feat(catalog): add function catalog (#827)

Signed-off-by: Michael Xu <[email protected]>

.gitignore

@@ -5,3 +5,4 @@
 /.vscode
 .gdb_history
 risinglight
+/.DS_Store

src/binder/create_function.rs

@@ -0,0 +1,56 @@
+// Copyright 2024 RisingLight Project Authors. Licensed under Apache-2.0.
+
+use std::fmt;
+use std::str::FromStr;
+
+use pretty_xmlish::helper::delegate_fmt;
+use pretty_xmlish::Pretty;
+use serde::{Deserialize, Serialize};
+
+use super::*;
+
+#[derive(Debug, PartialEq, Eq, PartialOrd, Ord, Hash, Clone, Serialize, Deserialize)]
+pub struct CreateFunction {
+    name: String,
+    arg_types: Vec<DataType>,
+    return_types: DataType,
+    language: String,
+    body: String,
+}
+
+impl fmt::Display for CreateFunction {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let explainer = Pretty::childless_record("CreateFunction", self.pretty_function());
+        delegate_fmt(&explainer, f, String::with_capacity(1000))
+    }
+}
+
+impl FromStr for CreateFunction {
+    type Err = ();
+
+    fn from_str(_s: &str) -> std::result::Result<Self, Self::Err> {
+        Err(())
+    }
+}
+
+impl CreateFunction {
+    pub fn pretty_function<'a>(&self) -> Vec<(&'a str, Pretty<'a>)> {
+        vec![
+            ("name", Pretty::display(&self.name)),
+            ("language", Pretty::display(&self.language)),
+            ("body", Pretty::display(&self.body)),
+        ]
+    }
+}
+
+impl Binder {
+    pub(super) fn bind_create_function(
+        &mut self,
+        _name: ObjectName,
+        _args: Option<Vec<OperateFunctionArg>>,
+        _return_type: Option<DataType>,
+        _params: CreateFunctionBody,
+    ) -> Result {
+        todo!()
+    }
+}

src/binder/mod.rs

@@ -15,6 +15,7 @@ use crate::planner::{Expr as Node, RecExpr, TypeError, TypeSchemaAnalysis};
 use crate::types::{DataTypeKind, DataValue};
 
 pub mod copy;
+mod create_function;
 mod create_table;
 mod delete;
 mod drop;
@@ -23,6 +24,7 @@ mod insert;
 mod select;
 mod table;
 
+pub use self::create_function::*;
 pub use self::create_table::*;
 pub use self::drop::*;
 
@@ -154,6 +156,13 @@ impl Binder {
                 constraints,
                 ..
             } => self.bind_create_table(name, &columns, &constraints),
+            Statement::CreateFunction {
+                name,
+                args,
+                return_type,
+                params,
+                ..
+            } => self.bind_create_function(name, args, return_type, params),
             Statement::Drop {
                 object_type,
                 if_exists,

src/catalog/function.rs

@@ -0,0 +1,42 @@
+// Copyright 2024 RisingLight Project Authors. Licensed under Apache-2.0.
+
+use crate::types::DataType;
+
+#[derive(Clone, PartialEq, Eq, Hash, Debug)]
+pub struct FunctionCatalog {
+    name: String,
+    arg_types: Vec<DataType>,
+    return_type: DataType,
+    language: String,
+    body: String,
+}
+
+impl FunctionCatalog {
+    pub fn new(
+        name: String,
+        arg_types: Vec<DataType>,
+        return_type: DataType,
+        language: String,
+        body: String,
+    ) -> Self {
+        Self {
+            name,
+            arg_types,
+            return_type,
+            language,
+            body,
+        }
+    }
+
+    pub fn body(&self) -> String {
+        self.body.clone()
+    }
+
+    pub fn name(&self) -> String {
+        self.name.clone()
+    }
+
+    pub fn language(&self) -> String {
+        self.language.clone()
+    }
+}

src/catalog/mod.rs

@@ -18,6 +18,7 @@ static CONTRIBUTORS_TABLE_NAME: &str = "contributors";
 pub const CONTRIBUTORS_TABLE_ID: TableId = 0;
 
 mod column;
+mod function;
 mod root;
 mod schema;
 mod table;

src/catalog/schema.rs

@@ -3,6 +3,7 @@
 use std::collections::HashMap;
 use std::sync::Arc;
 
+use super::function::FunctionCatalog;
 use super::*;
 
 /// The catalog of a schema.
@@ -13,6 +14,8 @@ pub struct SchemaCatalog {
     table_idxs: HashMap<String, TableId>,
     tables: HashMap<TableId, Arc<TableCatalog>>,
     next_table_id: TableId,
+    /// Currently indexed by function name
+    functions: HashMap<String, Arc<FunctionCatalog>>,
 }
 
 impl SchemaCatalog {
@@ -23,6 +26,7 @@ impl SchemaCatalog {
             table_idxs: HashMap::new(),
             tables: HashMap::new(),
             next_table_id: 0,
+            functions: HashMap::new(),
         }
     }
 
@@ -81,6 +85,10 @@ impl SchemaCatalog {
     pub fn id(&self) -> SchemaId {
         self.id
     }
+
+    pub fn get_function_by_name(&self, name: &str) -> Option<Arc<FunctionCatalog>> {
+        self.functions.get(name).cloned()
+    }
 }
 
 #[cfg(test)]

src/planner/explain.rs

@@ -341,6 +341,10 @@ impl<'a> Explain<'a> {
                 let fields = t.pretty_table().with(cost, rows);
                 Pretty::childless_record("CreateTable", fields)
             }
+            CreateFunction(f) => {
+                let v = f.pretty_function();
+                Pretty::childless_record("CreateFunction", v)
+            }
             Drop(t) => {
                 let fields = t.pretty_table().with(cost, rows);
                 Pretty::childless_record("Drop", fields)

src/planner/mod.rs

@@ -3,7 +3,7 @@
 use egg::{define_language, Id, Symbol};
 
 use crate::binder::copy::ExtSource;
-use crate::binder::{BoundDrop, CreateTable};
+use crate::binder::{BoundDrop, CreateFunction, CreateTable};
 use crate::catalog::{ColumnRefId, TableRefId};
 use crate::parser::{BinaryOperator, UnaryOperator};
 use crate::types::{ColumnIndex, DataTypeKind, DataValue, DateTimeField};
@@ -117,6 +117,7 @@ define_language! {
         "window" = Window([Id; 2]),             // (window [over..] child)
                                                     // output = child || exprs
         CreateTable(CreateTable),
+        CreateFunction(CreateFunction),
         Drop(BoundDrop),
         "insert" = Insert([Id; 3]),             // (insert table [column..] child)
         "delete" = Delete([Id; 2]),             // (delete table child)