Merge pull request #35 from vincent-laurent/main

[DEV] Improve performance

Author: vincent-laurent

.coveragerc

@@ -21,5 +21,6 @@ exclude_lines =
     # Don't complain if non-runnable code isn't run:
     if 0:
     if __name__ == .__main__.:
+    break
 [run]
     omit = *tests*

.github/workflows/pytest.yml

@@ -1,6 +1,7 @@
 name: Pytest
 permissions:
   contents: write
+  pages: write
 
 on:
   push:

.static/logoEurobiosMewsLabs.png

@@ -1,4 +1,4 @@
-<img src="./.static/logo.png" width="200"/>
+<img src=".static/logo.png" width="200"/>
 
 ### _Project for Automated Learning MAchine_ 
 

README.md

@@ -9,7 +9,7 @@
 import sys
 
 sys.path.insert(0, os.path.abspath('../'))
-sys.path.insert(0, os.path.abspath('.'))
+sys.path.insert(0, os.path.abspath("../../"))
 
 project = 'palma'
 copyright = '2024, Eurobios-Mews-Labs'
@@ -38,7 +38,8 @@
 # -- Options for HTML output -------------------------------------------------
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#options-for-html-output
 
-html_theme = 'sphinx_book_theme'
+html_theme = "sphinx_book_theme"
+# html_style = '_static/palma.css'
 html_static_path = ['../../.static']
 source_suffix = {
     '.rst': 'restructuredtext',

docs/source/conf.py

@@ -2,7 +2,8 @@
    sphinx-quickstart on Thu Jan  4 17:25:24 2024.
    You can adapt this file completely to your liking, but it should at least
    contain the root `toctree` directive.
-.. image:: ../../.static/logo.png
+
+.. image:: https://raw.githubusercontent.com/eurobios-mews-labs/palma/main/.static/logo.png
    :width: 200
 
 \

docs/source/index.rst

@@ -16,6 +16,8 @@
 from flaml import AutoML
 from sklearn import base
 
+from palma.base.splitting_strategy import ValidationStrategy
+
 try:
     from autosklearn.classification import AutoSklearnClassifier
     from autosklearn.regression import AutoSklearnRegressor
@@ -29,7 +31,8 @@ def __init__(self, engine_parameters: dict) -> None:
         self.__engine_parameters = engine_parameters
 
     @abstractmethod
-    def optimize(self, X: pd.DataFrame, y: pd.Series, splitter=None
+    def optimize(self, X: pd.DataFrame, y: pd.Series,
+                 splitter: "ValidationStrategy" = None
                  ) -> None:
         ...
 
@@ -87,13 +90,13 @@ def optimize(self, X: pd.DataFrame, y: pd.Series, splitter=None) -> None:
     @property
     def optimizer(self) -> Union[
         'AutoSklearnClassifier',
-            'AutoSklearnRegressor']:
+        'AutoSklearnRegressor']:
         return self.__optimizer
 
     @property
     def estimator_(self) -> Union[
         'AutoSklearnClassifier',
-            'AutoSklearnRegressor']:
+        'AutoSklearnRegressor']:
         return self.__optimizer.get_models_with_weights()
 
     @property
@@ -111,14 +114,20 @@ def __init__(self, problem: str, engine_parameters: dict) -> None:
             )
         engine_parameters["task"] = problem
 
-    def optimize(self, X: pd.DataFrame, y: pd.DataFrame, splitter=None
+    def optimize(self, X: pd.DataFrame, y: pd.DataFrame,
+                 splitter: ValidationStrategy = None
                  ) -> None:
+        split_type = None if splitter is None else splitter.splitter
+        groups = None if splitter is None else splitter.groups
+        groups = groups if groups is None else groups[splitter.train_index]
+
         self.allowing_splitter(splitter)
         self.__optimizer = AutoML()
         self.__optimizer.fit(
             X_train=pd.DataFrame(X.values, index=range(len(X))),
             y_train=pd.Series(y.values, index=range(len(X))),
-            split_type=splitter, mlflow_logging=False,
+            split_type=split_type, groups=groups,
+            mlflow_logging=False,
             **self.engine_parameters
         )
 

examples/classification.ipynb

@@ -47,14 +47,20 @@ def start(self, project: "Project"):
         self.engine_.optimize(
             project.X.iloc[project.validation_strategy.train_index],
             project.y.iloc[project.validation_strategy.train_index],
-            splitter=project.validation_strategy.splitter
+            splitter=project.validation_strategy,
         )
         self.best_model_ = self.engine_.estimator_
         logging.basicConfig(level=logging.DEBUG)
 
         logger.logger.log_artifact(
             self.engine_.estimator_,
             self.__run_id)
+        try:
+            logger.logger.log_metrics(
+                {"best_estimator": str(self.best_model_)}, 'model_selection'
+            )
+        except:
+            pass
 
     @property
     def run_id(self) -> str:

palma/base/engine.py

@@ -90,7 +90,6 @@ def start(
         from palma.utils.checker import ProjectPlanChecker
         from palma import logger
         self.__validation_strategy = ValidationStrategy(splitter)
-        self.__base_index = list(range(len(X)))
         self.__X, self.__y = self.__validation_strategy(
             X=X, y=y, X_test=X_test, y_test=y_test,
             groups=groups)
@@ -108,10 +107,6 @@ def __call_components(self, object_: "Project") -> None:
         for _, component in self.components.items():
             component(object_)
 
-    @property
-    def base_index(self) -> List[int]:
-        return self.__base_index
-
     @property
     def components(self) -> dict:
         return self.__components