Data models

The crader.models module defines dataclasses used across the pipeline.

Core entities

• Repository: repository identity and snapshot pointer.
• Snapshot: immutable version tied to a commit hash.
• FileRecord: metadata for a file in a snapshot, including parse status.
• ChunkNode: code chunk with byte range, line range, and metadata.
• ChunkContent: content-addressable storage for chunk text.
• CodeRelation: directed edge between two chunks.
• ParsingResult: aggregate container for parser output.

Retrieval

• RetrievedContext: enriched search result returned by CodeRetriever.
• Includes content, scores, semantic labels, and navigation hints.
• render() returns a Markdown-formatted string.