hydraflow 0.8.0__tar.gz → 0.9.1__tar.gz

{hydraflow-0.8.0 → hydraflow-0.9.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hydraflow
-Version: 0.8.0
+Version: 0.9.1
 Summary: Hydraflow integrates Hydra and MLflow to manage and track machine learning experiments.
 Project-URL: Documentation, https://daizutabi.github.io/hydraflow/
 Project-URL: Source, https://github.com/daizutabi/hydraflow
@@ -41,6 +41,7 @@ Requires-Dist: mlflow>=2.15
 Requires-Dist: omegaconf
 Requires-Dist: rich
 Requires-Dist: typer
+Requires-Dist: ulid
 Description-Content-Type: text/markdown
 
 # Hydraflow
@@ -93,31 +94,29 @@ pip install hydraflow
 Here is a simple example to get you started with Hydraflow:
 
 ```python
-import hydra
-import hydraflow
-import mlflow
+from __future__ import annotations
+
 from dataclasses import dataclass
-from hydra.core.config_store import ConfigStore
 from pathlib import Path
+from typing import TYPE_CHECKING
 
-@dataclass
-class MySQLConfig:
-    host: str = "localhost"
-    port: int = 3306
+import hydraflow
 
-cs = ConfigStore.instance()
-cs.store(name="config", node=MySQLConfig)
+if TYPE_CHECKING:
+    from mlflow.entities import Run
+
+
+@dataclass
+class Config:
+    count: int = 1
+    name: str = "a"
 
-@hydra.main(config_name="config", version_base=None)
-def my_app(cfg: MySQLConfig) -> None:
-    # Set experiment by Hydra job name.
-    hydraflow.set_experiment()
 
-    # Automatically log Hydra config as params.
-    with hydraflow.start_run(cfg):
-        # Your app code below.
+@hydraflow.main(Config)
+def app(run: Run, cfg: Config):
+    """Your app code here."""
 
 
 if __name__ == "__main__":
-    my_app()
+    app()
 ```

{hydraflow-0.8.0 → hydraflow-0.9.1}/README.md

@@ -48,31 +48,29 @@ pip install hydraflow
 Here is a simple example to get you started with Hydraflow:
 
 ```python
-import hydra
-import hydraflow
-import mlflow
+from __future__ import annotations
+
 from dataclasses import dataclass
-from hydra.core.config_store import ConfigStore
 from pathlib import Path
+from typing import TYPE_CHECKING
 
-@dataclass
-class MySQLConfig:
-    host: str = "localhost"
-    port: int = 3306
+import hydraflow
 
-cs = ConfigStore.instance()
-cs.store(name="config", node=MySQLConfig)
+if TYPE_CHECKING:
+    from mlflow.entities import Run
+
+
+@dataclass
+class Config:
+    count: int = 1
+    name: str = "a"
 
-@hydra.main(config_name="config", version_base=None)
-def my_app(cfg: MySQLConfig) -> None:
-    # Set experiment by Hydra job name.
-    hydraflow.set_experiment()
 
-    # Automatically log Hydra config as params.
-    with hydraflow.start_run(cfg):
-        # Your app code below.
+@hydraflow.main(Config)
+def app(run: Run, cfg: Config):
+    """Your app code here."""
 
 
 if __name__ == "__main__":
-    my_app()
+    app()
 ```

{hydraflow-0.8.0 → hydraflow-0.9.1}/docs/usage/quickstart.md

@@ -12,16 +12,6 @@ There are two main steps to using Hydraflow:
 --8<-- "apps/quickstart.py"
 ```
 
-### Set the MLflow experiment
-
-[`hydraflow.set_experiment`][] sets the MLflow experiment using the Hydra job name.
-Optionally, it can also set the tracking URI with `uri` argument.
-For example,
-
-```python
-hydraflow.set_experiment(uri="sqlite:///mlruns.db")
-```
-
 ### Start a new MLflow run
 
 [`hydraflow.start_run`][] starts a new MLflow run that logs the Hydra configuration.
@@ -64,10 +54,8 @@ $ python apps/quickstart.py -m width=400,600 height=100,200,300
 ### Run collection
 
 ```pycon exec="1" source="console" session="quickstart"
->>> import mlflow
->>> mlflow.set_experiment("quickstart")
 >>> import hydraflow
->>> rc = hydraflow.list_runs()
+>>> rc = hydraflow.list_runs("quickstart")
 >>> print(rc)
 ```
 
@@ -107,16 +95,6 @@ $ python apps/quickstart.py -m width=400,600 height=100,200,300
 >>> print(filtered)
 ```
 
-```pycon exec="1" source="console" session="quickstart"
->>> run = rc.find(height=100)
->>> print(run.data.params)
-```
-
-```pycon exec="1" source="console" session="quickstart"
->>> run = rc.find_last(height=100)
->>> print(run.data.params)
-```
-
 ### Group runs
 
 ```pycon exec="1" source="console" session="quickstart"

{hydraflow-0.8.0 → hydraflow-0.9.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "hydraflow"
-version = "0.8.0"
+version = "0.9.1"
 description = "Hydraflow integrates Hydra and MLflow to manage and track machine learning experiments."
 readme = "README.md"
 license = { file = "LICENSE" }
@@ -19,7 +19,14 @@ classifiers = [
   "Programming Language :: Python :: 3.13",
 ]
 requires-python = ">=3.10"
-dependencies = ["hydra-core>=1.3", "mlflow>=2.15", "omegaconf", "rich", "typer"]
+dependencies = [
+  "hydra-core>=1.3",
+  "mlflow>=2.15",
+  "omegaconf",
+  "rich",
+  "typer",
+  "ulid",
+]
 
 [project.urls]
 Documentation = "https://daizutabi.github.io/hydraflow/"
@@ -44,6 +51,7 @@ addopts = [
   "--cov=hydraflow",
   "--cov-report=lcov:lcov.info",
   "--dist=loadgroup",
+  "--doctest-modules",
   "-n8",
 ]
 filterwarnings = [
@@ -67,6 +75,7 @@ ignore = [
   "ANN003",
   "ANN401",
   "B904",
+  "D104",
   "D105",
   "D107",
   "D203",
@@ -79,13 +88,15 @@ ignore = [
   "PLR0913",
   "PLR1704",
   "PLR2004",
+  "S603",
   "SIM102",
   "SIM108",
   "TRY003",
 ]
 
 [tool.ruff.lint.per-file-ignores]
-"tests/*" = ["A001", "ANN", "ARG", "D", "FBT", "PD", "PLR", "PT", "S", "SLF"]
 "apps/*.py" = ["D", "G", "INP"]
-"src/hydraflow/main.py" = ["ANN201", "D401"]
 "src/hydraflow/cli.py" = ["ANN", "D"]
+"src/hydraflow/core/main.py" = ["ANN201", "D401"]
+"src/hydraflow/executor/conf.py" = ["ANN", "D"]
+"tests/*" = ["A001", "ANN", "ARG", "D", "FBT", "PD", "PLR", "PT", "S", "SLF"]

hydraflow-0.9.1/src/hydraflow/__init__.py

@@ -0,0 +1,37 @@
+"""Integrate Hydra and MLflow to manage and track machine learning experiments."""
+
+from hydraflow.core.context import chdir_artifact, log_run, start_run
+from hydraflow.core.io import (
+    get_artifact_dir,
+    get_artifact_path,
+    get_hydra_output_dir,
+    iter_artifact_paths,
+    iter_artifacts_dirs,
+    iter_experiment_dirs,
+    iter_run_dirs,
+    load_config,
+    remove_run,
+)
+from hydraflow.core.main import main
+from hydraflow.core.mlflow import list_run_ids, list_run_paths, list_runs
+from hydraflow.entities.run_collection import RunCollection
+
+__all__ = [
+    "RunCollection",
+    "chdir_artifact",
+    "get_artifact_dir",
+    "get_artifact_path",
+    "get_hydra_output_dir",
+    "iter_artifact_paths",
+    "iter_artifacts_dirs",
+    "iter_experiment_dirs",
+    "iter_run_dirs",
+    "list_run_ids",
+    "list_run_paths",
+    "list_runs",
+    "load_config",
+    "log_run",
+    "main",
+    "remove_run",
+    "start_run",
+]

hydraflow-0.9.1/src/hydraflow/cli.py

@@ -0,0 +1,67 @@
+"""Hydraflow CLI."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Annotated
+
+import typer
+from rich.console import Console
+from typer import Argument, Option
+
+from hydraflow.executor.io import load_config
+
+if TYPE_CHECKING:
+    from hydraflow.executor.job import Job
+
+app = typer.Typer(add_completion=False)
+console = Console()
+
+
+def get_job(name: str) -> Job:
+    cfg = load_config()
+    job = cfg.jobs[name]
+
+    if not job.name:
+        job.name = name
+
+    return job
+
+
+@app.command()
+def run(
+    name: Annotated[str, Argument(help="Job name.", show_default=False)],
+) -> None:
+    """Run a job."""
+    import mlflow
+
+    from hydraflow.executor.job import multirun
+
+    job = get_job(name)
+    mlflow.set_experiment(job.name)
+    multirun(job)
+
+
+@app.command()
+def show(
+    name: Annotated[str, Argument(help="Job name.", show_default=False)],
+) -> None:
+    """Show a job."""
+    from hydraflow.executor.job import show
+
+    job = get_job(name)
+    show(job)
+
+
+@app.callback(invoke_without_command=True)
+def callback(
+    *,
+    version: Annotated[
+        bool,
+        Option("--version", help="Show the version and exit."),
+    ] = False,
+) -> None:
+    if version:
+        import importlib.metadata
+
+        typer.echo(f"hydraflow {importlib.metadata.version('hydraflow')}")
+        raise typer.Exit

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/core}/context.py

@@ -12,8 +12,9 @@ import mlflow
 import mlflow.artifacts
 from hydra.core.hydra_config import HydraConfig
 
-from hydraflow.mlflow import log_params, log_text
-from hydraflow.utils import get_artifact_dir
+from hydraflow.core.io import get_artifact_dir
+
+from .mlflow import log_params, log_text
 
 if TYPE_CHECKING:
     from collections.abc import Iterator

hydraflow-0.8.0/src/hydraflow/utils.py → hydraflow-0.9.1/src/hydraflow/core/io.py

@@ -15,7 +15,7 @@ from mlflow.entities import Run
 from omegaconf import DictConfig, ListConfig, OmegaConf
 
 if TYPE_CHECKING:
-    from collections.abc import Iterable
+    from collections.abc import Iterable, Iterator
 
 
 def file_uri_to_path(uri: str) -> Path:
@@ -147,3 +147,74 @@ def remove_run(run: Run | Iterable[Run]) -> None:
         return
 
     shutil.rmtree(get_artifact_dir(run).parent)
+
+
+def get_root_dir(uri: str | Path | None = None) -> Path:
+    """Get the root directory for the MLflow tracking server."""
+    if uri is not None:
+        return Path(uri).absolute()
+
+    uri = mlflow.get_tracking_uri()
+
+    if uri.startswith("file:"):
+        return file_uri_to_path(uri)
+
+    return Path(uri).absolute()
+
+
+def get_experiment_name(path: Path) -> str | None:
+    """Get the experiment name from the meta file."""
+    metafile = path / "meta.yaml"
+    if not metafile.exists():
+        return None
+    lines = metafile.read_text().splitlines()
+    for line in lines:
+        if line.startswith("name:"):
+            return line.split(":")[1].strip()
+    return None
+
+
+def iter_experiment_dirs(
+    experiment_names: str | list[str] | None = None,
+    root_dir: str | Path | None = None,
+) -> Iterator[Path]:
+    """Iterate over the experiment directories in the root directory."""
+    if isinstance(experiment_names, str):
+        experiment_names = [experiment_names]
+
+    root_dir = get_root_dir(root_dir)
+    for path in root_dir.iterdir():
+        if path.is_dir() and path.name not in [".trash", "0"]:
+            if name := get_experiment_name(path):
+                if experiment_names is None or name in experiment_names:
+                    yield path
+
+
+def iter_run_dirs(
+    experiment_names: str | list[str] | None = None,
+    root_dir: str | Path | None = None,
+) -> Iterator[Path]:
+    """Iterate over the run directories in the root directory."""
+    for experiment_dir in iter_experiment_dirs(experiment_names, root_dir):
+        for path in experiment_dir.iterdir():
+            if path.is_dir() and (path / "artifacts").exists():
+                yield path
+
+
+def iter_artifacts_dirs(
+    experiment_names: str | list[str] | None = None,
+    root_dir: str | Path | None = None,
+) -> Iterator[Path]:
+    """Iterate over the artifacts directories in the root directory."""
+    for path in iter_run_dirs(experiment_names, root_dir):
+        yield path / "artifacts"
+
+
+def iter_artifact_paths(
+    artifact_path: str | Path,
+    experiment_names: str | list[str] | None = None,
+    root_dir: str | Path | None = None,
+) -> Iterator[Path]:
+    """Iterate over the artifact paths in the root directory."""
+    for path in iter_artifacts_dirs(experiment_names, root_dir):
+        yield path / artifact_path

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/core}/main.py

@@ -7,6 +7,7 @@ management.
 
 The main functionality is provided through the `main` decorator, which can be
 used to wrap experiment entry points. This decorator handles:
+
 - Configuration management via Hydra
 - Experiment tracking via MLflow
 - Run deduplication based on configurations
@@ -44,11 +45,12 @@ from mlflow.entities import RunStatus
 from omegaconf import OmegaConf
 
 import hydraflow
-from hydraflow.utils import file_uri_to_path
+from hydraflow.core.io import file_uri_to_path
 
 if TYPE_CHECKING:
     from collections.abc import Callable
     from pathlib import Path
+    from typing import Any
 
     from mlflow.entities import Run
 
@@ -115,7 +117,7 @@ def main(
     return decorator
 
 
-def get_run_id(uri: str, config: object, overrides: list[str] | None) -> str | None:
+def get_run_id(uri: str, config: Any, overrides: list[str] | None) -> str | None:
     """Try to get the run ID for the given configuration.
 
     If the run is not found, the function will return None.
@@ -137,7 +139,7 @@ def get_run_id(uri: str, config: object, overrides: list[str] | None) -> str | N
     return None
 
 
-def equals(run_dir: Path, config: object, overrides: list[str] | None) -> bool:
+def equals(run_dir: Path, config: Any, overrides: list[str] | None) -> bool:
     """Check if the run directory matches the given configuration or overrides.
 
     Args:

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/core}/mlflow.py

@@ -13,9 +13,10 @@ import joblib
 import mlflow
 import mlflow.artifacts
 
-from hydraflow.config import iter_params
-from hydraflow.run_collection import RunCollection
-from hydraflow.utils import file_uri_to_path, get_artifact_dir
+from hydraflow.core.io import file_uri_to_path, get_artifact_dir
+from hydraflow.entities.run_collection import RunCollection
+
+from .config import iter_params
 
 if TYPE_CHECKING:
     from pathlib import Path

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/entities}/run_collection.py

@@ -25,12 +25,13 @@ from typing import TYPE_CHECKING, Any, overload
 
 from mlflow.entities import RunStatus
 
-import hydraflow.param
-from hydraflow.config import iter_params, select_config, select_overrides
-from hydraflow.param import get_params, get_values
-from hydraflow.run_data import RunCollectionData
-from hydraflow.run_info import RunCollectionInfo
-from hydraflow.utils import load_config
+import hydraflow.core.param
+from hydraflow.core.config import iter_params, select_config, select_overrides
+from hydraflow.core.io import load_config
+from hydraflow.core.param import get_params, get_values
+
+from .run_data import RunCollectionData
+from .run_info import RunCollectionInfo
 
 if TYPE_CHECKING:
     from collections.abc import Callable, Iterator
@@ -478,7 +479,7 @@ def _param_matches(run: Run, key: str, value: Any) -> bool:
     if param == "None":
         return value is None or value == "None"
 
-    return hydraflow.param.match(param, value)
+    return hydraflow.core.param.match(param, value)
 
 
 def filter_runs(

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/entities}/run_data.py

@@ -6,14 +6,14 @@ from typing import TYPE_CHECKING
 
 from pandas import DataFrame
 
-from hydraflow.config import iter_params
-from hydraflow.utils import load_config
+from hydraflow.core.config import iter_params
+from hydraflow.core.io import load_config
 
 if TYPE_CHECKING:
     from collections.abc import Iterable
     from typing import Any
 
-    from hydraflow.run_collection import RunCollection
+    from .run_collection import RunCollection
 
 
 class RunCollectionData:

{hydraflow-0.8.0/src/hydraflow → hydraflow-0.9.1/src/hydraflow/entities}/run_info.py

@@ -4,12 +4,12 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING
 
-from hydraflow.utils import get_artifact_dir
+from hydraflow.core.io import get_artifact_dir
 
 if TYPE_CHECKING:
     from pathlib import Path
 
-    from hydraflow.run_collection import RunCollection
+    from .run_collection import RunCollection
 
 
 class RunCollectionInfo:

hydraflow-0.9.1/src/hydraflow/executor/conf.py

@@ -0,0 +1,23 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+
+
+@dataclass
+class Step:
+    args: str = ""
+    batch: str = ""
+    options: str = ""
+
+
+@dataclass
+class Job:
+    name: str = ""
+    run: str = ""
+    call: str = ""
+    steps: list[Step] = field(default_factory=list)
+
+
+@dataclass
+class HydraflowConf:
+    jobs: dict[str, Job] = field(default_factory=dict)

hydraflow-0.9.1/src/hydraflow/executor/io.py

@@ -0,0 +1,34 @@
+"""Hydraflow jobs IO."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+from omegaconf import OmegaConf
+
+from .conf import HydraflowConf
+
+
+def find_config_file() -> Path | None:
+    """Find the hydraflow config file."""
+    if Path("hydraflow.yaml").exists():
+        return Path("hydraflow.yaml")
+
+    if Path("hydraflow.yml").exists():
+        return Path("hydraflow.yml")
+
+    return None
+
+
+def load_config() -> HydraflowConf:
+    """Load the hydraflow config."""
+    schema = OmegaConf.structured(HydraflowConf)
+
+    path = find_config_file()
+
+    if path is None:
+        return schema
+
+    cfg = OmegaConf.load(path)
+
+    return OmegaConf.merge(schema, cfg)  # type: ignore