Struct EncoderConfig 

pub struct EncoderConfig {
    pub quantize_coords_m: Option<f64>,
    pub quantize_attrs: HashMap<String, f64>,
    pub quantize_attrs_auto: bool,
    pub vector_groups: Vec<VectorGroup>,
    pub point_elevation_column: String,
    pub vertex_time_max_step_ms: u32,
    pub format_version: u32,
    pub template_collector: Option<Arc<TemplateCollector>>,
}

Resolved, explicit encoder settings — the values the tile encoder reads at encode time (coordinate + attribute quantization, vector grouping, the point-elevation fold, vertex-time precision).

Historically these lived only in process-wide mutable statics set via the set_* functions above; that made a new encode caller silently inherit whatever the last set_* left behind (the origin of the stt-serve parity bug) and made it impossible to encode two different configurations in one process (blocking multi-dataset serve + parallel per-config tests). Passing an EncoderConfig explicitly to encode_tile_with / encode_layer_with removes both problems. The globals + the no-arg encode_tile / encode_layer wrappers remain for the one-shot CLI and existing callers; EncoderConfig::from_globals snapshots them.

Fields

quantize_coords_m: Option<f64>

Fixed-point coordinate quantization ground precision in meters (None = Float64 GeoArrow coordinates, the default).

quantize_attrs: HashMap<String, f64>

Per-property explicit fixed-point precisions (name → precision).

quantize_attrs_auto: bool

Range-adaptive UInt16 quantization for every un-listed Float64 property.

vector_groups: Vec<VectorGroup>

Scalar columns to fuse into interleaved FixedSizeList vector columns.

point_elevation_column: String

Property folded into POINT geometry as the z coordinate (empty = none).

vertex_time_max_step_ms: u32

Ceiling (ms) on the per-vertex time u16-delta quantization step.

format_version: u32

Layer-frame format version (FORMAT_VERSION_V1 | FORMAT_VERSION_V2). The SINGLE branch point between the frozen 0.3.x v1 frame and the sectioned v2 frame (design §1 ★F3): everything upstream of frame assembly + metadata placement is shared. Defaults to v1 so every existing caller stays byte-identical; stt-build passes 2.

template_collector: Option<Arc<TemplateCollector>>

v2 only: when set, layer schemas are hoisted into this collector and frames carry 16-byte template-hash references (the packed-dataset mode — PackWriter::finalize publishes the collected templates in manifest.schemas). None under v2 emits self-contained frames with inline schema sections. Ignored under v1.

Implementations

impl EncoderConfig

pub fn from_globals() -> Self

Snapshot the current process-wide encoder globals into an explicit config. Used by the no-arg encode_tile / encode_layer back-compat wrappers.

Trait Implementations

impl Clone for EncoderConfig

fn clone(&self) -> EncoderConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for EncoderConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for EncoderConfig

fn default() -> Self

Returns the “default value” for a type.