This visualisation uses AI Generated code, finetuned for the best visualisation, not code quality

Interactive C + Raylib visualizer for understanding approximate nearest-neighbour search through the FAISS pipeline, from coarse clustering and inverted lists to product quantization, LUT scoring, and reranking.

• Coarse partition training and centroid assignment
• IVF list construction and query-time nprobe search
• Product quantization and query lookup-table scoring
• Approximate shortlist generation followed by exact reranking
• A final side-by-side comparison of FAISS-style and ScaNN-style pipelines

flowchart LR
    A["Dataset Vectors"]
    B["Coarse k-means"]
    C["IVF Inverted Lists"]
    D["PQ Codes"]
    E["Query LUT / Approx Scan"]
    F["Refine + Rerank"]

    A --> B
    B --> C
    C --> D
    D --> E
    E --> F

• q: quit
• Page-specific controls are shown in the footer of each stage
• Query points, centroids, and step-through states can be dragged or advanced on the relevant pages

make run