1. add compute_cost 2. add todo documentation 3. improve type like ta…

…bleid 4. CMSL->S

optd-cost-model/src/common/nodes.rs

@@ -10,6 +10,7 @@ use super::{
     values::Value,
 };
 
+/// TODO: documentation
 #[derive(Debug, Copy, Clone, PartialEq, Eq, Hash)]
 pub enum JoinType {
     Inner = 1,
@@ -23,6 +24,7 @@ pub enum JoinType {
     RightAnti,
 }
 
+/// TODO: documentation
 #[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
 pub enum PhysicalNodeType {
     PhysicalProjection,
@@ -42,6 +44,7 @@ impl std::fmt::Display for PhysicalNodeType {
     }
 }
 
+/// TODO: documentation
 #[derive(Debug, Clone, PartialEq, Eq, Hash)]
 pub enum PredicateType {
     List,
@@ -68,6 +71,7 @@ impl std::fmt::Display for PredicateType {
 
 pub type ArcPredicateNode = Arc<PredicateNode>;
 
+/// TODO: documentation
 #[derive(Clone, Debug, Hash, PartialEq, Eq)]
 pub struct PredicateNode {
     /// A generic predicate node type

optd-cost-model/src/common/predicates/bin_op_pred.rs

@@ -1,3 +1,4 @@
+/// TODO: documentation
 #[derive(Copy, Clone, PartialEq, Eq, Hash, Debug)]
 pub enum BinOpType {
     // numerical

optd-cost-model/src/common/predicates/constant_pred.rs

@@ -1,5 +1,6 @@
 use serde::{Deserialize, Serialize};
 
+/// TODO: documentation
 #[derive(Copy, Clone, PartialEq, Eq, Hash, Debug, Serialize, Deserialize)]
 pub enum ConstantType {
     Bool,

optd-cost-model/src/common/predicates/func_pred.rs

@@ -1,3 +1,4 @@
+/// TODO: documentation
 #[derive(Clone, PartialEq, Eq, Hash, Debug)]
 pub enum FuncType {
     Scalar(datafusion_expr::BuiltinScalarFunction),

optd-cost-model/src/common/predicates/log_op_pred.rs

@@ -1,5 +1,6 @@
 use std::fmt::Display;
 
+/// TODO: documentation
 #[derive(Copy, Clone, PartialEq, Eq, Hash, Debug)]
 pub enum LogOpType {
     And,

optd-cost-model/src/common/predicates/sort_order_pred.rs

@@ -1,5 +1,6 @@
 use std::fmt::Display;
 
+/// TODO: documentation
 #[derive(Copy, Clone, PartialEq, Eq, Hash, Debug)]
 pub enum SortOrderType {
     Asc,

optd-cost-model/src/common/predicates/un_op_pred.rs

@@ -1,5 +1,6 @@
 use std::fmt::Display;
 
+/// TODO: documentation
 #[derive(Copy, Clone, PartialEq, Eq, Hash, Debug)]
 pub enum UnOpType {
     Neg = 1,

optd-cost-model/src/common/types.rs

@@ -1,13 +1,24 @@
 use std::fmt::Display;
 
+/// TODO: documentation
 #[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug, Default, Hash)]
 pub struct GroupId(pub usize);
 
+/// TODO: documentation
 #[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug, Default, Hash)]
 pub struct ExprId(pub usize);
 
+/// TODO: documentation
 #[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug, Default, Hash)]
-pub struct PredId(pub usize);
+pub struct TableId(pub usize);
+
+/// TODO: documentation
+#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug, Default, Hash)]
+pub struct AttrId(pub usize);
+
+/// TODO: documentation
+#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug, Default, Hash)]
+pub struct EpochId(pub usize);
 
 impl Display for GroupId {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
@@ -21,8 +32,20 @@ impl Display for ExprId {
     }
 }
 
-impl Display for PredId {
+impl Display for TableId {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Table#{}", self.0)
+    }
+}
+
+impl Display for AttrId {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "Attr#{}", self.0)
+    }
+}
+
+impl Display for EpochId {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "P{}", self.0)
+        write!(f, "Epoch#{}", self.0)
     }
 }

optd-cost-model/src/common/values.rs

@@ -4,9 +4,11 @@ use ordered_float::OrderedFloat;
 use serde::{Deserialize, Deserializer, Serialize, Serializer};
 use std::sync::Arc;
 
+/// TODO: documentation
 #[derive(Clone, Debug, PartialEq, Eq, Hash, PartialOrd, Ord)]
 pub struct SerializableOrderedF64(pub OrderedFloat<f64>);
 
+/// TODO: documentation
 impl Serialize for SerializableOrderedF64 {
     fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
     where
@@ -17,6 +19,7 @@ impl Serialize for SerializableOrderedF64 {
     }
 }
 
+/// TODO: documentation
 impl<'de> Deserialize<'de> for SerializableOrderedF64 {
     fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
     where
@@ -28,6 +31,7 @@ impl<'de> Deserialize<'de> for SerializableOrderedF64 {
     }
 }
 
+/// TODO: documentation
 #[derive(Clone, Debug, PartialEq, Eq, Hash, Serialize, Deserialize, PartialOrd, Ord)]
 pub enum Value {
     UInt8(u8),
@@ -69,6 +73,7 @@ impl std::fmt::Display for Value {
     }
 }
 
+/// TODO: documentation
 /// The `as_*()` functions do not perform conversions. This is *unlike* the `as`
 /// keyword in rust.
 ///

optd-cost-model/src/cost_model.rs

@@ -10,20 +10,22 @@ use optd_persistent::{
 use crate::{
     common::{
         nodes::{ArcPredicateNode, PhysicalNodeType},
-        types::ExprId,
+        types::{AttrId, EpochId, ExprId, TableId},
     },
     storage::CostModelStorageManager,
-    ComputeCostContext, CostModel, CostModelResult, EstimatedStatistic,
+    ComputeCostContext, Cost, CostModel, CostModelResult, EstimatedStatistic, StatValue,
 };
 
-pub struct CostModelImpl<CMSL: CostModelStorageLayer> {
-    storage_manager: CostModelStorageManager<CMSL>,
+/// TODO: documentation
+pub struct CostModelImpl<S: CostModelStorageLayer> {
+    storage_manager: CostModelStorageManager<S>,
     default_catalog_source: CatalogSource,
 }
 
-impl<CMSL: CostModelStorageLayer> CostModelImpl<CMSL> {
+impl<S: CostModelStorageLayer> CostModelImpl<S> {
+    /// TODO: documentation
     pub fn new(
-        storage_manager: CostModelStorageManager<CMSL>,
+        storage_manager: CostModelStorageManager<S>,
         default_catalog_source: CatalogSource,
     ) -> Self {
         Self {
@@ -33,7 +35,17 @@ impl<CMSL: CostModelStorageLayer> CostModelImpl<CMSL> {
     }
 }
 
-impl<CMSL: CostModelStorageLayer + std::marker::Sync + 'static> CostModel for CostModelImpl<CMSL> {
+impl<S: CostModelStorageLayer + std::marker::Sync + 'static> CostModel for CostModelImpl<S> {
+    fn compute_operation_cost(
+        &self,
+        node: &PhysicalNodeType,
+        predicates: &[ArcPredicateNode],
+        children_stats: &[Option<&EstimatedStatistic>],
+        context: Option<ComputeCostContext>,
+    ) -> CostModelResult<Cost> {
+        todo!()
+    }
+
     fn derive_statistics(
         &self,
         node: PhysicalNodeType,
@@ -56,27 +68,26 @@ impl<CMSL: CostModelStorageLayer + std::marker::Sync + 'static> CostModel for Co
     fn get_table_statistic_for_analysis(
         &self,
         // TODO: i32 should be changed to TableId.
-        table_id: i32,
+        table_id: TableId,
         stat_type: StatType,
-        epoch_id: Option<i32>,
-    ) -> CostModelResult<Option<crate::StatValue>> {
+        epoch_id: Option<EpochId>,
+    ) -> CostModelResult<Option<StatValue>> {
         todo!()
     }
 
     fn get_attribute_statistic_for_analysis(
         &self,
-        // TODO: i32 should be changed to AttrId or EpochId.
-        attr_ids: Vec<i32>,
+        attr_ids: Vec<AttrId>,
         stat_type: StatType,
-        epoch_id: Option<i32>,
-    ) -> CostModelResult<Option<crate::StatValue>> {
+        epoch_id: Option<EpochId>,
+    ) -> CostModelResult<Option<StatValue>> {
         todo!()
     }
 
     fn get_cost_for_analysis(
         &self,
         expr_id: ExprId,
-        epoch_id: Option<i32>,
+        epoch_id: Option<EpochId>,
     ) -> CostModelResult<Option<crate::Cost>> {
         todo!()
     }

optd-cost-model/src/lib.rs

@@ -1,6 +1,6 @@
 use common::{
     nodes::{ArcPredicateNode, PhysicalNodeType},
-    types::{ExprId, GroupId},
+    types::{AttrId, EpochId, ExprId, GroupId, TableId},
 };
 use optd_persistent::cost_model::interface::{Stat, StatType};
 
@@ -38,9 +38,20 @@ pub enum CostModelError {
 }
 
 pub trait CostModel: 'static + Send + Sync {
+    /// TODO: documentation
+    fn compute_operation_cost(
+        &self,
+        node: &PhysicalNodeType,
+        predicates: &[ArcPredicateNode],
+        children_stats: &[Option<&EstimatedStatistic>],
+        context: Option<ComputeCostContext>,
+    ) -> CostModelResult<Cost>;
+
     /// TODO: documentation
     /// It is for cardinality estimation. The output should be the estimated
     /// statistic calculated by the cost model.
+    /// TODO: Consider make it a helper function, so we can store Cost in the
+    /// ORM more easily.
     fn derive_statistics(
         &self,
         node: PhysicalNodeType,
@@ -59,25 +70,26 @@ pub trait CostModel: 'static + Send + Sync {
         data: String,
     ) -> CostModelResult<()>;
 
+    /// TODO: documentation
     fn get_table_statistic_for_analysis(
         &self,
-        // TODO: i32 should be changed to TableId.
-        table_id: i32,
+        table_id: TableId,
         stat_type: StatType,
-        epoch_id: Option<i32>,
+        epoch_id: Option<EpochId>,
     ) -> CostModelResult<Option<StatValue>>;
 
+    /// TODO: documentation
     fn get_attribute_statistic_for_analysis(
         &self,
-        // TODO: i32 should be changed to AttrId or EpochId.
-        attr_ids: Vec<i32>,
+        attr_ids: Vec<AttrId>,
         stat_type: StatType,
-        epoch_id: Option<i32>,
+        epoch_id: Option<EpochId>,
     ) -> CostModelResult<Option<StatValue>>;
 
+    /// TODO: documentation
     fn get_cost_for_analysis(
         &self,
         expr_id: ExprId,
-        epoch_id: Option<i32>,
+        epoch_id: Option<EpochId>,
     ) -> CostModelResult<Option<Cost>>;
 }

optd-cost-model/src/storage.rs

@@ -2,13 +2,15 @@ use std::sync::Arc;
 
 use optd_persistent::CostModelStorageLayer;
 
-pub struct CostModelStorageManager<CMSL: CostModelStorageLayer> {
-    pub backend_manager: Arc<CMSL>,
+/// TODO: documentation
+pub struct CostModelStorageManager<S: CostModelStorageLayer> {
+    pub backend_manager: Arc<S>,
     // TODO: in-memory cache
 }
 
-impl<CMSL: CostModelStorageLayer> CostModelStorageManager<CMSL> {
-    pub fn new(backend_manager: Arc<CMSL>) -> Self {
+impl<S: CostModelStorageLayer> CostModelStorageManager<S> {
+    /// TODO: documentation
+    pub fn new(backend_manager: Arc<S>) -> Self {
         Self { backend_manager }
     }
 }