Struct MetadataExtractor 

pub struct MetadataExtractor;

Metadata extractor for resource files.

Extracts dependency information embedded in resource files:

• Markdown files (.md): YAML frontmatter between --- delimiters
• JSON files (.json): dependencies field in the JSON structure
• Other files: No dependencies supported

Implementations

impl MetadataExtractor

pub fn extract( path: &Path, content: &str, variant_inputs: Option<&Value>, context: Option<&OperationContext>, ) -> Result<DependencyMetadata>

Extract dependency metadata from a file’s content.

Uses operation-scoped context for warning deduplication when provided.

Arguments

• path - Path to the file (used to determine file type)
• content - Content of the file
• variant_inputs - Optional template variables (contains project config and any overrides)
• context - Optional operation context for warning deduplication

Returns

• DependencyMetadata - Extracted metadata (may be empty)

Template Support

If variant_inputs is provided, frontmatter is rendered as a Tera template before parsing, allowing references like: {{ project.language }} or {{ config.model }}

Examples

use agpm_cli::core::OperationContext;
use agpm_cli::metadata::MetadataExtractor;
use std::path::Path;

let ctx = OperationContext::new();
let path = Path::new("agent.md");
let content = "---\ndependencies:\n  agents:\n    - path: helper.md\n---\n# Agent";

let metadata = MetadataExtractor::extract(
    path,
    content,
    None,
    Some(&ctx)
).unwrap();

pub fn extract_auto(content: &str) -> Result<DependencyMetadata>

Extract metadata from file content without knowing the file type.

Tries to detect the format automatically.