Biotope Annotate

Draft stage

Biotope is in draft stage. Functionality may be missing or incomplete. The API is subject to change.

Overview

The biotope annotate command creates dataset metadata definitions in Croissant format. It provides interactive and programmatic ways to annotate your data with consistent metadata.

Features

Interactive Annotation

The biotope annotate interactive command provides a user-friendly interface for creating metadata:

• Project Metadata Pre-fill: Automatically pre-fills fields with project-level metadata if available
• Staged Files: Annotate files that have been staged with biotope add or biotope get
• Incomplete Annotations: Complete or update existing metadata files
• Validation: Built-in validation of metadata structure and content

Project Metadata Integration

When using interactive annotation, biotope automatically loads project-level metadata from .biotope/config/biotope.yaml and pre-fills the annotation form with:

• Description: Project description (if not already specified)
• URL: Project URL
• Creator: Project creator information
• License: Project license
• Citation: Project citation information

This makes the annotation process faster and ensures consistency across all datasets in the project.

Commands

biotope annotate interactive

Interactive annotation process for files with project metadata pre-fill.

Usage:

biotope annotate interactive [OPTIONS]

Options: - --file, -f: Specific file to annotate - --prefill-metadata: JSON string with metadata to pre-fill - --staged: Annotate all staged files - --incomplete: Complete incomplete annotations

Examples:

# Interactive annotation with project metadata pre-fill
biotope annotate interactive

# Annotate a specific file
biotope annotate interactive --file data/raw/experiment.csv

# Annotate all staged files
biotope annotate interactive --staged

# Complete incomplete annotations
biotope annotate interactive --incomplete

# Pre-fill with custom metadata
biotope annotate interactive --prefill-metadata '{"description": "Custom description"}'

biotope annotate create

Create metadata from command line parameters (non-interactive mode).

Usage:

biotope annotate create [OPTIONS]

Required Options: - --name, -n: Name of the dataset - --data-source, -s: URL or path to the data source - --access-restrictions, -a: Note on access restrictions (e.g., public, restricted, private)

Optional Options: - --output, -o: Output file path (default: metadata.json) - --description, -d: Description of the dataset - --contact, -c: Responsible contact person (default: current user) - --date: Date of creation in ISO format (default: today) - --format, -f: Description of file format - --legal-obligations, -l: Note on legal obligations - --collaboration-partner, -p: Collaboration partner and institute

Examples:

# Create metadata with required fields only
biotope annotate create \
  --name "RNA-seq Dataset" \
  --data-source "https://example.com/rna-seq-data" \
  --access-restrictions "Public"

# Create metadata with all fields
biotope annotate create \
  --name "Proteomics Dataset" \
  --description "Mass spectrometry data from protein samples" \
  --data-source "https://example.com/proteomics-data" \
  --contact "researcher@university.edu" \
  --date "2024-01-15" \
  --access-restrictions "Restricted to academic use" \
  --format "mzML" \
  --legal-obligations "Data sharing agreement required" \
  --collaboration-partner "Proteomics Center, University Hospital" \
  --output "proteomics_metadata.json"

Generated Metadata Structure:

{
  "@context": {
    "@vocab": "https://schema.org/",
    "cr": "https://mlcommons.org/croissant/",
    "ml": "http://ml-schema.org/",
    "sc": "https://schema.org/",
    "dct": "http://purl.org/dc/terms/",
    "data": "https://mlcommons.org/croissant/data/",
    "rai": "https://mlcommons.org/croissant/rai/",
    "format": "https://mlcommons.org/croissant/format/",
    "citeAs": "https://mlcommons.org/croissant/citeAs/",
    "conformsTo": "https://mlcommons.org/croissant/conformsTo/",
    "@language": "en",
    "repeated": "https://mlcommons.org/croissant/repeated/",
    "field": "https://mlcommons.org/croissant/field/",
    "examples": "https://mlcommons.org/croissant/examples/",
    "recordSet": "https://mlcommons.org/croissant/recordSet/",
    "fileObject": "https://mlcommons.org/croissant/fileObject/",
    "fileSet": "https://mlcommons.org/croissant/fileSet/",
    "source": "https://mlcommons.org/croissant/source/",
    "references": "https://mlcommons.org/croissant/references/",
    "key": "https://mlcommons.org/croissant/key/",
    "parentField": "https://mlcommons.org/croissant/parentField/",
    "isLiveDataset": "https://mlcommons.org/croissant/isLiveDataset/",
    "separator": "https://mlcommons.org/croissant/separator/",
    "extract": "https://mlcommons.org/croissant/extract/",
    "subField": "https://mlcommons.org/croissant/subField/",
    "regex": "https://mlcommons.org/croissant/regex/",
    "column": "https://mlcommons.org/croissant/column/",
    "path": "https://mlcommons.org/croissant/path/",
    "fileProperty": "https://mlcommons.org/croissant/fileProperty/",
    "md5": "https://mlcommons.org/croissant/md5/",
    "jsonPath": "https://mlcommons.org/croissant/jsonPath/",
    "transform": "https://mlcommons.org/croissant/transform/",
    "replace": "https://mlcommons.org/croissant/replace/",
    "dataType": "https://mlcommons.org/croissant/dataType/"
  },
  "@type": "Dataset",
  "name": "Proteomics Dataset",
  "description": "Mass spectrometry data from protein samples",
  "url": "https://example.com/proteomics-data",
  "creator": {
    "@type": "Person",
    "name": "researcher@university.edu"
  },
  "dateCreated": "2024-01-15",
  "datePublished": "2024-01-15",
  "version": "1.0",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "citation": "Please cite this dataset as: Proteomics Dataset (2024)",
  "cr:accessRestrictions": "Restricted to academic use",
  "encodingFormat": "mzML",
  "cr:legalObligations": "Data sharing agreement required",
  "cr:collaborationPartner": "Proteomics Center, University Hospital",
  "distribution": []
}

biotope annotate validate

Validate a Croissant metadata file using the mlcroissant CLI.

Usage:

biotope annotate validate --jsonld <file_name.json>

Options: - --jsonld, -j: Path to the JSON-LD metadata file to validate (required)

Examples:

# Validate a metadata file
biotope annotate validate --jsonld metadata.json

# Validate with full path
biotope annotate validate --jsonld /path/to/dataset_metadata.json

Output:

Validation successful! The metadata file is valid.

Error Handling: - If validation fails, the command will exit with an error code and display validation errors - Warnings are displayed but don't cause the command to fail - The command uses the mlcroissant CLI under the hood for validation

biotope annotate load

Load records from a dataset using its Croissant metadata.

Usage:

biotope annotate load [OPTIONS]

Required Options: - --jsonld, -j: Path to the JSON-LD metadata file - --record-set, -r: Name of the record set to load

Optional Options: - --num-records, -n: Number of records to load (default: 10)

Examples:

# Load 10 records from the default record set
biotope annotate load \
  --jsonld metadata.json \
  --record-set samples

# Load 50 records from a specific record set
biotope annotate load \
  --jsonld metadata.json \
  --record-set measurements \
  --num-records 50

Output:

Record 1: {'patient_id': 'P001', 'gene_expression': [0.1, 0.2, 0.3]}
Record 2: {'patient_id': 'P002', 'gene_expression': [0.4, 0.5, 0.6]}
...
Loaded 10 records from record set 'samples'

Metadata Pre-fill Priority

When pre-filling metadata, the following priority order is used:

1. Command-line prefill: Metadata provided via --prefill-metadata
2. Project metadata: Metadata from .biotope/config/biotope.yaml
3. Default values: Built-in defaults for required fields

This allows you to override project-level metadata when needed while still benefiting from the convenience of pre-filled values.

Croissant ML Format

Biotope uses the Croissant ML metadata format for dataset annotations. This format provides:

• Standardized structure: Consistent metadata across different datasets
• Rich annotations: Support for complex metadata relationships
• Validation: Built-in validation of metadata structure
• Interoperability: Compatible with other ML metadata tools

Key Croissant ML Concepts

1. Dataset: The top-level container for all metadata
2. Distribution: How the dataset is distributed (FileObject, FileSet)
3. RecordSet: Structure of data records
4. Field: Individual data fields with types and constraints
5. Context: JSON-LD context for semantic meaning

Scientific Metadata Fields

Biotope includes additional scientific metadata fields:

• cr:accessRestrictions: Data access restrictions
• cr:legalObligations: Legal requirements for data usage
• cr:collaborationPartner: Collaboration partner information
• cr:projectName: Project name for context

Configuration

Project metadata is managed through the biotope config command:

# Set project metadata for pre-filling
biotope config set-project-metadata

# View current project metadata
biotope config show-project-metadata

Best Practices

1. Use Interactive Mode: For most use cases, biotope annotate interactive provides the best user experience
2. Set Project Metadata: Configure project-level metadata early for consistent annotations
3. Validate Regularly: Use biotope annotate validate to ensure metadata quality
4. Use Descriptive Names: Choose clear, descriptive dataset names
5. Include All Required Fields: Ensure all required metadata fields are completed
6. Test with Load: Use biotope annotate load to verify metadata works with actual data

Integration with Git Workflow

The annotate commands integrate seamlessly with biotope's Git workflow:

# Add files to staging
biotope add data/raw/experiment.csv

# Create metadata interactively
biotope annotate interactive --staged

# Validate the metadata
biotope annotate validate --jsonld .biotope/datasets/experiment.csv.jsonld

# Commit the changes
biotope commit -m "Add experiment dataset with metadata"

Troubleshooting

Validation Errors

• Ensure all required fields are present
• Check that field values meet validation requirements
• Verify JSON-LD syntax is correct

Load Errors

• Ensure the record set name exists in the metadata
• Check that the data source is accessible
• Verify the metadata file is valid

Interactive Mode Issues

• Make sure you're in a biotope project directory
• Check that files are properly staged with biotope add
• Verify project metadata is configured if using pre-fill

Command for creating dataset metadata definitions in Croissant format.

annotate()

Create dataset metadata definitions in Croissant format.

Source code in biotope/commands/annotate.py

@click.group()
def annotate() -> None:
    """Create dataset metadata definitions in Croissant format."""

create(output, name, description, data_source, contact, date, access_restrictions, format, legal_obligations, collaboration_partner)

Create a new Croissant metadata file with required scientific metadata fields.

Source code in biotope/commands/annotate.py

@annotate.command()
@click.option(
    "--output",
    "-o",
    type=click.Path(),
    default="metadata.json",
    help="Output file path for the metadata JSON-LD.",
)
@click.option(
    "--name",
    "-n",
    required=True,
    help="Name of the dataset.",
)
@click.option(
    "--description",
    "-d",
    default="",
    help="Description of the dataset.",
)
@click.option(
    "--data-source",
    "-s",
    required=True,
    help="URL or path to the data source.",
)
@click.option(
    "--contact",
    "-c",
    default=getpass.getuser(),
    help="Responsible contact person for the dataset.",
)
@click.option(
    "--date",
    default=datetime.datetime.now(tz=datetime.timezone.utc).date().isoformat(),
    help="Date of creation (ISO format: YYYY-MM-DD).",
)
@click.option(
    "--access-restrictions",
    "-a",
    required=True,
    help="Note on access restrictions (e.g., public, restricted, private).",
)
@click.option(
    "--format",
    "-f",
    help="Description of file format.",
)
@click.option(
    "--legal-obligations",
    "-l",
    help="Note on legal obligations.",
)
@click.option(
    "--collaboration-partner",
    "-p",
    help="Collaboration partner and institute.",
)
def create(
    output,
    name,
    description,
    data_source,
    contact,
    date,
    access_restrictions,
    format,
    legal_obligations,
    collaboration_partner,
):
    """Create a new Croissant metadata file with required scientific metadata fields."""
    # Create a basic metadata structure with proper Croissant context
    metadata = {
        "@context": {
            "@vocab": "https://schema.org/",
            "cr": "https://mlcommons.org/croissant/",
            "ml": "http://ml-schema.org/",
            "sc": "https://schema.org/",
            "dct": "http://purl.org/dc/terms/",
            "data": "https://mlcommons.org/croissant/data/",
            "rai": "https://mlcommons.org/croissant/rai/",
            "format": "https://mlcommons.org/croissant/format/",
            "citeAs": "https://mlcommons.org/croissant/citeAs/",
            "conformsTo": "https://mlcommons.org/croissant/conformsTo/",
            "@language": "en",
            "repeated": "https://mlcommons.org/croissant/repeated/",
            "field": "https://mlcommons.org/croissant/field/",
            "examples": "https://mlcommons.org/croissant/examples/",
            "recordSet": "https://mlcommons.org/croissant/recordSet/",
            "fileObject": "https://mlcommons.org/croissant/fileObject/",
            "fileSet": "https://mlcommons.org/croissant/fileSet/",
            "source": "https://mlcommons.org/croissant/source/",
            "references": "https://mlcommons.org/croissant/references/",
            "key": "https://mlcommons.org/croissant/key/",
            "parentField": "https://mlcommons.org/croissant/parentField/",
            "isLiveDataset": "https://mlcommons.org/croissant/isLiveDataset/",
            "separator": "https://mlcommons.org/croissant/separator/",
            "extract": "https://mlcommons.org/croissant/extract/",
            "subField": "https://mlcommons.org/croissant/subField/",
            "regex": "https://mlcommons.org/croissant/regex/",
            "column": "https://mlcommons.org/croissant/column/",
            "path": "https://mlcommons.org/croissant/path/",
            "fileProperty": "https://mlcommons.org/croissant/fileProperty/",
            "md5": "https://mlcommons.org/croissant/md5/",
            "jsonPath": "https://mlcommons.org/croissant/jsonPath/",
            "transform": "https://mlcommons.org/croissant/transform/",
            "replace": "https://mlcommons.org/croissant/replace/",
            "dataType": "https://mlcommons.org/croissant/dataType/",
        },
        "@type": "Dataset",
        "name": name,
        "description": description,
        "url": data_source,  # Changed from dataSource to url for schema.org compatibility
        "creator": {
            "@type": "Person",
            "name": contact,
        },
        "dateCreated": date,
        # Add recommended properties
        "datePublished": date,  # Use creation date as publication date by default
        "version": "1.0",  # Default version
        "license": "https://creativecommons.org/licenses/by/4.0/",  # Default license
        "citation": f"Please cite this dataset as: {name} ({date.split('-')[0]})",  # Simple citation
    }

    # Add custom fields with proper namespacing
    metadata["cr:accessRestrictions"] = access_restrictions

    # Add optional fields if provided
    if format:
        metadata["encodingFormat"] = format  # Using schema.org standard property
    if legal_obligations:
        metadata["cr:legalObligations"] = legal_obligations
    if collaboration_partner:
        metadata["cr:collaborationPartner"] = collaboration_partner

    # Add distribution property with empty array for FileObjects/FileSets
    metadata["distribution"] = []

    # Write to file
    with open(output, "w") as f:
        json.dump(metadata, f, indent=2)

    # Stage the changes in Git if we're in a biotope project
    try:
        biotope_root = find_biotope_root()
        if biotope_root:
            import subprocess
            subprocess.run(
                ["git", "add", ".biotope/"],
                cwd=biotope_root,
                check=True
            )
            click.echo(f"✅ Staged changes in Git")
    except (subprocess.CalledProcessError, FileNotFoundError):
        pass  # Not in a biotope project or Git not available

    click.echo(f"Created Croissant metadata file at {output}")

get_staged_files(biotope_root)

Get list of staged files from Git.

Source code in biotope/commands/annotate.py

def get_staged_files(biotope_root: Path) -> list:
    """Get list of staged files from Git."""
    import json
    import subprocess
    staged_files = []

    try:
        # Get staged files from Git
        result = subprocess.run(
            ["git", "diff", "--cached", "--name-only"],
            cwd=biotope_root,
            capture_output=True,
            text=True,
            check=True
        )

        for file_path in result.stdout.splitlines():
            if file_path.startswith(".biotope/datasets/") and file_path.endswith(".jsonld"):
                # Read the metadata file to get file information
                metadata_file = biotope_root / file_path
                try:
                    with open(metadata_file) as f:
                        metadata = json.load(f)
                        for distribution in metadata.get("distribution", []):
                            if distribution.get("@type") == "sc:FileObject":
                                staged_files.append({
                                    "file_path": distribution.get("contentUrl"),
                                    "sha256": distribution.get("sha256"),
                                    "size": distribution.get("contentSize")
                                })
                except (json.JSONDecodeError, KeyError):
                    continue

    except subprocess.CalledProcessError:
        pass

    return staged_files

get_standard_context()

Get the standard Croissant context.

Source code in biotope/commands/annotate.py

def get_standard_context() -> dict:
    """Get the standard Croissant context."""
    return {
        "@vocab": "https://schema.org/",
        "cr": "https://mlcommons.org/croissant/",
        "ml": "http://ml-schema.org/",
        "sc": "https://schema.org/",
        "dct": "http://purl.org/dc/terms/",
        "data": "https://mlcommons.org/croissant/data/",
        "rai": "https://mlcommons.org/croissant/rai/",
        "format": "https://mlcommons.org/croissant/format/",
        "citeAs": "https://mlcommons.org/croissant/citeAs/",
        "conformsTo": "https://mlcommons.org/croissant/conformsTo/",
        "@language": "en",
        "repeated": "https://mlcommons.org/croissant/repeated/",
        "field": "https://mlcommons.org/croissant/field/",
        "examples": "https://mlcommons.org/croissant/examples/",
        "recordSet": "https://mlcommons.org/croissant/recordSet/",
        "fileObject": "https://mlcommons.org/croissant/fileObject/",
        "fileSet": "https://mlcommons.org/croissant/fileSet/",
        "source": "https://mlcommons.org/croissant/source/",
        "references": "https://mlcommons.org/croissant/references/",
        "key": "https://mlcommons.org/croissant/key/",
        "parentField": "https://mlcommons.org/croissant/parentField/",
        "isLiveDataset": "https://mlcommons.org/croissant/isLiveDataset/",
        "separator": "https://mlcommons.org/croissant/separator/",
        "extract": "https://mlcommons.org/croissant/extract/",
        "subField": "https://mlcommons.org/croissant/subField/",
        "regex": "https://mlcommons.org/croissant/regex/",
        "column": "https://mlcommons.org/croissant/column/",
        "path": "https://mlcommons.org/croissant/path/",
        "fileProperty": "https://mlcommons.org/croissant/fileProperty/",
        "md5": "https://mlcommons.org/croissant/md5/",
        "jsonPath": "https://mlcommons.org/croissant/jsonPath/",
        "transform": "https://mlcommons.org/croissant/transform/",
        "replace": "https://mlcommons.org/croissant/replace/",
        "dataType": "https://mlcommons.org/croissant/dataType/",
        "includes": "https://mlcommons.org/croissant/includes/",
        "excludes": "https://mlcommons.org/croissant/excludes/",
    }

interactive(file_path=None, prefill_metadata=None, staged=False, incomplete=False)

Interactive annotation process for files.

Source code in biotope/commands/annotate.py

@annotate.command()
@click.option(
    "--file-path",
    "-f",
    type=click.Path(exists=True),
    help="Path to the file to annotate",
)
@click.option(
    "--prefill-metadata",
    "-p",
    type=str,
    help="JSON string containing pre-filled metadata",
)
@click.option(
    "--staged",
    "-s",
    is_flag=True,
    help="Annotate all staged files",
)
@click.option(
    "--incomplete",
    "-i",
    is_flag=True,
    help="Annotate all tracked files with incomplete metadata",
)
def interactive(file_path: str | None = None, prefill_metadata: str | None = None, staged: bool = False, incomplete: bool = False) -> None:
    """Interactive annotation process for files."""
    console = Console()

    # Initialize metadata with pre-filled values if provided
    dynamic_metadata = json.loads(prefill_metadata) if prefill_metadata else {}

    # Load project-level metadata for pre-filling if we're in a biotope project
    biotope_root = find_biotope_root()
    if biotope_root:
        from biotope.utils import load_project_metadata
        project_metadata = load_project_metadata(biotope_root)

        # Merge project metadata with any provided prefill metadata
        # Project metadata takes precedence for common fields
        for key, value in project_metadata.items():
            if key not in dynamic_metadata:
                dynamic_metadata[key] = value

    # Merge with standard context and structure
    metadata = merge_metadata(dynamic_metadata)

    # Handle staged files
    if staged:
        if not biotope_root:
            click.echo("❌ Not in a biotope project. Run 'biotope init' first.")
            raise click.Abort

        staged_files = get_staged_files(biotope_root)
        if not staged_files:
            click.echo("❌ No files staged. Use 'biotope add <file>' first.")
            raise click.Abort

        console.print(f"[bold blue]Annotating {len(staged_files)} staged file(s)[/]")

        for i, file_info in enumerate(staged_files):
            file_path = biotope_root / file_info["file_path"]
            console.print(f"\n[bold green]File {i+1}/{len(staged_files)}: {file_path.name}[/]")

            # Find the existing metadata file for this data file
            datasets_dir = biotope_root / ".biotope" / "datasets"
            relative_path = file_path.relative_to(biotope_root)
            metadata_file = datasets_dir / relative_path.with_suffix('.jsonld')

            # Check if metadata file exists
            if metadata_file.exists():
                # Load existing metadata to pre-fill
                try:
                    with open(metadata_file) as f:
                        existing_metadata = json.load(f)
                except (json.JSONDecodeError, IOError):
                    existing_metadata = {}

                # Extract file information from existing metadata
                file_metadata = {
                    "name": existing_metadata.get("name", file_path.stem),
                    "description": existing_metadata.get("description", f"Dataset for {file_path.name}"),
                    "distribution": existing_metadata.get("distribution", [])
                }

                # Merge with project metadata
                if biotope_root:
                    from biotope.utils import load_project_metadata
                    project_metadata = load_project_metadata(biotope_root)
                    for key, value in project_metadata.items():
                        if key not in file_metadata and key not in existing_metadata:
                            file_metadata[key] = value

                # Run interactive annotation for this file (updating existing)
                _run_interactive_annotation(console, metadata_file, file_metadata, biotope_root, update_existing=True)
            else:
                # Pre-fill with file information for new metadata
                file_metadata = {
                    "name": file_path.stem,
                    "description": f"Dataset for {file_path.name}",
                    "distribution": [
                        {
                            "@type": "sc:FileObject",
                            "@id": f"file_{file_info['sha256'][:8]}",
                            "name": file_path.name,
                            "contentUrl": str(file_path.relative_to(biotope_root)),
                            "sha256": file_info["sha256"],
                            "contentSize": file_info["size"]
                        }
                    ]
                }

                # Merge with project metadata
                if biotope_root:
                    from biotope.utils import load_project_metadata
                    project_metadata = load_project_metadata(biotope_root)
                    for key, value in project_metadata.items():
                        if key not in file_metadata:
                            file_metadata[key] = value

                # Run interactive annotation for this file (creating new)
                _run_interactive_annotation(console, file_path, file_metadata, biotope_root)

        return

    # Handle incomplete files
    if incomplete:
        if not biotope_root:
            click.echo("❌ Not in a biotope project. Run 'biotope init' first.")
            raise click.Abort

        # Get all tracked files and check their annotation status
        from biotope.validation import get_all_tracked_files, get_annotation_status_for_files

        tracked_files = get_all_tracked_files(biotope_root)
        if not tracked_files:
            click.echo("❌ No tracked files found. Use 'biotope add <file>' first.")
            raise click.Abort

        annotation_status = get_annotation_status_for_files(biotope_root, tracked_files)
        incomplete_files = [
            file_path for file_path, (is_annotated, _) in annotation_status.items() 
            if not is_annotated
        ]

        if not incomplete_files:
            click.echo("✅ All tracked files are properly annotated!")
            return

        console.print(f"[bold blue]Found {len(incomplete_files)} file(s) with incomplete annotation[/]")

        for i, file_path in enumerate(incomplete_files):
            metadata_file = biotope_root / file_path
            console.print(f"\n[bold green]File {i+1}/{len(incomplete_files)}: {metadata_file.stem}[/]")

            # Load existing metadata to pre-fill
            try:
                with open(metadata_file) as f:
                    existing_metadata = json.load(f)
            except (json.JSONDecodeError, IOError):
                existing_metadata = {}

            # Extract file information from existing metadata
            file_info = {
                "name": existing_metadata.get("name", metadata_file.stem),
                "description": existing_metadata.get("description", f"Dataset for {metadata_file.stem}"),
                "distribution": existing_metadata.get("distribution", [])
            }

            # Merge with project metadata for missing fields
            if biotope_root:
                from biotope.utils import load_project_metadata
                project_metadata = load_project_metadata(biotope_root)
                for key, value in project_metadata.items():
                    if key not in file_info and key not in existing_metadata:
                        file_info[key] = value

            # Run interactive annotation for this file (updating existing)
            _run_interactive_annotation(console, metadata_file, file_info, biotope_root, update_existing=True)

        return

    # If file path is provided, use it
    if file_path:
        metadata["file_path"] = file_path

    # Create a nice header
    console.print(
        Panel(
            "[bold blue]Biotope Dataset Metadata Creator[/]",
            subtitle="Create scientific dataset metadata in Croissant format",
        ),
    )

    console.print(Markdown("This wizard will help you document your scientific dataset with standardized metadata."))
    console.print()

    # Show project metadata info if available
    if biotope_root:
        from biotope.utils import load_project_metadata
        project_metadata = load_project_metadata(biotope_root)
        if project_metadata:
            console.print("[bold green]Project Metadata Available[/]")
            console.print("─" * 50)
            console.print("The following project-level metadata will be used as defaults:")

            table = Table(show_header=False)
            table.add_column("Field", style="cyan")
            table.add_column("Value", style="green")

            for key, value in project_metadata.items():
                if key == "creator" and isinstance(value, dict):
                    display_value = value.get("name", str(value))
                else:
                    display_value = str(value)
                table.add_row(key, display_value)

            console.print(table)
            console.print()

    # Section: Basic Information
    console.print("[bold green]Basic Dataset Information[/]")
    console.print("─" * 50)

    # Use pre-filled name if available, otherwise prompt
    dataset_name = metadata.get("name", "")
    if not dataset_name:
        dataset_name = click.prompt(
            "Dataset name (a short, descriptive title; no spaces allowed)",
            default="",
        )
    else:
        dataset_name = click.prompt(
            "Dataset name (a short, descriptive title; no spaces allowed)",
            default=dataset_name,
        )

    description = click.prompt(
        "Dataset description (what does this dataset contain and what is it used for?)",
        default=metadata.get("description", ""),
    )

    # Section: Source Information
    console.print("\n[bold green]Data Source Information[/]")
    console.print("─" * 50)
    console.print("Where did this data come from? (e.g., a URL, database name, or experiment)")
    data_source = click.prompt("Data source", default=metadata.get("url", ""))

    # Section: Ownership and Dates
    console.print("\n[bold green]Ownership and Dates[/]")
    console.print("─" * 50)

    project_name = click.prompt(
        "Project name",
        default=metadata.get("cr:projectName", Path.cwd().name),
    )

    contact = click.prompt(
        "Contact person (email preferred)",
        default=metadata.get("creator", {}).get("name", getpass.getuser()),
    )

    date = click.prompt(
        "Creation date (YYYY-MM-DD)",
        default=metadata.get("dateCreated", datetime.date.today().isoformat()),
    )

    # Section: Access Information
    console.print("\n[bold green]Access Information[/]")
    console.print("─" * 50)

    # Create a table for examples
    table = Table(title="Access Restriction Examples")
    table.add_column("Type", style="cyan")
    table.add_column("Description", style="green")
    table.add_row("Public", "Anyone can access and use the data")
    table.add_row("Academic", "Restricted to academic/research use only")
    table.add_row("Approval", "Requires explicit approval from data owner")
    table.add_row("Embargo", "Will become public after a specific date")
    console.print(table)

    has_access_restrictions = Confirm.ask(
        "Does this dataset have access restrictions?",
        default=bool(metadata.get("cr:accessRestrictions")),
    )

    access_restrictions = None
    if has_access_restrictions:
        access_restrictions = Prompt.ask(
            "Please describe the access restrictions",
            default=metadata.get("cr:accessRestrictions", ""),
        )
        if not access_restrictions.strip():
            access_restrictions = None

    # Section: Additional Information
    console.print("\n[bold green]Additional Information[/]")
    console.print("─" * 50)
    console.print("[italic]The following fields are optional but recommended for scientific datasets[/]")

    # Get default format from distribution if available
    default_format = ""
    distribution = metadata.get("distribution", [])
    if distribution and len(distribution) > 0:
        default_format = distribution[0].get("encodingFormat", "")

    format = click.prompt(
        "File format (MIME type, e.g., text/csv, application/json, application/x-hdf5, application/fastq)",
        default=metadata.get("encodingFormat")
        or metadata.get("format")
        or default_format,
    )

    legal_obligations = click.prompt(
        "Legal obligations (e.g., citation requirements, licenses)",
        default=metadata.get("cr:legalObligations", ""),
    )

    collaboration_partner = click.prompt(
        "Collaboration partner and institute",
        default=metadata.get("cr:collaborationPartner", ""),
    )

    # Section: Publication Information
    console.print("\n[bold green]Publication Information[/]")
    console.print("─" * 50)
    console.print("[italic]The following fields are recommended for proper dataset citation[/]")

    publication_date = click.prompt(
        "Publication date (YYYY-MM-DD)",
        default=metadata.get("datePublished", date),  # Use creation date as default
    )

    version = click.prompt(
        "Dataset version",
        default=metadata.get("version", "1.0"),
    )

    license_url = click.prompt(
        "License URL",
        default=metadata.get("license", "https://creativecommons.org/licenses/by/4.0/"),
    )

    citation = click.prompt(
        "Citation text",
        default=metadata.get("citation", f"Please cite this dataset as: {dataset_name} ({date.split('-')[0]})"),
    )

    # Update metadata with new values while preserving any existing fields
    new_metadata = {
        "@context": get_standard_context(),  # Use the standard context
        "@type": "Dataset",
        "name": dataset_name,
        "description": description,
        "url": data_source,
        "creator": {
            "@type": "Person",
            "name": contact,
        },
        "dateCreated": date,
        "cr:projectName": project_name,
        "datePublished": publication_date,
        "version": version,
        "license": license_url,
        "citation": citation,
    }

    # Only add access restrictions if they exist
    if access_restrictions:
        new_metadata["cr:accessRestrictions"] = access_restrictions

    # Add optional fields if provided
    if format:
        new_metadata["encodingFormat"] = format
    if legal_obligations:
        new_metadata["cr:legalObligations"] = legal_obligations
    if collaboration_partner:
        new_metadata["cr:collaborationPartner"] = collaboration_partner

    # Update metadata while preserving pre-filled values
    for key, value in new_metadata.items():
        if key not in ["distribution"]:  # Don't overwrite distribution
            metadata[key] = value

    # Initialize distribution array for FileObjects/FileSets if it doesn't exist
    if "distribution" not in metadata:
        metadata["distribution"] = []

    # Section: File Resources
    console.print("\n[bold green]File Resources[/]")
    console.print("─" * 50)
    console.print("Croissant datasets can include file resources (FileObject) and file collections (FileSet).")

    # If we have pre-filled distribution, use it
    if prefill_metadata and "distribution" in dynamic_metadata:
        # Create a table to display pre-filled file information
        table = Table(title="Pre-filled File Resources")
        table.add_column("Type", style="cyan")
        table.add_column("Name", style="green")
        table.add_column("Format", style="yellow")
        table.add_column("Hash", style="magenta")

        for resource in dynamic_metadata["distribution"]:
            resource_type = resource.get("@type", "").replace("sc:", "").replace("cr:", "")
            name = resource.get("name", "")
            format = resource.get("encodingFormat", "")
            hash = resource.get("sha256", "")[:8] + "..." if resource.get("sha256") else ""

            table.add_row(resource_type, name, format, hash)

        console.print(table)

        if click.confirm("Would you like to use these pre-filled file resources?", default=True):
            metadata["distribution"] = dynamic_metadata["distribution"]
            console.print("[bold green]Using pre-filled file resources[/]")
        else:
            console.print("[yellow]You can now add new file resources manually[/]")
            metadata["distribution"] = []
    elif click.confirm("Would you like to add file resources to your dataset?", default=True):
        while True:
            resource_type = click.prompt(
                "Resource type",
                type=click.Choice(["FileObject", "FileSet"]),
                default="FileObject",
            )

            if resource_type == "FileObject":
                file_id = click.prompt("File ID (unique identifier for this file)")
                file_name = click.prompt("File name (including extension)")
                content_url = click.prompt("Content URL (where the file can be accessed)")
                encoding_format = click.prompt(
                    "Encoding format (MIME type, e.g., text/csv, application/json, application/x-hdf5, application/fastq)",
                )

                file_object = {
                    "@type": "sc:FileObject",
                    "@id": file_id,
                    "name": file_name,
                    "contentUrl": content_url,
                    "encodingFormat": encoding_format,
                    "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
                }

                # Optional SHA256 checksum
                if click.confirm("Add SHA256 checksum?", default=False):
                    sha256 = click.prompt("SHA256 checksum")
                    file_object["sha256"] = sha256

                # Optional containedIn property
                if click.confirm("Is this file contained in another file (e.g., in an archive)?", default=False):
                    container_id = click.prompt("Container file ID")
                    file_object["containedIn"] = {"@id": container_id}

                metadata["distribution"].append(file_object)

            else:  # FileSet
                fileset_id = click.prompt("FileSet ID (unique identifier for this file set)")

                # Container information
                container_id = click.prompt("Container file ID (archive or directory)")

                fileset = {
                    "@type": "cr:FileSet",
                    "@id": fileset_id,
                    "containedIn": {"@id": container_id},
                }

                # File pattern information
                encoding_format = click.prompt(
                    "Encoding format of files in this set (MIME type, e.g., text/csv, application/json, application/x-hdf5, application/fastq)",
                    default="",
                )
                if encoding_format:
                    fileset["encodingFormat"] = encoding_format

                includes_pattern = click.prompt("Include pattern (e.g., *.jpg, data/*.csv)", default="")
                if includes_pattern:
                    fileset["includes"] = includes_pattern

                # Optional exclude pattern
                if click.confirm("Add exclude pattern?", default=False):
                    excludes_pattern = click.prompt("Exclude pattern")
                    fileset["excludes"] = excludes_pattern

                metadata["distribution"].append(fileset)

            if not click.confirm("Add another file resource?", default=False):
                break

    # Section: Data Structure
    console.print("\n[bold green]Data Structure[/]")
    console.print("─" * 50)

    # Create a table for record set examples
    table = Table(title="Record Set Examples")
    table.add_column("Dataset Type", style="cyan")
    table.add_column("Record Sets", style="green")
    table.add_row("Genomics", "patients, samples, gene_expressions")
    table.add_row("Climate", "locations, time_series, measurements")
    table.add_row("Medical", "patients, visits, treatments, outcomes")
    console.print(table)

    console.print("Record sets describe the structure of your data.")

    if click.confirm("Would you like to add a record set to describe your data structure?", default=True):
        metadata["cr:recordSet"] = []

        while True:
            record_set_name = click.prompt("Record set name (e.g., 'patients', 'samples')")
            record_set_description = click.prompt(f"Description of the '{record_set_name}' record set", default="")

            # Create record set with proper Croissant format
            record_set = {
                "@type": "cr:RecordSet",
                "@id": f"#{record_set_name}",
                "name": record_set_name,
                "description": record_set_description,
            }

            # Ask about data type
            if click.confirm(
                f"Would you like to specify a data type for the '{record_set_name}' record set?",
                default=False,
            ):
                data_type = click.prompt("Data type (e.g., sc:GeoCoordinates, sc:Person)")
                record_set["dataType"] = data_type

            # Ask about fields with examples
            console.print(f"\n[bold]Fields in '{record_set_name}'[/]")
            console.print("Fields describe the data columns or attributes in this record set.")

            if click.confirm(f"Would you like to add fields to the '{record_set_name}' record set?", default=True):
                record_set["cr:field"] = []

                while True:
                    field_name = click.prompt("Field name (column or attribute name)")
                    field_description = click.prompt(f"Description of '{field_name}'", default="")

                    # Create field with proper Croissant format
                    field = {
                        "@type": "cr:Field",
                        "@id": f"#{record_set_name}/{field_name}",
                        "name": field_name,
                        "description": field_description,
                    }

                    # Ask about data type
                    if click.confirm(
                        f"Would you like to specify a data type for the '{field_name}' field?",
                        default=False,
                    ):
                        data_type = click.prompt("Data type (e.g., sc:Text, sc:Integer, sc:Float, sc:ImageObject)")
                        field["dataType"] = data_type

                    # Ask about source
                    if click.confirm(
                        f"Would you like to specify a data source for the '{field_name}' field?",
                        default=False,
                    ):
                        source_type = click.prompt(
                            "Source type",
                            type=click.Choice(["FileObject", "FileSet"]),
                            default="FileObject",
                        )
                        source_id = click.prompt(f"{source_type} ID")

                        source = {"source": {}}
                        if source_type == "FileObject":
                            source["source"]["fileObject"] = {"@id": source_id}
                        else:
                            source["source"]["fileSet"] = {"@id": source_id}

                        # Ask about extraction method
                        extract_type = click.prompt(
                            "Extraction method",
                            type=click.Choice(["column", "jsonPath", "fileProperty", "none"]),
                            default="none",
                        )

                        if extract_type != "none":
                            source["source"]["extract"] = {}
                            if extract_type == "column":
                                column_name = click.prompt("Column name")
                                source["source"]["extract"]["column"] = column_name
                            elif extract_type == "jsonPath":
                                json_path = click.prompt("JSONPath expression")
                                source["source"]["extract"]["jsonPath"] = json_path
                            elif extract_type == "fileProperty":
                                file_property = click.prompt(
                                    "File property",
                                    type=click.Choice(["fullpath", "filename", "content", "lines", "lineNumbers"]),
                                )
                                source["source"]["extract"]["fileProperty"] = file_property

                        # Add source to field
                        for key, value in source["source"].items():
                            field[key] = value

                    # Ask if the field is repeated (array)
                    if click.confirm(f"Is '{field_name}' a repeated field (array/list)?", default=False):
                        field["repeated"] = True

                    # Ask if the field references another field
                    if click.confirm(f"Does '{field_name}' reference another field (foreign key)?", default=False):
                        ref_record_set = click.prompt("Referenced record set name")
                        ref_field = click.prompt("Referenced field name")
                        field["references"] = {"@id": f"#{ref_record_set}/{ref_field}"}

                    # Add field to record set
                    record_set["cr:field"].append(field)

                    if not click.confirm("Add another field?", default=True):
                        break

            # Ask about key fields
            if click.confirm(
                f"Would you like to specify key fields for the '{record_set_name}' record set?",
                default=False,
            ):
                record_set["key"] = []
                while True:
                    key_field = click.prompt("Key field name")
                    record_set["key"].append({"@id": f"#{record_set_name}/{key_field}"})

                    if not click.confirm("Add another key field?", default=False):
                        break

            # Add record set to metadata
            metadata["cr:recordSet"].append(record_set)

            if not click.confirm("Add another record set?", default=False):
                break

    # Save metadata with a suggested filename
    default_filename = f"{dataset_name.lower().replace(' ', '_')}_metadata.json"
    output_path = click.prompt("Output file path", default=default_filename)

    with open(output_path, "w") as f:
        json.dump(metadata, f, indent=2)

    # Stage the changes in Git if we're in a biotope project
    try:
        biotope_root = find_biotope_root()
        if biotope_root:
            import subprocess
            subprocess.run(
                ["git", "add", ".biotope/"],
                cwd=biotope_root,
                check=True
            )
            console.print(f"✅ Staged changes in Git")
    except (subprocess.CalledProcessError, FileNotFoundError):
        pass  # Not in a biotope project or Git not available

    # Final success message with rich formatting
    console.print()
    console.print(
        Panel(
            f"[bold green]✅ Created Croissant metadata file at:[/] [blue]{output_path}[/]",
            title="Success",
            border_style="green",
        ),
    )

    console.print("[italic]Validate this file with:[/]")
    console.print(f"[bold yellow]biotope annotate validate --jsonld {output_path}[/]")

load(jsonld, record_set, num_records)

Load records from a dataset using its Croissant metadata.

Source code in biotope/commands/annotate.py

@annotate.command()
@click.option(
    "--jsonld",
    "-j",
    type=click.Path(exists=True),
    required=True,
    help="Path to the JSON-LD metadata file.",
)
@click.option(
    "--record-set",
    "-r",
    required=True,
    help="Name of the record set to load.",
)
@click.option(
    "--num-records",
    "-n",
    type=int,
    default=10,
    help="Number of records to load.",
)
def load(jsonld, record_set, num_records):
    """Load records from a dataset using its Croissant metadata."""
    try:
        # Use mlcroissant CLI to load the dataset
        result = subprocess.run(
            [
                "mlcroissant",
                "load",
                "--jsonld",
                jsonld,
                "--record_set",
                record_set,
                "--num_records",
                str(num_records),
            ],
            capture_output=True,
            text=True,
            check=True,
        )

        # Display the output
        if result.stdout:
            click.echo(result.stdout)

        click.echo(f"Loaded {num_records} records from record set '{record_set}'")
    except subprocess.CalledProcessError as e:
        click.echo(f"Error loading dataset: {e.stderr}", err=True)
        exit(1)
    except Exception as e:
        click.echo(f"Error running load command: {e!s}", err=True)
        exit(1)

merge_metadata(dynamic_metadata)

Merge dynamic metadata with standard context and structure.

Source code in biotope/commands/annotate.py

def merge_metadata(dynamic_metadata: dict) -> dict:
    """Merge dynamic metadata with standard context and structure."""
    # Start with standard context
    metadata = {
        "@context": get_standard_context(),
        "@type": "Dataset",
    }

    # Update with dynamic content
    metadata.update(dynamic_metadata)

    return metadata

validate(jsonld)

Validate a Croissant metadata file.

Source code in biotope/commands/annotate.py

@annotate.command()
@click.option(
    "--jsonld",
    "-j",
    type=click.Path(exists=True),
    required=True,
    help="Path to the JSON-LD metadata file to validate.",
)
def validate(jsonld):
    """Validate a Croissant metadata file."""
    try:
        # Use mlcroissant CLI to validate the file
        result = subprocess.run(
            ["mlcroissant", "validate", "--jsonld", jsonld],
            capture_output=True,
            text=True,
            check=True,
        )
        click.echo("Validation successful! The metadata file is valid.")
        if result.stdout:
            # Filter out informational log messages
            filtered_output = "\n".join(
                line for line in result.stdout.splitlines() if not line.startswith("I") or not line.endswith("Done.")
            )
            if filtered_output:
                click.echo(f"Output: {filtered_output}")
        if result.stderr:
            # Filter out informational log messages
            filtered_stderr = "\n".join(
                line for line in result.stderr.splitlines() if not line.startswith("I") or not line.endswith("Done.")
            )
            if filtered_stderr:
                click.echo(f"Warnings: {filtered_stderr}")
    except subprocess.CalledProcessError as e:
        click.echo(f"Validation failed: {e.stderr}", err=True)
        exit(1)
    except Exception as e:
        click.echo(f"Error running validation: {e!s}", err=True)
        exit(1)