feat(postgres): add TSVECTOR datatype and `@@` operator (#12955)

Co-authored-by: Jano Kacer <jan@vestberry.com>
Co-authored-by: Sébastien BRAMILLE <2752200+oktapodia@users.noreply.github.com>

docs/manual/core-concepts/model-basics.md

@@ -270,6 +270,7 @@ DataTypes.STRING.BINARY      // VARCHAR BINARY
 DataTypes.TEXT               // TEXT
 DataTypes.TEXT               // TEXT
 DataTypes.TEXT('tiny')       // TINYTEXT
 DataTypes.TEXT('tiny')       // TINYTEXT
 DataTypes.CITEXT             // CITEXT          PostgreSQL and SQLite only.
 DataTypes.CITEXT             // CITEXT          PostgreSQL and SQLite only.
+DataTypes.TSVECTOR           // TSVECTOR        PostgreSQL only.
 ```
 ```
 
 
 ### Boolean
 ### Boolean

docs/manual/core-concepts/model-querying-basics.md

@@ -261,6 +261,7 @@ Post.findAll({
       [Op.notIRegexp]: '^[h|a|t]',             // !~* '^[h|a|t]' (PG only)
       [Op.notIRegexp]: '^[h|a|t]',             // !~* '^[h|a|t]' (PG only)
 
 
       [Op.any]: [2, 3],                        // ANY ARRAY[2, 3]::INTEGER (PG only)
       [Op.any]: [2, 3],                        // ANY ARRAY[2, 3]::INTEGER (PG only)
+      [Op.match]: Sequelize.fn('to_tsquery', 'fat & rat') // match text search for strings 'fat' and 'rat' (PG only)
 
 
       // In Postgres, Op.like/Op.iLike/Op.notLike can be combined to Op.any:
       // In Postgres, Op.like/Op.iLike/Op.notLike can be combined to Op.any:
       [Op.like]: { [Op.any]: ['cat', 'hat'] }  // LIKE ANY ARRAY['cat', 'hat']
       [Op.like]: { [Op.any]: ['cat', 'hat'] }  // LIKE ANY ARRAY['cat', 'hat']

lib/data-types.js

@@ -941,6 +941,21 @@ class MACADDR extends ABSTRACT {
 }
 }
 
 
 /**
 /**
+ * The TSVECTOR type stores text search vectors.
+ *
+ * Only available for Postgres
+ *
+ */
+class TSVECTOR extends ABSTRACT {
+  validate(value) {
+    if (typeof value !== 'string') {
+      throw new sequelizeErrors.ValidationError(util.format('%j is not a valid string', value));
+    }
+    return true;
+  }
+}
+
+/**
  * A convenience class holding commonly used data types. The data types are used when defining a new model using `Sequelize.define`, like this:
  * A convenience class holding commonly used data types. The data types are used when defining a new model using `Sequelize.define`, like this:
  * ```js
  * ```js
  * sequelize.define('model', {
  * sequelize.define('model', {
@@ -1023,7 +1038,8 @@ const DataTypes = module.exports = {
   CIDR,
   CIDR,
   INET,
   INET,
   MACADDR,
   MACADDR,
-  CITEXT
+  CITEXT,
+  TSVECTOR
 };
 };
 
 
 _.each(DataTypes, (dataType, name) => {
 _.each(DataTypes, (dataType, name) => {

lib/dialects/abstract/query-generator/operators.js

@@ -42,7 +42,8 @@ const OperatorHelpers = {
     [Op.and]: ' AND ',
     [Op.and]: ' AND ',
     [Op.or]: ' OR ',
     [Op.or]: ' OR ',
     [Op.col]: 'COL',
     [Op.col]: 'COL',
-    [Op.placeholder]: '$$PLACEHOLDER$$'
+    [Op.placeholder]: '$$PLACEHOLDER$$',
+    [Op.match]: '@@'
   },
   },
 
 
   OperatorsAliasMap: {},
   OperatorsAliasMap: {},

lib/dialects/postgres/data-types.js

@@ -36,6 +36,7 @@ module.exports = BaseTypes => {
   BaseTypes.CIDR.types.postgres = ['cidr'];
   BaseTypes.CIDR.types.postgres = ['cidr'];
   BaseTypes.INET.types.postgres = ['inet'];
   BaseTypes.INET.types.postgres = ['inet'];
   BaseTypes.MACADDR.types.postgres = ['macaddr'];
   BaseTypes.MACADDR.types.postgres = ['macaddr'];
+  BaseTypes.TSVECTOR.types.postgres = ['tsvector'];
   BaseTypes.JSON.types.postgres = ['json'];
   BaseTypes.JSON.types.postgres = ['json'];
   BaseTypes.JSONB.types.postgres = ['jsonb'];
   BaseTypes.JSONB.types.postgres = ['jsonb'];
   BaseTypes.TIME.types.postgres = ['time'];
   BaseTypes.TIME.types.postgres = ['time'];

lib/dialects/postgres/index.js

@@ -57,6 +57,7 @@ PostgresDialect.prototype.supports = _.merge(_.cloneDeep(AbstractDialect.prototy
   JSON: true,
   JSON: true,
   JSONB: true,
   JSONB: true,
   HSTORE: true,
   HSTORE: true,
+  TSVECTOR: true,
   deferrableConstraints: true,
   deferrableConstraints: true,
   searchPath: true
   searchPath: true
 });
 });

lib/operators.js

@@ -84,7 +84,8 @@ const Op = {
   values: Symbol.for('values'),
   values: Symbol.for('values'),
   col: Symbol.for('col'),
   col: Symbol.for('col'),
   placeholder: Symbol.for('placeholder'),
   placeholder: Symbol.for('placeholder'),
-  join: Symbol.for('join')
+  join: Symbol.for('join'),
+  match: Symbol.for('match')
 };
 };
 
 
 module.exports = Op;
 module.exports = Op;

test/integration/data-types.test.js

@@ -362,6 +362,18 @@ describe(Support.getTestDialectTeaser('DataTypes'), () => {
 
 
   });
   });
 
 
+  if (current.dialect.supports.TSVECTOR) {
+    it('calls parse and stringify for TSVECTOR', async () => {
+      const Type = new Sequelize.TSVECTOR();
+
+      if (['postgres'].includes(dialect)) {
+        await testSuccess(Type, 'swagger');
+      } else {
+        testFailure(Type);
+      }
+    });
+  }
+
   it('calls parse and stringify for ENUM', async () => {
   it('calls parse and stringify for ENUM', async () => {
     const Type = new Sequelize.ENUM('hat', 'cat');
     const Type = new Sequelize.ENUM('hat', 'cat');
 
 

test/integration/model/create.test.js

@@ -1007,6 +1007,31 @@ describe(Support.getTestDialectTeaser('Model'), () => {
       });
       });
     }
     }
 
 
+    if (dialect === 'postgres') {
+      it('allows the creation of a TSVECTOR field', async function() {
+        const User = this.sequelize.define('UserWithTSVECTOR', {
+          name: Sequelize.TSVECTOR
+        });
+
+        await User.sync({ force: true });
+        await User.create({ name: 'John Doe' });
+      });
+
+      it('TSVECTOR only allow string', async function() {
+        const User = this.sequelize.define('UserWithTSVECTOR', {
+          username: { type: Sequelize.TSVECTOR }
+        });
+
+        try {
+          await User.sync({ force: true });
+          await User.create({ username: 42 });
+        } catch (err) {
+          if (!(err instanceof Sequelize.ValidationError)) throw err;
+          expect(err).to.be.ok;
+        }
+      });
+    }
+
     if (current.dialect.supports.index.functionBased) {
     if (current.dialect.supports.index.functionBased) {
       it("doesn't allow duplicated records with unique function based indexes", async function() {
       it("doesn't allow duplicated records with unique function based indexes", async function() {
         const User = this.sequelize.define('UserWithUniqueUsernameFunctionIndex', {
         const User = this.sequelize.define('UserWithUniqueUsernameFunctionIndex', {

test/integration/model/findOne.test.js

@@ -273,6 +273,22 @@ describe(Support.getTestDialectTeaser('Model'), () => {
           expect(user.username).to.equal('longUserNAME');
           expect(user.username).to.equal('longUserNAME');
         });
         });
       }
       }
+
+      if (dialect === 'postgres') {
+        it('should allow case-sensitive find on TSVECTOR type', async function() {
+          const User = this.sequelize.define('UserWithCaseInsensitiveName', {
+            username: Sequelize.TSVECTOR
+          });
+
+          await User.sync({ force: true });
+          await User.create({ username: 'longUserNAME' });
+          const user = await User.findOne({
+            where: { username: 'longUserNAME' }
+          });
+          expect(user).to.exist;
+          expect(user.username).to.equal("'longUserNAME'");
+        });
+      }
     });
     });
 
 
     describe('eager loading', () => {
     describe('eager loading', () => {

test/unit/sql/where.test.js

@@ -1206,6 +1206,20 @@ describe(Support.getTestDialectTeaser('SQL'), () => {
       }
       }
     }
     }
 
 
+    if (current.dialect.supports.TSVESCTOR) {
+      describe('Op.match', () => {
+        testsql(
+          'username',
+          {
+            [Op.match]: Support.sequelize.fn('to_tsvector', 'swagger')
+          },
+          {
+            postgres: "[username] @@ to_tsvector('swagger')"
+          }
+        );
+      });
+    }
+
     describe('fn', () => {
     describe('fn', () => {
       it('{name: this.sequelize.fn(\'LOWER\', \'DERP\')}', function() {
       it('{name: this.sequelize.fn(\'LOWER\', \'DERP\')}', function() {
         expectsql(sql.whereQuery({ name: this.sequelize.fn('LOWER', 'DERP') }), {
         expectsql(sql.whereQuery({ name: this.sequelize.fn('LOWER', 'DERP') }), {