hydraflow 0.2.4__tar.gz → 0.2.5__tar.gz

{hydraflow-0.2.4 → hydraflow-0.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: hydraflow
-Version: 0.2.4
+Version: 0.2.5
 Summary: Hydraflow integrates Hydra and MLflow to manage and track machine learning experiments.
 Project-URL: Documentation, https://github.com/daizutabi/hydraflow
 Project-URL: Source, https://github.com/daizutabi/hydraflow
@@ -48,7 +48,7 @@ Description-Content-Type: text/markdown
 
 ## Overview
 
-Hydraflow is a powerful library designed to seamlessly integrate
+Hydraflow is a library designed to seamlessly integrate
 [Hydra](https://hydra.cc/) and [MLflow](https://mlflow.org/), making it easier to
 manage and track machine learning experiments. By combining the flexibility of
 Hydra's configuration management with the robust experiment tracking capabilities

{hydraflow-0.2.4 → hydraflow-0.2.5}/README.md

@@ -17,7 +17,7 @@
 
 ## Overview
 
-Hydraflow is a powerful library designed to seamlessly integrate
+Hydraflow is a library designed to seamlessly integrate
 [Hydra](https://hydra.cc/) and [MLflow](https://mlflow.org/), making it easier to
 manage and track machine learning experiments. By combining the flexibility of
 Hydra's configuration management with the robust experiment tracking capabilities

{hydraflow-0.2.4 → hydraflow-0.2.5}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "hydraflow"
-version = "0.2.4"
+version = "0.2.5"
 description = "Hydraflow integrates Hydra and MLflow to manage and track machine learning experiments."
 readme = "README.md"
 license = "MIT"

{hydraflow-0.2.4 → hydraflow-0.2.5}/src/hydraflow/runs.py

@@ -501,6 +501,33 @@ class RunCollection:
         """
         return (func(download_artifacts(run_id=run.info.run_id)) for run in self._runs)
 
+    def group_by(
+        self, names: list[str] | None = None, *args
+    ) -> dict[tuple[str, ...], RunCollection]:
+        """
+        Group the runs by the specified parameter names and return a dictionary
+        where the keys are the parameter values and the values are the runs.
+
+        Args:
+            names (list[str] | None): The parameter names to group by.
+            *args: Additional positional arguments to specify parameter names.
+
+        Returns:
+            A dictionary where the keys are the parameter values and the values
+            are the runs.
+        """
+        names = names[:] if names else []
+        names.extend(args)
+
+        grouped_runs = {}
+        for run in self._runs:
+            key = get_params(run, names)
+            if key not in grouped_runs:
+                grouped_runs[key] = []
+            grouped_runs[key].append(run)
+
+        return {key: RunCollection(runs) for key, runs in grouped_runs.items()}
+
 
 def _param_matches(run: Run, key: str, value: Any) -> bool:
     """
@@ -765,6 +792,13 @@ def try_get_run(runs: list[Run], config: object | None = None, **kwargs) -> Run
     raise ValueError(msg)
 
 
+def get_params(run: Run, names: list[str] | None = None, *args) -> tuple[str, ...]:
+    names = names[:] if names else []
+    names.extend(args)
+
+    return tuple(run.data.params[name] for name in names)
+
+
 def get_param_names(runs: list[Run]) -> list[str]:
     """
     Get the parameter names from the runs.

{hydraflow-0.2.4 → hydraflow-0.2.5}/tests/test_runs.py

@@ -419,6 +419,21 @@ def test_run_collection_contains(runs: RunCollection, i: int):
     assert runs._runs[i] in runs
 
 
+def test_run_collection_group_by(runs: RunCollection):
+    grouped = runs.group_by(["p"])
+    assert len(grouped) == 6
+    assert all(isinstance(group, RunCollection) for group in grouped.values())
+    assert all(len(group) == 1 for group in grouped.values())
+    assert grouped[("0",)][0] == runs[0]
+    assert grouped[("1",)][0] == runs[1]
+
+    grouped = runs.group_by(["q"])
+    assert len(grouped) == 2
+
+    grouped = runs.group_by(["r"])
+    assert len(grouped) == 3
+
+
 # def test_hydra_output_dir_error(runs_list: list[Run]):
 #     from hydraflow.runs import get_hydra_output_dir