Clean up returned Codegen code #1371
Conversation
Codecov Report
@@ Coverage Diff @@
## main #1371 +/- ##
==========================================
- Coverage 99.95% 99.95% -0.00%
==========================================
Files 213 213
Lines 13938 13936 -2
==========================================
- Hits 13931 13929 -2
Misses 7 7
Continue to review full report at Codecov.
|
| @@ -170,9 +170,6 @@ def generate_pipeline_code(element): | |||
| raise ValueError("Element must be a pipeline instance, received {}".format(type(element))) | |||
|
|
|||
| component_graph_string = ', '.join([com.__class__.__name__ if com.__class__ not in all_components() else "'{}'".format(com.name) for com in element.component_graph]) | |||
| import_strings = [com.__class__.__name__ for com in element.component_graph if com.__class__ in all_components()] | |||
There was a problem hiding this comment.
Removed this because for EvalML components, we don't need to import them in order to reference and use them in the component graph of a pipeline
bchen1116
requested review from
dsherry,
freddyaboulton,
christopherbunn,
eccabay and
jeremyliweishih
and removed request for
dsherry
|
|
bchen1116
force-pushed
the
bc_fix_codegen
branch
from
e287dcf
to
d5b6181
Compare
| @@ -189,15 +188,15 @@ def generate_pipeline_code(element): | |||
| pipeline_string = "\t" + "\n\t".join(pipeline_list) + "\n" if len(pipeline_list) else "" | |||
| # create the base string for the pipeline | |||
| base_string = "\nclass {0}({1}):\n" \ | |||
| "\tcomponent_graph = [{2}]\n" \ | |||
| "\tcomponent_graph = [\n\t\t{2}\n\t]\n" \ | |||
| '\nparameters = {\n\t"Imputer": {\n\t\t"categorical_impute_strategy": "most_frequent",\n\t\t"numeric_impute_strategy": "mean",\n\t\t"categorical_fill_value": None,\n\t\t"numeric_fill_value": None\n\t},' \ | ||
| '\n\t"Random Forest Classifier": {\n\t\t"n_estimators": 100,\n\t\t"max_depth": 6,\n\t\t"n_jobs": -1\n\t}\n}\n' \ | ||
| 'pipeline = MockBinaryPipeline(parameters)' | ||
| pipeline = generate_pipeline_code(mock_binary_pipeline) |
|
Arghh codecov!!! You only touched files with 100% coverage. I have no idea why codecov would complain about this... will look further. Merging! |
|
Oops when I merged somehow my browser filled in the commit text as "Fix CLA" 🤦 |
|
@dsherry Codecov is failing because the total number of lines in evalml have decreased by 2, so the percentage overall decreased. I'm running into the same thing in #1388 😅 |
fix #1375
Added formatting for code generation (both for param dictionary and component graph). Remove unnecessary import code from
generate_code_pipelines