Upload, download, and inspect datasets stored in MLflow on the BioLM platform.
Datasets are backed by MLflow runs and artifacts; install the mlflow extra and
authenticate with biolm login before using these commands.
Usage
biolm dataset [OPTIONS] COMMAND [ARGS]...
biolm dataset download
Download dataset artifacts from MLflow to a local directory.
Fetches all artifacts by default, or a single path when --artifact-path is set.
Usage
biolm dataset download [OPTIONS] DATASET_ID [OUTPUT_PATH]
Options
- --experiment <experiment>
MLflow experiment name (default: {username}/datasets)
- --artifact-path <artifact_path>
Specific artifact path to download (default: all artifacts)
- --mlflow-uri <mlflow_uri>
MLflow tracking URI
Arguments
- DATASET_ID
Required argument
- OUTPUT_PATH
Optional argument
biolm dataset list
List datasets in your MLflow experiment with optional JSON or CSV export.
By default lists runs tagged as datasets under {username}/datasets.
Usage
biolm dataset list [OPTIONS]
Options
- --experiment <experiment>
MLflow experiment name (default: {username}/datasets)
- --format <format>
Output format
- Options:
table | json | csv
- -o, --output <output>
Save output to file
- --mlflow-uri <mlflow_uri>
MLflow tracking URI
- --all-runs
List all runs in experiment, not just datasets (for debugging)
biolm dataset show
Show metadata, tags, metrics, and artifact listings for a dataset by ID.
Resolves the dataset in your default {username}/datasets experiment unless overridden.
Usage
biolm dataset show [OPTIONS] DATASET_ID
Options
- --experiment <experiment>
MLflow experiment name (default: {username}/datasets)
- --format <format>
Output format
- Options:
table | json | yaml
- -o, --output <output>
Save output to file
- --mlflow-uri <mlflow_uri>
MLflow tracking URI
Arguments
- DATASET_ID
Required argument
biolm dataset upload
Upload a file or directory to a dataset, creating the dataset run if needed.
Artifacts are stored in MLflow; use --recursive for directory uploads.
Usage
biolm dataset upload [OPTIONS] DATASET_ID FILE_PATH
Options
- --experiment <experiment>
MLflow experiment name (default: {username}/datasets)
- --name <name>
Dataset name/description (stored as run name)
- -r, --recursive
Upload directory recursively
- --mlflow-uri <mlflow_uri>
MLflow tracking URI
Arguments
- DATASET_ID
Required argument
- FILE_PATH
Required argument