DOC Update plot_column_transformer to notebook style #17028
Conversation
| @@ -39,95 +33,130 @@ | |||
| from sklearn.compose import ColumnTransformer | |||
| from sklearn.svm import LinearSVC | |||
|
|
|||
| ############################################################################## | |||
There was a problem hiding this comment.
Hi @lucyleeow there is a discussion about changing this syntax starting from 0.23 (#17068). Maybe, keep an eye on it and be ready to change this one if it is merged. Thanks!
There was a problem hiding this comment.
No problem, easy change!
|
Looks good. I was a bit confused that the results change but then saw you also change the newsgroups. Maybe @ogrisel wants to have a look, he might have written the original? |
|
Thanks @amueller!
This was because I wanted to print an example of a sample but the posts in 'alt.atheism' and 'talk.religion.misc' were a little controversial... |
| ############################################################################## | ||
| # We will also create a transformer that extracts the | ||
| # length of the text and the number of sentences. | ||
|
|
There was a problem hiding this comment.
Nit: do we need two new lines here?
There was a problem hiding this comment.
I remember why I put 2 lines now. Lint complains if it's only 1.
| return [{'length': len(text), | ||
| 'num_sentences': text.count('.')} | ||
| for text in posts] | ||
|
|
There was a problem hiding this comment.
Nit: Do we need two new lines here?
| for text in posts] | ||
|
|
||
|
|
||
| TextStats = FunctionTransformer(text_stats) |
There was a problem hiding this comment.
Since this is an instance:
| TextStats = FunctionTransformer(text_stats) | |
| text_stats_transformer = FunctionTransformer(text_stats) |
| ('body_stats', Pipeline([ | ||
| ('stats', TextStats()), # returns a list of dicts | ||
| ('stats', TextStats), # returns a list of dicts |
There was a problem hiding this comment.
| ('stats', TextStats), # returns a list of dicts | |
| ('stats', text_stats_transformer), # returns a list of dicts |
|
|
||
| SubjectBodyExtractor = FunctionTransformer(subject_body_extractor) |
There was a problem hiding this comment.
Since this is an instance:
| SubjectBodyExtractor = FunctionTransformer(subject_body_extractor) | |
| subject_body_extractor = FunctionTransformer(subject_body_extractor) |
There was a problem hiding this comment.
subject_body_transformer so it's not the same name as the function? Not that this matters, though if we make it the same name we should also do this for text_stats_transformer to be consistent.
|
|
||
| pipeline = Pipeline([ | ||
| # Extract subject & body | ||
| ('subjectbody', SubjectBodyExtractor), |
There was a problem hiding this comment.
| ('subjectbody', SubjectBodyExtractor), | |
| ('subjectbody', subject_body_extractor), |
| return_X_y=True) | ||
| ############################################################################## | ||
| # Finally, we fit our pipeline on the training data and use it to predict | ||
| # topics for ``X_test``. Performance metrics of our pipeline is then printed. |
There was a problem hiding this comment.
| # topics for ``X_test``. Performance metrics of our pipeline is then printed. | |
| # topics for ``X_test``. Performance metrics of our pipeline are then printed. |
Reference Issues/PRs
None
What does this implement/fix? Explain your changes.
plot_column_transformer.pyto use notebook styleFunctionTransformeras stateless transformers defined hereAny other comments?