Skip to main content

Module ingest_pipeline

Module ingest_pipeline 

Source
Expand description

High-level ingestion pipeline — workspace-wide self-ingest and graph coherence verification.

Provides:

  • ingest_workspace(): walk all crates in a Cargo workspace and ingest them
  • verify_graph(): check graph coherence (no dangling edges, no duplicate IDs)
  • write_graph_parquet(): persist CodeNode + CodeEdge batches to Parquet files

Structs§

GraphViolations
Graph coherence violations found by verify_graph().
WorkspaceIngestResult
Workspace-wide ingestion result.

Functions§

discover_workspace_crates
Discover all crate directories in a Cargo workspace.
ingest_workspace
Ingest all crates in a Cargo workspace.
load_nodes_from_parquet
Load CodeNodes from a Parquet graph directory produced by write_graph_parquet().
verify_graph
Verify graph coherence of merged CodeNode + CodeEdge RecordBatches.
write_graph_parquet
Write CodeNodes and CodeEdges RecordBatches to Parquet files.