medpipe 0.1.2__tar.gz → 0.2.2.dev0__tar.gz

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: medpipe
-Version: 0.1.2
+Version: 0.2.2.dev0
 Summary: A package for clinical AI pipeline development
 Author-email: Mathias Roesler <mathias.roesler@auckland.co.nz>
 License-Expression: Apache-2.0
@@ -22,6 +22,7 @@ Requires-Dist: pytest==9.0.1
 Requires-Dist: scikit_learn==1.8.0
 Requires-Dist: scipy==1.17.0
 Requires-Dist: pyarrow==23.0.1
+Requires-Dist: lightgbm==4.6.0
 Dynamic: license-file
 
 # medpipe
@@ -45,7 +46,11 @@ The **medpipe** package is a layer to help create AI models for clinical applica
 ___
 ## Installation
 
-To install **medpipe** clone the GitHub repository and install the package with pip: 
+To install **medpipe** from the PyPI use pip:
+```
+$ pip install medpipe
+```
+or clone the GitHub repository and install the package with pip: 
 ```
 $ git clone git@github.com:Surgical-Recovery-and-Safety-Lab/medpipe.git
 $ cd medpipe

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/README.md

@@ -19,7 +19,11 @@ The **medpipe** package is a layer to help create AI models for clinical applica
 ___
 ## Installation
 
-To install **medpipe** clone the GitHub repository and install the package with pip: 
+To install **medpipe** from the PyPI use pip:
+```
+$ pip install medpipe
+```
+or clone the GitHub repository and install the package with pip: 
 ```
 $ git clone git@github.com:Surgical-Recovery-and-Safety-Lab/medpipe.git
 $ cd medpipe

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/pyproject.toml

@@ -4,7 +4,7 @@ requires = [ "setuptools>=80.9" ]
 
 [project]
 name = "medpipe"
-version = "0.1.2"
+version = "0.2.2.dev0"
 description = "A package for clinical AI pipeline development"
 readme = "README.md"
 license = "Apache-2.0"

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/requirements.txt

@@ -7,3 +7,4 @@ pytest==9.0.1
 scikit_learn==1.8.0
 scipy==1.17.0
 pyarrow==23.0.1
+lightgbm==4.6.0

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/src/medpipe/models/core.py

@@ -14,6 +14,7 @@ Functions:
 
 import joblib
 import numpy as np
+from lightgbm import LGBMClassifier
 from sklearn.ensemble import HistGradientBoostingClassifier
 from sklearn.isotonic import IsotonicRegression
 from sklearn.linear_model import LogisticRegression
@@ -36,9 +37,10 @@ def create_model(
 
     Parameters
     ----------
-    model_type : {"hgb-c", "logistic", "isotonic"}
+    model_type : {"hgb-c", "lgbm-c", "logistic", "isotonic"}
         Type of model to create.
             hgb-c: histogram gradient boosting classifier.
+            lgbm-c: LightGBM classifier.
             logistic: logistic regression.
             isotonic: isotonic regression.
     quiet : bool, default: False
@@ -48,7 +50,7 @@ def create_model(
 
     Returns
     -------
-    model : HistGradBoostingClassifier
+    model : HistGradBoostingClassifier, LGBMClassifier
             LogisticRegression, IsotonicRegression,
         Created model.
 
@@ -73,6 +75,14 @@ def create_model(
                     SCRIPT_NAME,
                 )
             model = HistGradientBoostingClassifier(**config_params)
+        case "lgbm-c":
+            if not quiet:
+                print_message(
+                    "Creating a LGBMClassifier",
+                    logger,
+                    SCRIPT_NAME,
+                )
+            model = LGBMClassifier(**config_params)
         case "logistic":
             if not quiet:
                 print_message(

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/src/medpipe.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: medpipe
-Version: 0.1.2
+Version: 0.2.2.dev0
 Summary: A package for clinical AI pipeline development
 Author-email: Mathias Roesler <mathias.roesler@auckland.co.nz>
 License-Expression: Apache-2.0
@@ -22,6 +22,7 @@ Requires-Dist: pytest==9.0.1
 Requires-Dist: scikit_learn==1.8.0
 Requires-Dist: scipy==1.17.0
 Requires-Dist: pyarrow==23.0.1
+Requires-Dist: lightgbm==4.6.0
 Dynamic: license-file
 
 # medpipe
@@ -45,7 +46,11 @@ The **medpipe** package is a layer to help create AI models for clinical applica
 ___
 ## Installation
 
-To install **medpipe** clone the GitHub repository and install the package with pip: 
+To install **medpipe** from the PyPI use pip:
+```
+$ pip install medpipe
+```
+or clone the GitHub repository and install the package with pip: 
 ```
 $ git clone git@github.com:Surgical-Recovery-and-Safety-Lab/medpipe.git
 $ cd medpipe

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/src/medpipe.egg-info/requires.txt

@@ -7,3 +7,4 @@ pytest==9.0.1
 scikit_learn==1.8.0
 scipy==1.17.0
 pyarrow==23.0.1
+lightgbm==4.6.0

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/test/test_models_core.py

@@ -9,7 +9,10 @@ Test functions for the models.core module
 import pathlib
 
 import pytest
+from lightgbm import LGBMClassifier
 from sklearn.ensemble import HistGradientBoostingClassifier
+from sklearn.isotonic import IsotonicRegression
+from sklearn.linear_model import LogisticRegression
 
 from medpipe.models.core import create_model
 from medpipe.utils.config import get_configuration, split_version_number
@@ -20,10 +23,17 @@ DATA_DIR = CWD / "test/test_data/"
 
 
 @pytest.mark.parametrize(
-    "model_type, config_file",
-    [("hgb-c", "config/HGBc_config.toml"), ("hgb-c", "")],
+    "model_type, config_file, model_class",
+    [
+        ("hgb-c", "config/HGBc_config.toml", HistGradientBoostingClassifier),
+        ("hgb-c", "", HistGradientBoostingClassifier),
+        ("isotonic", "", IsotonicRegression),
+        ("logistic", "", LogisticRegression),
+        ("lgbm-c", "config/LGBMc_config.toml", LGBMClassifier),
+        ("lgbm-c", "", LGBMClassifier),
+    ],
 )
-def test_create_model_HGBc_success(model_type, config_file):
+def test_create_model_success(model_type, config_file, model_class):
     if config_file:
         general_params = read_toml_configuration(str(DATA_DIR / config_file))
         _, model_version = split_version_number(general_params["version"])
@@ -39,13 +49,7 @@ def test_create_model_HGBc_success(model_type, config_file):
     else:
         model = create_model(model_type, logger=None)
 
-    if model_type == "hgb-c":
-        assert isinstance(model, HistGradientBoostingClassifier)
-
-
-def test_create_model_not_valid_model():
-    with pytest.raises(ValueError):
-        create_model("not_valid_model")
+    assert isinstance(model, model_class)
 
 
 @pytest.mark.parametrize("model_type", [123, [], {}, 1.5, None])
@@ -55,7 +59,7 @@ def test_create_model_invalid_model_type(model_type):
 
 
 # Test invalid configuration parameters for the model
-@pytest.mark.parametrize("model_type", ["hgb-c"])
+@pytest.mark.parametrize("model_type", ["hgb-c", "isotonic", "logistic"])
 def test_create_model_invalid_config(model_type):
     model_config = {"invalid": None}
     # Expecting a failure due to invalid config
@@ -65,20 +69,3 @@ def test_create_model_invalid_config(model_type):
             logger=None,
             **model_config,
         )
-
-
-# Test passing a logger to see if log messages are printed
-@pytest.mark.parametrize("model_type", ["hgb-c"])
-def test_create_model_with_logger(model_type):
-    # Here we'll check if logger prints the expected message to stdout/stderr
-    logger = None  # Use a mock or None for simplicity in this case
-    create_model(model_type, logger=logger)
-
-
-# Test `create_model` when no config file is passed
-@pytest.mark.parametrize("model_type", ["hgb-c"])
-def test_create_model_without_config(model_type):
-    model = create_model(model_type, logger=None)
-
-    if model_type == "hgb-c":
-        assert isinstance(model, HistGradientBoostingClassifier)

{medpipe-0.1.2 → medpipe-0.2.2.dev0}/test/test_pipeline_Pipeline.py

@@ -19,7 +19,7 @@ DATA_DIR = CWD / "test/test_data/"
 
 @pytest.mark.parametrize(
     "config_file",
-    ["config/HGBc_config.toml"],
+    ["config/HGBc_config.toml", "config/LGBMc_config.toml"],
 )
 def test_create_Pipeline_success(config_file):
     general_config = read_toml_configuration(str(DATA_DIR / config_file))
@@ -50,6 +50,30 @@ def test_run_Pipeline_HGBc_single_label(version_number):
     pipeline.run(X_train)
 
 
+@pytest.mark.parametrize(
+    "version_number",
+    [
+        "v0.1.1.2-a.1.0.1",
+        "v0.1.0.2-a.2.1.1",
+        "v0.1.1.2-a.3.2.1",
+        "v0.1.0.2-a.4.0.1",
+        "v0.1.0.2-a.5.1.1",
+        "v0.1.2.2-a.1.1.1",
+        "v0.1.3.2-a.2.1.1",
+    ],
+)
+def test_run_Pipeline_LGBMc_single_label(version_number):
+    config_file = "config/LGBMc_config.toml"
+    data = load_data_from_csv("test/test_data/test_data.csv")
+    general_config = read_toml_configuration(str(DATA_DIR / config_file))
+    general_config["version"] = version_number
+    pipeline = Pipeline(general_config)
+
+    pipeline.preprocessor.fit(data)
+    X_train, _ = pipeline.get_test_data(data)
+    pipeline.run(X_train)
+
+
 @pytest.mark.parametrize(
     "version_number",
     [
@@ -72,3 +96,27 @@ def test_run_Pipeline_HGBc_multi_label(version_number):
     pipeline.preprocessor.fit(data)
     X_train, _ = pipeline.get_test_data(data)
     pipeline.run(X_train)
+
+
+@pytest.mark.parametrize(
+    "version_number",
+    [
+        "v0.1.1.2-b.1.0.1",
+        "v0.1.0.2-b.2.1.1",
+        "v0.1.1.2-b.3.2.1",
+        "v0.1.0.2-b.4.0.1",
+        "v0.1.0.2-b.5.1.1",
+        "v0.1.2.2-b.5.1.1",
+        "v0.1.3.2-b.2.1.1",
+    ],
+)
+def test_run_Pipeline_LGBMc_multi_label(version_number):
+    config_file = "config/LGBMc_config.toml"
+    data = load_data_from_csv("test/test_data/test_data.csv")
+    general_config = read_toml_configuration(str(DATA_DIR / config_file))
+    general_config["version"] = version_number
+    pipeline = Pipeline(general_config)
+
+    pipeline.preprocessor.fit(data)
+    X_train, _ = pipeline.get_test_data(data)
+    pipeline.run(X_train)