review Comment:
Why do we need to return here?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
