Struct LakehouseContext 

pub struct LakehouseContext { /* private fields */ }

Bundles all runtime resources needed for lakehouse query execution.

This struct holds the data lake connection, metadata cache, and DataFusion runtime, providing a single context object that can be passed through the query path. Parquet byte-range caching is handled by the in-process L1 cache wrapped around the reader factory’s object store (see object_cache::l1_wrap), not by this struct.

Implementations

impl LakehouseContext

pub async fn from_connection( lake: Arc<DataLakeConnection>, ) -> Result<Arc<LakehouseContext>, Error>

Builds a lakehouse context from an already-connected DataLakeConnection.

Runs migrate_lakehouse on the supplied connection (idempotent) and creates the DataFusion runtime. The caller is responsible for running migrate_db (ingestion schema) before this if both migrations are needed — the monolith does this via connect_to_remote_data_lake.

pub async fn from_env( policy: WritablePolicy, ) -> Result<Arc<LakehouseContext>, Error>

Reads MICROMEGAS_SQL_CONNECTION_STRING and MICROMEGAS_OBJECT_STORE_URI, connects to the data lake under policy, runs lakehouse migrations, and creates the runtime environment.

pub fn new( lake: Arc<DataLakeConnection>, runtime: Arc<RuntimeEnv>, ) -> Result<LakehouseContext, Error>

Creates a new lakehouse context with a default-sized metadata cache.

Fails if MICROMEGAS_DEFAULT_AUDIENCE is set to a malformed value – every caller is a startup path, so a typo in that knob stops the role rather than silently relabelling legacy data (see default_audience_from_env).

pub fn with_caches( lake: Arc<DataLakeConnection>, runtime: Arc<RuntimeEnv>, metadata_cache: Arc<MetadataCache>, default_audience: Arc<str>, ) -> LakehouseContext

Creates a new lakehouse context with a custom metadata cache.

Takes default_audience rather than re-reading the environment: a context built this way must resolve the same default as the one it is derived from (see the field’s doc comment).

pub fn default_audience(&self) -> Arc<str>

Returns the audience a never-stamped process is read as (MICROMEGAS_DEFAULT_AUDIENCE).

pub fn lake(&self) -> &Arc<DataLakeConnection>

Returns the data lake connection.

pub fn metadata_cache(&self) -> &Arc<MetadataCache>

Returns the metadata cache.

pub fn runtime(&self) -> &Arc<RuntimeEnv>

Returns the DataFusion runtime environment.

pub fn with_runtime(&self, runtime: Arc<RuntimeEnv>) -> Arc<LakehouseContext>

Clones this context with runtime swapped, sharing the metadata cache and reader factory. Deliberately a struct-update clone rather than a call to new()/ with_caches() — those rebuild the MetadataCache and ReaderFactory, which would throw away the shared metadata cache per query.

pub fn reader_factory(&self) -> &Arc<ReaderFactory>

Returns the shared ReaderFactory.

pub fn audience_index(&self) -> &Arc<AudienceIndex>

Returns the shared call-level-guard audience index.

pub fn query_denials(&self) -> &Arc<QueryDenyList>

Returns the shared query deny list.

Trait Implementations

impl Clone for LakehouseContext

fn clone(&self) -> LakehouseContext

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LakehouseContext

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

Formats the value using the given formatter.