Struct ScriptState 

pub struct ScriptState {
    driver: Driver,
    client: PersistClient,
    loc: PersistLocation,
    schemas: BTreeMap<String, RelationDesc>,
    shards: BTreeMap<String, ShardId>,
    indexes: BTreeMap<u64, IndexEntry>,
    mv_outputs: BTreeMap<u64, CollectionMetadata>,
    next_internal: u64,
}

Mutable state threaded through a script run.

Fields

driver: Driver

client: PersistClient

loc: PersistLocation

schemas: BTreeMap<String, RelationDesc>

Named schemas declared via define_schema.

shards: BTreeMap<String, ShardId>

Alias-to-shard map; aliases are allocated lazily on first use.

indexes: BTreeMap<u64, IndexEntry>

Exported indexes, by global id, for later import / count assertions.

mv_outputs: BTreeMap<u64, CollectionMetadata>

Materialized-view sink outputs, by sink global id: the target shard’s metadata, so a peek of the sink id reads its shard via a persist peek.

next_internal: u64

Next ephemeral id for the count sugar’s dataflows.

Implementations

impl ScriptState

pub async fn new(driver: Driver, loc: PersistLocation) -> Result<Self>

Build the state from a connected driver and its persist location, opening a persist client.

fn shard_id(&mut self, alias: &str) -> ShardId

Resolve a shard alias, allocating a fresh ShardId on first use.

fn alloc_internal(&mut self) -> GlobalId

Allocate a fresh ephemeral global id for an internally-built dataflow.

async fn count_via_reduce(&mut self, index_id: u64, ts: u64) -> Result<u64>

Count the rows of a registered index at ts by running a count(*) Reduce over it: build an ephemeral dataflow that index-imports index_id, schedule and hydrate it, then peek its single-row output. An empty result (the reduce emits no row over empty input) reads as a count of 0.

fn resolve_schema(&self, name: &Option<String>) -> Result<RelationDesc>

Resolve a schema name, defaulting to the built-in sample schema when absent.

fn check_sink_schema( &self, builder: &DataflowBuilder, on_id: u64, desc: &RelationDesc, ) -> Result<()>

Validate that a sink’s declared output schema desc matches the column types of the object on_id it exports. Compares column types (not inferred keys), so a mismatched arity or type fails before the dataflow is submitted.

pub async fn execute(&mut self, cmd: Command) -> Result<String>

Execute a single command, returning its golden output text.