Crate stt_optimize 

stt-optimize as a library.

stt-build --auto calls into recommend_for to pick zoom levels and a temporal bucket from an input file before building. The stt-optimize CLI (a binary in the spatiotemporal-tiles facade crate) is a thin wrapper around the same functions.

Re-exports

pub use analysis::inspect::InspectReport;

pub use diff::DiffReport;

pub use doctor::DoctorReport;

pub use order_audit::OrderAuditReport;

pub use loader::DataSource;

pub use packed::PackedTileset;

pub use recommend::Recommendations;

Modules

advisors

Evidence-based flag advisors: each module inspects the analysis result (and, where useful, trial-encodes the loader sample through the real encoder) and returns Advice entries for stt-build flags beyond the zoom/bucket basics.

analysis

Analysis modules for spatiotemporal data

diff

Tileset-to-tileset comparison for stt-optimize diff: totals, per-zoom directory stats, and per-column costs between two InspectReports (typically before/after a re-encode or fleet reprocess), each as absolute and percent deltas.

doctor

Tileset doctor: severity-ranked findings over a built (packed) tileset, each with a concrete remediation flag and a projection derived from the inspect report’s measured per-column costs.

loader

Data loading for GeoParquet files

measure

Sample-encode measurement: push a deterministic sample of source features through the real stt-core encoder (+ zstd) to replace formula-based size estimates with measured bytes.

order_audit

Blob-ordering audit over a built archive: measure per-ordering range-read cost and recommend --blob-ordering.

packed

Packed-tileset reading for post-build analysis (stt-optimize inspect/diff).

recommend

Parameter recommendation engine

report

Report generation (text and JSON formats)

Functions

recommend_for

Analyze a GeoParquet or STT input and produce build recommendations.