# KozhinDev ClearML Wrapper
A set of utility functions and decorators for simplifying ClearML integration in your ML projects
## Features
- **`log_metrics`** – decorator for logging scalar metrics
- **`get_local_dataset_path`** – get local path to a dataset file/folder
- **`prepare_task`** – initialize or retrieve ClearML tasks (local or remote)
- **`get_local_model_path`** – retrieve a local copy of a model artifact
First, install the library: ```pip install kozhindev_clearml_wrapper```
# Usage
1. Logging metrics
Use the @log_metrics decorator to log scalar metrics to ClearML:
```python
from clearml import Task
from kozhindev_clearml_wrapper import log_metrics
task = Task.init(project_name="Demo", task_name="Log Metrics Example")
@log_metrics(task)
def train():
for i in range(5):
yield "loss", "train", 0.1 * i, i # (title, series, metric, iteration)
train()
```
2. Get Local Dataset Path
Retrieve a dataset by ID and get the local path to a file or folder inside it:
```python
from kozhindev_clearml_wrapper import get_local_dataset_path
dataset_path = get_local_dataset_path(
dataset_id="your-dataset-id",
dataset_name="data.csv"
)
print(dataset_path)
```
3. Prepare ClearML Task
Initialize a new task (local) or connect to an existing one (remote):
```python
from kozhindev_clearml_wrapper import prepare_task
# Local task (creates a new one)
task = prepare_task(
task_type="local",
project_name="Demo",
task_name="Local Task Example"
)
# Remote task (connects training parameters to the current ClearML task)
train_params = {"learning_rate": 0.001, "epochs": 10}
task = prepare_task(
task_type="remote",
train_params=train_params
)
```
4. Get Local Model Path
Download and get the local path to a model artifact from a specific task:
```python
from kozhindev_clearml_wrapper import get_local_model_path
path_to_model = get_local_model_path(
task_id="your-task-id",
artifact_name="trained_models"
)
print(path_to_model)
```
Raw data
{
"_id": null,
"home_page": null,
"name": "kozhindev-clearml-wrapper",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "clearml-wrapper clearml wrapper",
"author": "YVoskanyan",
"author_email": "yuvoskanyan@team.kozhin.dev",
"download_url": "https://files.pythonhosted.org/packages/03/b2/8f4d793f034bbd677f605e44a37f16344bf466e79d314cf4dcca92c93cdf/kozhindev_clearml_wrapper-0.0.2.tar.gz",
"platform": null,
"description": "# KozhinDev ClearML Wrapper\nA set of utility functions and decorators for simplifying ClearML integration in your ML projects \n## Features\n\n- **`log_metrics`** \u2013 decorator for logging scalar metrics \n- **`get_local_dataset_path`** \u2013 get local path to a dataset file/folder \n- **`prepare_task`** \u2013 initialize or retrieve ClearML tasks (local or remote) \n- **`get_local_model_path`** \u2013 retrieve a local copy of a model artifact\n\nFirst, install the library: ```pip install kozhindev_clearml_wrapper```\n# Usage\n1. Logging metrics \nUse the @log_metrics decorator to log scalar metrics to ClearML:\n```python\nfrom clearml import Task\nfrom kozhindev_clearml_wrapper import log_metrics\n\ntask = Task.init(project_name=\"Demo\", task_name=\"Log Metrics Example\")\n\n@log_metrics(task)\ndef train():\n for i in range(5):\n yield \"loss\", \"train\", 0.1 * i, i # (title, series, metric, iteration)\n\ntrain()\n```\n2. Get Local Dataset Path \nRetrieve a dataset by ID and get the local path to a file or folder inside it:\n```python\nfrom kozhindev_clearml_wrapper import get_local_dataset_path\n\ndataset_path = get_local_dataset_path(\n dataset_id=\"your-dataset-id\",\n dataset_name=\"data.csv\"\n)\n\nprint(dataset_path)\n```\n3. Prepare ClearML Task \nInitialize a new task (local) or connect to an existing one (remote):\n```python\nfrom kozhindev_clearml_wrapper import prepare_task\n\n# Local task (creates a new one)\ntask = prepare_task(\n task_type=\"local\",\n project_name=\"Demo\",\n task_name=\"Local Task Example\"\n)\n\n# Remote task (connects training parameters to the current ClearML task)\ntrain_params = {\"learning_rate\": 0.001, \"epochs\": 10}\ntask = prepare_task(\n task_type=\"remote\",\n train_params=train_params\n)\n```\n4. Get Local Model Path \nDownload and get the local path to a model artifact from a specific task:\n```python\nfrom kozhindev_clearml_wrapper import get_local_model_path\n\npath_to_model = get_local_model_path(\n task_id=\"your-task-id\",\n artifact_name=\"trained_models\"\n)\n\nprint(path_to_model)\n```\n",
"bugtrack_url": null,
"license": null,
"summary": "Wrapper over clearml, developed by ML team of KozhinDev company",
"version": "0.0.2",
"project_urls": null,
"split_keywords": [
"clearml-wrapper",
"clearml",
"wrapper"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "6c6bbc6901249656fe9d7c564605a5899734b000e1a33c920ea1be949cb25489",
"md5": "c42def215dd497fdaba159042f66d2aa",
"sha256": "4ee16aa05ec92b6b237da08dfac66af2cdacfc5627d44d7c6d7fed1397f530bf"
},
"downloads": -1,
"filename": "kozhindev_clearml_wrapper-0.0.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "c42def215dd497fdaba159042f66d2aa",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 4705,
"upload_time": "2025-09-03T05:19:48",
"upload_time_iso_8601": "2025-09-03T05:19:48.735610Z",
"url": "https://files.pythonhosted.org/packages/6c/6b/bc6901249656fe9d7c564605a5899734b000e1a33c920ea1be949cb25489/kozhindev_clearml_wrapper-0.0.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "03b28f4d793f034bbd677f605e44a37f16344bf466e79d314cf4dcca92c93cdf",
"md5": "c8f603cfcbf47faaf2d3e2751be5de00",
"sha256": "e0150e99a06c14d26dfc1e152a8e4b95cc6430801268f948472b22d1d676cd95"
},
"downloads": -1,
"filename": "kozhindev_clearml_wrapper-0.0.2.tar.gz",
"has_sig": false,
"md5_digest": "c8f603cfcbf47faaf2d3e2751be5de00",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 4024,
"upload_time": "2025-09-03T05:19:50",
"upload_time_iso_8601": "2025-09-03T05:19:50.078464Z",
"url": "https://files.pythonhosted.org/packages/03/b2/8f4d793f034bbd677f605e44a37f16344bf466e79d314cf4dcca92c93cdf/kozhindev_clearml_wrapper-0.0.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-09-03 05:19:50",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "kozhindev-clearml-wrapper"
}
JSON