MetonaSqlark/src/sql/parser.ts

/**
 * metona-sqlark SQL Parser — 递归下降语法分析器
 * @module sql/parser
 *
 * Token 流 → AST Statement。
 * 支持的语法是标准 SQL 的子集。
 */

import { Lexer } from './lexer';
import { TokenType, type Token } from './tokens';
import type {
  Statement,
  SelectStatement,
  InsertStatement,
  UpdateStatement,
  DeleteStatement,
  CreateTableStatement,
  DropTableStatement,
  ASTColumnDef,
} from '../query/ast';
import type { WhereCondition, OrderBy, SortDirection } from '../constants';
import { DatabaseError } from '../constants';

// ---------------------------------------------------------------------------
//  Parser
// ---------------------------------------------------------------------------

export class Parser {
  private lexer: Lexer;
  private curToken!: Token;
  private peekToken!: Token;

  constructor(sql: string) {
    this.lexer = new Lexer(sql);
    // 预读两个 token
    this.nextToken();
    this.nextToken();
  }

  /** 解析完整 SQL 语句 */
  parseStatement(): Statement {
    switch (this.curToken.type) {
      case TokenType.SELECT:
        return this.parseSelect();
      case TokenType.INSERT:
        return this.parseInsert();
      case TokenType.UPDATE:
        return this.parseUpdate();
      case TokenType.DELETE:
        return this.parseDelete();
      case TokenType.CREATE:
        return this.parseCreateTable();
      case TokenType.DROP:
        return this.parseDropTable();
      default:
        throw this.error(`Unexpected token "${this.curToken.value}"`);
    }
  }

  // ===================================================================
  //  SELECT
  // ===================================================================

  private parseSelect(): SelectStatement {
    this.expect(TokenType.SELECT);

    // DISTINCT（可选）
    let distinct = false;
    if (this.curTokenIs(TokenType.DISTINCT)) {
      distinct = true;
      this.nextToken();
    }

    // 列
    const columns: string[] = [];
    if (this.curTokenIs(TokenType.STAR)) {
      columns.push('*');
      this.nextToken();
    } else {
      columns.push(...this.parseColumnList());
    }

    // FROM
    this.expect(TokenType.FROM);
    const tableName = this.expectIdentifier('table name');

    // 表别名（可选）
    let alias: string | undefined;
    if (this.curTokenIs(TokenType.AS)) {
      this.nextToken();
      alias = this.expectIdentifier('alias');
    } else if (this.curToken.type === TokenType.IDENTIFIER && !this._isReservedAfterFrom()) {
      alias = this.curToken.value;
      this.nextToken();
    }

    const stmt: SelectStatement = {
      type: 'SELECT',
      columns,
      distinct: distinct || undefined,
      from: tableName,
      alias,
      where: {},
    };

    // JOIN 子句（可选，支持多个）
    const joins = this.parseJoinClauses();
    if (joins.length > 0) {
      stmt.joins = joins;
    }

    // WHERE（可选）
    if (this.curTokenIs(TokenType.WHERE)) {
      this.nextToken();
      stmt.where = this.parseCondition();
    }

    // GROUP BY（可选）
    if (this.curTokenIs(TokenType.GROUP)) {
      this.nextToken();
      this.expect(TokenType.BY);
      stmt.groupBy = this.parseIdentifierList();
    }

    // HAVING（可选）
    if (this.curTokenIs(TokenType.HAVING)) {
      this.nextToken();
      stmt.having = this.parseCondition();
    }

    // ORDER BY（可选）
    if (this.curTokenIs(TokenType.ORDER)) {
      this.nextToken();
      this.expect(TokenType.BY);
      stmt.orderBy = this.parseOrderByList();
    }

    // LIMIT（可选）
    if (this.curTokenIs(TokenType.LIMIT)) {
      this.nextToken();
      stmt.limit = this.expectNumber('LIMIT value');
    }

    // OFFSET（可选）
    if (this.curTokenIs(TokenType.OFFSET)) {
      this.nextToken();
      stmt.offset = this.expectNumber('OFFSET value');
    }

    return stmt;
  }

  /** 解析 JOIN 子句列表 */
  private parseJoinClauses(): import('../query/ast').JoinClause[] {
    const joins: import('../query/ast').JoinClause[] = [];
    while (this._isJoinKeyword()) {
      joins.push(this.parseJoinClause());
    }
    return joins;
  }

  private _isJoinKeyword(): boolean {
    return (
      this.curTokenIs(TokenType.INNER) ||
      this.curTokenIs(TokenType.LEFT) ||
      this.curTokenIs(TokenType.RIGHT) ||
      this.curTokenIs(TokenType.CROSS) ||
      this.curTokenIs(TokenType.JOIN)
    );
  }

  /** 解析单个 JOIN 子句 */
  private parseJoinClause(): import('../query/ast').JoinClause {
    let type: import('../query/ast').JoinType = 'INNER';

    if (this.curTokenIs(TokenType.INNER)) {
      type = 'INNER';
      this.nextToken();
    } else if (this.curTokenIs(TokenType.LEFT)) {
      type = 'LEFT';
      this.nextToken();
      if (this.curTokenIs(TokenType.OUTER)) this.nextToken(); // 可选 OUTER
    } else if (this.curTokenIs(TokenType.RIGHT)) {
      type = 'RIGHT';
      this.nextToken();
      if (this.curTokenIs(TokenType.OUTER)) this.nextToken();
    } else if (this.curTokenIs(TokenType.CROSS)) {
      type = 'CROSS';
      this.nextToken();
    }

    this.expect(TokenType.JOIN);
    const tableName = this.expectIdentifier('table name');

    // JOIN 表别名（可选）
    let alias: string | undefined;
    if (this.curTokenIs(TokenType.AS)) {
      this.nextToken();
      alias = this.expectIdentifier('alias');
    } else if (this.curToken.type === TokenType.IDENTIFIER && !this._isJoinReserved()) {
      alias = this.curToken.value;
      this.nextToken();
    }

    // ON 条件（CROSS JOIN 不需要 ON）
    let on = {};
    if (type !== 'CROSS' && this.curTokenIs(TokenType.ON)) {
      this.nextToken();
      on = this.parseCondition();
    }

    return { type, table: tableName, alias, on };
  }

  /** 判断当前 token 是否为 FROM 之后的保留字 */
  private _isReservedAfterFrom(): boolean {
    return (
      this.curTokenIs(TokenType.WHERE) ||
      this.curTokenIs(TokenType.ORDER) ||
      this.curTokenIs(TokenType.LIMIT) ||
      this.curTokenIs(TokenType.OFFSET) ||
      this.curTokenIs(TokenType.GROUP) ||
      this._isJoinKeyword()
    );
  }

  private _isJoinReserved(): boolean {
    return (
      this.curTokenIs(TokenType.ON) ||
      this.curTokenIs(TokenType.WHERE) ||
      this.curTokenIs(TokenType.ORDER) ||
      this.curTokenIs(TokenType.LIMIT) ||
      this._isJoinKeyword()
    );
  }

  // ===================================================================
  //  INSERT
  // ===================================================================

  private parseInsert(): InsertStatement {
    this.expect(TokenType.INSERT);
    this.expect(TokenType.INTO);

    const tableName = this.expectIdentifier('table name');

    // 列名（可选）
    let columns: string[] | undefined;
    if (this.curTokenIs(TokenType.LPAREN)) {
      this.nextToken();
      columns = this.parseIdentifierList();
      this.expect(TokenType.RPAREN);
    }

    // VALUES
    this.expect(TokenType.VALUES);

    // 值列表
    const values: unknown[][] = [];
    do {
      if (this.curTokenIs(TokenType.COMMA)) {
        this.nextToken();
      }
      this.expect(TokenType.LPAREN);
      const rowValues = this.parseValueList();
      this.expect(TokenType.RPAREN);
      values.push(rowValues);
    } while (this.curTokenIs(TokenType.COMMA));

    return {
      type: 'INSERT',
      into: tableName,
      columns,
      values,
    };
  }

  // ===================================================================
  //  UPDATE
  // ===================================================================

  private parseUpdate(): UpdateStatement {
    this.expect(TokenType.UPDATE);
    const tableName = this.expectIdentifier('table name');
    this.expect(TokenType.SET);

    // SET col=val, ...
    const sets: Record<string, unknown> = {};
    do {
      if (this.curTokenIs(TokenType.COMMA)) this.nextToken();
      const col = this.expectIdentifier('column name');
      this.expect(TokenType.EQ);
      sets[col] = this.parseValue();
    } while (this.curTokenIs(TokenType.COMMA));

    let where: WhereCondition = {};
    if (this.curTokenIs(TokenType.WHERE)) {
      this.nextToken();
      where = this.parseCondition();
    }

    return { type: 'UPDATE', table: tableName, sets, where };
  }

  // ===================================================================
  //  DELETE
  // ===================================================================

  private parseDelete(): DeleteStatement {
    this.expect(TokenType.DELETE);
    this.expect(TokenType.FROM);
    const tableName = this.expectIdentifier('table name');

    let where: WhereCondition = {};
    if (this.curTokenIs(TokenType.WHERE)) {
      this.nextToken();
      where = this.parseCondition();
    }

    return { type: 'DELETE', from: tableName, where };
  }

  // ===================================================================
  //  CREATE TABLE
  // ===================================================================

  private parseCreateTable(): CreateTableStatement {
    this.expect(TokenType.CREATE);
    this.expect(TokenType.TABLE);
    const tableName = this.expectIdentifier('table name');
    this.expect(TokenType.LPAREN);

    const columns: ASTColumnDef[] = [];
    do {
      if (this.curTokenIs(TokenType.COMMA)) this.nextToken();
      columns.push(this.parseColumnDef());
    } while (this.curTokenIs(TokenType.COMMA));

    this.expect(TokenType.RPAREN);

    return { type: 'CREATE_TABLE', name: tableName, columns };
  }

  private parseColumnDef(): ASTColumnDef {
    const name = this.expectIdentifier('column name');
    const type = this.expectIdentifier('column type').toLowerCase();

    const col: ASTColumnDef = { name, type };

    // 修饰符
    while (
      this.curTokenIs(TokenType.PRIMARY) ||
      this.curTokenIs(TokenType.UNIQUE) ||
      this.curTokenIs(TokenType.NOT) ||
      this.curTokenIs(TokenType.DEFAULT)
    ) {
      if (this.curTokenIs(TokenType.PRIMARY)) {
        this.nextToken();
        this.expect(TokenType.KEY);
        col.primaryKey = true;
      } else if (this.curTokenIs(TokenType.UNIQUE)) {
        this.nextToken();
        col.unique = true;
      } else if (this.curTokenIs(TokenType.NOT)) {
        this.nextToken();
        // 支持 NOT NULL → required
        this.expect(TokenType.NULL);
        col.required = true;
      } else if (this.curTokenIs(TokenType.DEFAULT)) {
        this.nextToken();
        col.default = this.parseValue();
      } else {
        break;
      }
    }

    return col;
  }

  // ===================================================================
  //  DROP TABLE
  // ===================================================================

  private parseDropTable(): DropTableStatement {
    this.expect(TokenType.DROP);
    this.expect(TokenType.TABLE);
    const tableName = this.expectIdentifier('table name');
    return { type: 'DROP_TABLE', name: tableName };
  }

  // ===================================================================
  //  条件表达式
  // ===================================================================

  /** condition → simple_cond ((AND|OR) simple_cond)* */
  private parseCondition(): WhereCondition {
    let left = this.parseSimpleCondition();

    while (this.curTokenIs(TokenType.AND) || this.curTokenIs(TokenType.OR)) {
      const isAnd = this.curTokenIs(TokenType.AND);
      this.nextToken();
      const right = this.parseSimpleCondition();

      if (isAnd) {
        // 合并到 $and
        left = { $and: [left, right] } as unknown as WhereCondition;
      } else {
        left = { $or: [left, right] } as unknown as WhereCondition;
      }
    }

    return left;
  }

  /** simple_cond → column op value | column IS [NOT] NULL | column [NOT] LIKE pattern
   *                | column [NOT] IN (values) | NOT condition | (condition) */
  private parseSimpleCondition(): WhereCondition {
    // NOT expr
    if (this.curTokenIs(TokenType.NOT)) {
      this.nextToken();
      const inner = this.parseSimpleCondition();
      return { $not: inner } as unknown as WhereCondition;
    }

    // (condition)
    if (this.curTokenIs(TokenType.LPAREN)) {
      this.nextToken();
      const inner = this.parseCondition();
      this.expect(TokenType.RPAREN);
      return inner;
    }

    // column
    const column = this.parseColumnRef();

    // IS NULL / IS NOT NULL
    if (this.curTokenIs(TokenType.IDENTIFIER) && this.curToken.value.toUpperCase() === 'IS') {
      this.nextToken();
      const isNot = this.curTokenIs(TokenType.NOT);
      if (isNot) this.nextToken();
      this.expect(TokenType.NULL);
      const result: WhereCondition = {};
      result[column] = isNot ? { $ne: null } : { $eq: null };
      return result;
    }

    // LIKE
    if (this.curTokenIs(TokenType.LIKE)) {
      this.nextToken();
      const pattern = this.parseValue();
      const result: WhereCondition = {};
      result[column] = { $like: pattern };
      return result;
    }

    // IN
    if (this.curTokenIs(TokenType.IN)) {
      this.nextToken();
      this.expect(TokenType.LPAREN);
      const values = this.parseValueList();
      this.expect(TokenType.RPAREN);
      const result: WhereCondition = {};
      result[column] = { $in: values };
      return result;
    }

    // 比较运算符
    const op = this.parseComparisonOp();

    // 尝试解析列引用（identifier DOT identifier 格式）
    let value: unknown;
    if (
      (this.curToken.type === TokenType.IDENTIFIER || this._isKeywordAsIdent()) &&
      this.peekTokenIs(TokenType.DOT)
    ) {
      const colRef = this.parseColumnRef();
      value = { $col: colRef };
    } else {
      value = this.parseValue();
    }

    const result: WhereCondition = {};
    result[column] = { [op]: value };
    return result;
  }

  private peekTokenIs(type: TokenType): boolean {
    return this.peekToken.type === type;
  }

  private parseComparisonOp(): string {
    switch (this.curToken.type) {
      case TokenType.EQ: this.nextToken(); return '$eq';
      case TokenType.NEQ: this.nextToken(); return '$ne';
      case TokenType.GT: this.nextToken(); return '$gt';
      case TokenType.GTE: this.nextToken(); return '$gte';
      case TokenType.LT: this.nextToken(); return '$lt';
      case TokenType.LTE: this.nextToken(); return '$lte';
      default:
        throw this.error(`Expected comparison operator, got "${this.curToken.value}"`);
    }
  }

  // ===================================================================
  //  辅助解析
  // ===================================================================

  private parseColumnList(): string[] {
    const cols: string[] = [];
    cols.push(this.parseColumnRef());
    while (this.curTokenIs(TokenType.COMMA)) {
      this.nextToken();
      cols.push(this.parseColumnRef());
    }
    return cols;
  }

  /** 解析列引用：支持 'col'、'table.col' 和 'COUNT(*)'/'SUM(col)' 等 */
  private parseColumnRef(): string {
    // 聚合函数?
    if (
      this.curTokenIs(TokenType.COUNT) ||
      this.curTokenIs(TokenType.SUM) ||
      this.curTokenIs(TokenType.AVG) ||
      this.curTokenIs(TokenType.MIN) ||
      this.curTokenIs(TokenType.MAX)
    ) {
      return this.parseAggregateCall();
    }

    const first = this.expectIdentifier('column name');
    if (this.curTokenIs(TokenType.DOT)) {
      this.nextToken();
      const second = this.expectIdentifier('column name');
      return `${first}.${second}`;
    }
    return first;
  }

  /** 解析聚合函数调用: COUNT(*), SUM(col), AVG(col), MIN(col), MAX(col) */
  private parseAggregateCall(): string {
    const func = this.curToken.value.toUpperCase();
    this.nextToken();
    this.expect(TokenType.LPAREN);

    let arg: string;
    if (this.curTokenIs(TokenType.STAR)) {
      arg = '*';
      this.nextToken();
    } else {
      arg = this.parseColumnRef();
    }

    this.expect(TokenType.RPAREN);

    // 可选别名: AS alias
    let alias = '';
    if (this.curTokenIs(TokenType.AS)) {
      this.nextToken();
      alias = this.expectIdentifier('alias');
    } else if (this.curToken.type === TokenType.IDENTIFIER && this._isAggregateAlias()) {
      alias = this.curToken.value;
      this.nextToken();
    }

    if (alias) {
      return `${func}(${arg}) AS ${alias}`;
    }
    return `${func}(${arg})`;
  }

  private _isAggregateAlias(): boolean {
    return !this._isReservedAfterFrom() && !this._isJoinKeyword();
  }

  private parseIdentifierList(): string[] {
    const ids: string[] = [];
    ids.push(this.expectIdentifier('identifier'));
    while (this.curTokenIs(TokenType.COMMA)) {
      this.nextToken();
      ids.push(this.expectIdentifier('identifier'));
    }
    return ids;
  }

  private parseValueList(): unknown[] {
    const vals: unknown[] = [];
    vals.push(this.parseValue());
    while (this.curTokenIs(TokenType.COMMA)) {
      this.nextToken();
      vals.push(this.parseValue());
    }
    return vals;
  }

  private parseOrderByList(): OrderBy[] {
    const list: OrderBy[] = [];
    list.push(this.parseOrderBy());
    while (this.curTokenIs(TokenType.COMMA)) {
      this.nextToken();
      list.push(this.parseOrderBy());
    }
    return list;
  }

  private parseOrderBy(): OrderBy {
    const column = this.expectIdentifier('column name');
    let direction: SortDirection = 'asc';
    if (this.curTokenIs(TokenType.ASC)) {
      this.nextToken();
    } else if (this.curTokenIs(TokenType.DESC)) {
      direction = 'desc';
      this.nextToken();
    }
    return { column, direction };
  }

  /** 解析字面量值 */
  private parseValue(): unknown {
    switch (this.curToken.type) {
      case TokenType.STRING: {
        const val = this.curToken.value;
        this.nextToken();
        return val;
      }
      case TokenType.NUMBER: {
        const val = Number(this.curToken.value);
        this.nextToken();
        return val;
      }
      case TokenType.TRUE: this.nextToken(); return true;
      case TokenType.FALSE: this.nextToken(); return false;
      case TokenType.NULL: this.nextToken(); return null;
      default:
        throw this.error(`Expected value, got "${this.curToken.value}"`);
    }
  }

  // ===================================================================
  //  Token 操作
  // ===================================================================

  private nextToken(): void {
    this.curToken = this.peekToken;
    this.peekToken = this.lexer.nextToken();
  }

  private curTokenIs(type: TokenType): boolean {
    return this.curToken.type === type;
  }

  private expect(type: TokenType): void {
    if (this.curTokenIs(type)) {
      this.nextToken();
      return;
    }
    throw this.error(`Expected ${type}, got "${this.curToken.value}"`);
  }

  private expectIdentifier(context: string): string {
    if (this.curToken.type === TokenType.IDENTIFIER || this._isKeywordAsIdent()) {
      const val = this.curToken.value;
      this.nextToken();
      return val;
    }
    throw this.error(`Expected ${context}, got "${this.curToken.value}"`);
  }

  /** 关键字可以作为标识符（如列名等于关键字） */
  private _isKeywordAsIdent(): boolean {
    return (
      this.curToken.type !== TokenType.EOF &&
      this.curToken.type !== TokenType.ILLEGAL &&
      this.curToken.type !== TokenType.STRING &&
      this.curToken.type !== TokenType.NUMBER &&
      this.curToken.type !== TokenType.COMMA &&
      this.curToken.type !== TokenType.LPAREN &&
      this.curToken.type !== TokenType.RPAREN &&
      this.curToken.type !== TokenType.SEMICOLON &&
      this.curToken.type !== TokenType.EQ &&
      this.curToken.type !== TokenType.NEQ &&
      this.curToken.type !== TokenType.GT &&
      this.curToken.type !== TokenType.GTE &&
      this.curToken.type !== TokenType.LT &&
      this.curToken.type !== TokenType.LTE &&
      this.curToken.type !== TokenType.DOT &&
      this.curToken.type !== TokenType.STAR
    );
  }

  private expectNumber(context: string): number {
    if (this.curToken.type === TokenType.NUMBER) {
      const val = Number(this.curToken.value);
      this.nextToken();
      return val;
    }
    throw this.error(`Expected ${context}, got "${this.curToken.value}"`);
  }

  private error(msg: string): DatabaseError {
    return new DatabaseError(
      `Parse error at position ${this.curToken.position}: ${msg}`,
      'PARSE_ERROR',
    );
  }
}

// ---------------------------------------------------------------------------
//  便捷方法
// ---------------------------------------------------------------------------

/** 解析 SQL 字符串为 AST Statement */
export function parse(sql: string): Statement {
  const parser = new Parser(sql);
  const stmt = parser.parseStatement();
  return stmt;
}