mentat/query-projector/src/pull.rs

122 lines
3.6 KiB
Rust
Raw Normal View History

// Copyright 2018 Mozilla
//
// Licensed under the Apache License, Version 2.0 (the "License"); you may not use
// this file except in compliance with the License. You may obtain a copy of the
// License at http://www.apache.org/licenses/LICENSE-2.0
// Unless required by applicable law or agreed to in writing, software distributed
// under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
// CONDITIONS OF ANY KIND, either express or implied. See the License for the
// specific language governing permissions and limitations under the License.
use std::collections::{
BTreeMap,
BTreeSet,
};
use mentat_core::{
Binding,
Entid,
Schema,
StructuredMap,
TypedValue,
ValueRc,
};
use mentat_query::{
PullAttributeSpec,
};
use mentat_query_pull::{
Puller,
};
use errors::{
Result,
};
use super::{
Index,
rusqlite,
};
#[derive(Clone, Debug)]
pub(crate) struct PullOperation(pub(crate) Vec<PullAttributeSpec>);
#[derive(Clone, Copy, Debug)]
pub(crate) struct PullIndices {
pub(crate) sql_index: Index, // SQLite column index.
pub(crate) output_index: usize,
}
impl PullIndices {
fn zero() -> PullIndices {
PullIndices {
sql_index: 0,
output_index: 0,
}
}
}
#[derive(Debug)]
pub(crate) struct PullTemplate {
pub(crate) indices: PullIndices,
pub(crate) op: PullOperation,
}
pub(crate) struct PullConsumer<'schema> {
indices: PullIndices,
schema: &'schema Schema,
puller: Puller,
entities: BTreeSet<Entid>,
results: BTreeMap<Entid, ValueRc<StructuredMap>>,
}
impl<'schema> PullConsumer<'schema> {
pub(crate) fn for_puller(puller: Puller, schema: &'schema Schema, indices: PullIndices) -> PullConsumer<'schema> {
PullConsumer {
indices: indices,
schema: schema,
puller: puller,
entities: Default::default(),
results: Default::default(),
}
}
pub(crate) fn for_template(schema: &'schema Schema, template: &PullTemplate) -> Result<PullConsumer<'schema>> {
let puller = Puller::prepare(schema, template.op.0.clone())?;
Ok(PullConsumer::for_puller(puller, schema, template.indices))
}
pub(crate) fn for_operation(schema: &'schema Schema, operation: &PullOperation) -> Result<PullConsumer<'schema>> {
let puller = Puller::prepare(schema, operation.0.clone())?;
Ok(PullConsumer::for_puller(puller, schema, PullIndices::zero()))
}
pub(crate) fn collect_entity<'a, 'stmt>(&mut self, row: &rusqlite::Row<'a, 'stmt>) -> Entid {
let entity = row.get(self.indices.sql_index);
self.entities.insert(entity);
entity
}
pub(crate) fn pull(&mut self, sqlite: &rusqlite::Connection) -> Result<()> {
let entities: Vec<Entid> = self.entities.iter().cloned().collect();
self.results = self.puller.pull(self.schema, sqlite, entities)?;
Ok(())
}
pub(crate) fn expand(&self, bindings: &mut [Binding]) {
if let Binding::Scalar(TypedValue::Ref(id)) = bindings[self.indices.output_index] {
if let Some(pulled) = self.results.get(&id).cloned() {
bindings[self.indices.output_index] = Binding::Map(pulled);
} else {
bindings[self.indices.output_index] = Binding::Map(ValueRc::new(Default::default()));
}
}
}
// TODO: do we need to include empty maps for entities that didn't match any pull?
pub(crate) fn into_coll_results(self) -> Vec<Binding> {
self.results.values().cloned().map(|vrc| Binding::Map(vrc)).collect()
}
}