Struct ParseQuery 

pub struct ParseQuery {
    pub opening_filter: OpeningFilterMode,
    pub tessellation_quality: Option<String>,
    pub parquet_layout: ParquetLayout,
    pub stream_shapes: StreamShapes,
    pub data_model_entities: DataModelEntities,
    pub sha256: Option<String>,
}

Query parameters shared by all parse endpoints.

Fields

opening_filter: OpeningFilterMode

Opening filter mode: “default”, “ignore_all”, or “ignore_opaque”.

tessellation_quality: Option<String>

Tessellation detail level (#976): “lowest” | “low” | “medium” | “high” | “highest”. Omitted = “medium” (byte-identical to the historical output — and to what the wasm path produces without setTessellationQuality, keeping client and server meshes in parity).

parquet_layout: ParquetLayout

Flat-Parquet mesh-table layout (#3888): “flat” (default) or “shared-shapes” — see ParquetLayout for why it is opt-in. A query parameter rather than a header because every endpoint it has to reach (both parse routes, the cache check, the cached-geometry fetch) already takes this struct, so the signal travels with the cache identity.

stream_shapes: StreamShapes

Whether POST /api/v1/parse/parquet-stream may share shapes ACROSS batches (#5407): “batch-local” (default) or “cross-batch”, which needs parquet_layout=shared-shapes. See StreamShapes for why it is a second opt-in rather than implied by the layout. Ignored by every other route, and not part of the cache identity.

data_model_entities: DataModelEntities

Which rows the data model’s entities table carries (#6034): “all” (default, every STEP instance) or “rooted” (objects with a GlobalId, plus every instance another data-model table references). See DataModelEntities. Read by every route that writes the data model (/parse/parquet, /parse/parquet/optimized, /parse/parquet-stream) and by /cache/check, because it selects WHICH data-model entry has to exist; the geometry and metadata entries are the same for both.

sha256: Option<String>

SHA-256 of the file the client is asking about, hex, lowercase (#3901).

Read by POST /api/v1/parse/parquet-stream and (since #5128) POST /api/v1/parse/parquet/optimized, and only when the request carries no multipart body: see cached_replay::replay_by_client_hash and parquet_optimized_replay::replay_optimized_by_client_hash for what each does and what it is not allowed to do. It lives on this struct rather than in a header so it travels with the rest of the cache identity (opening_filter, tessellation_quality, parquet_layout) through the one place every parse route already parses. A hash paired with the wrong layout names a different entry, and splitting one identity across two transports is how such pairings drift apart.

Implementations

impl ParseQuery

fn resolved_stream_shapes(&self) -> Result<StreamShapes, ApiError>

The stream’s batch-sharing mode, refusing cross-batch sharing on the flat layout: that layout has no rotation columns and must stay byte-identical to v5, so it cannot share anything, and a client asking for it has misread the contract.

fn resolved_tessellation_quality(&self) -> Result<TessellationQuality, ApiError>

Resolve and validate the requested tessellation level.

Trait Implementations

impl Default for ParseQuery

fn default() -> ParseQuery

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ParseQuery

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.