easy-psql

Description

Welcome to the easy-psql documentation! easy-psql is a simple intermediary for querying data in PostgreSQL databases. Whether you're a beginner or an experienced developer, this documentation will help you get started with easy-psql and leverage its capabilities to interact with your PostgreSQL databases efficiently.

Note: This package is intended for personal usage, and no tests have been written for it. Therefore, it is recommended to use it with caution and at your own responsibility. Like any software, there may be unforeseen bugs or issues that could affect your application. It is advisable to thoroughly review the package's functionality and integrate it into your projects with careful consideration of potential risks. (This package was developed using nodejs 16)

Installation

To install easy-psql, you can use npm:

npm install easy-psql

Establishing Connection with PostgreSQL Database

const { DB } = require("easy-psql");

DB.registerConnectionConfig({
  user: "postgres",
  database: "postgres",
  schema: "public",
  password: "postgres",
  port: 5432,
  host: "localhost",
  min: 5, // pool size limits
  max: 10, // pool size limits
});

Defining Models and Relations for PostgreSQL Database

const { Model, Column, Relation } = require("easy-psql");

class Role extends Model {
  constructor(connection) {
    super("roles", connection, "public" /* defaults to public */);
  }

  columns = {
    id: new Column({
      name: "id",
      type: "uuid",
      primary: true,
      defaultValue: "gen_random_uuid()",
    }),
    name: new Column({
      name: "name",
      type: "text",
      nullable: false,
    }),
  };
}

class User extends Model {
  constructor(connection) {
    super("users", connection);
  }

  columns = {
    id: new Column({
      name: "id",
      type: "uuid",
      primary: true,
      defaultValue: "gen_random_uuid()",
    }),
    role_id: new Column({
      name: "role_id",
      type: "uuid",
    }),
    email: new Column({
      name: "email",
      type: "text",
      unique: true,
    }),
    password: new Column({
      name: "password",
      type: "text",
      nullable: false,
    }),
    grid_order: new Column({
      name: "grid_order",
      type: "int",
      nullable: false,
      defaultValue: 1,
    }),
  };

  relations = {
    role: new Relation({
      from_table: "users",
      from_column: "role_id",
      to_table: "roles",
      to_column: "id",
      alias: "role",
      type: "object", // object or array
    }),
  };
}

// Register the models in order to be able to use relations
DB.register(User);
DB.register(Role);

JSON / JSONB path helpers

For json and jsonb columns you can generate fragments for updates and where clauses. Import JsonUpdateOperators and JsonWhereClauseOperators from easy-psql. Each method returns an internal SQL object that the ORM expands when you pass it as a column value in update or where.

Use …Jsonb methods when the underlying column type is jsonb, and …Json when it is json (plain JSON). The json variants use different casts and operators in PostgreSQL, so pick the suffix that matches the column type.

Update helpers (JsonUpdateOperators)

| Method | Purpose | | -------- | -------- | | setJsonb(column, path, value, createMissingKeys?) | JSONB_SET | | unsetJsonb(column, path) / unsetJson(column, path) | delete key or path (jsonb / json delete operators) | | insertJsonb(column, path, value, insertAfter?) | jsonb_insert | | mergeJsonb(column, value) | merge object with \|\| | | arrayAppendJsonb, arrayPrependJsonb | append / prepend one element to an array at path | | arrayMultiAppendJsonb, arrayMultiPrependJsonb | append / prepend multiple elements | | arrayRemoveAtJsonb(column, path, index) | remove array element by index | | counterJsonb(column, path, delta?) | numeric increment at path | | booleanToggleJsonb(column, path) | flip a boolean at path |

path is either a string key or a string array for nested paths (e.g. ['tags', '0']).

Where helpers (JsonWhereClauseOperators)

| Method | Typical use | | -------- | ----------- | | equalsJsonb / equalsJson, notEqualsJsonb / notEqualsJson | equality on extracted JSON value | | greaterThanJsonb / greaterThanJson, greaterThanOrEqual… | ordered comparisons | | lessThanJsonb / lessThanJson, lessThanOrEqual… | ordered comparisons | | containsJsonb / containsJson | any of the given strings exists as key or array element (?|); pass a string array (not a JSON string) | | containedInJsonb / containedInJson | left JSON is contained in right (<@) |

Example:

const {
  DB,
  Model,
  Column,
  JsonUpdateOperators,
  JsonWhereClauseOperators,
} = require("easy-psql");

class Building extends Model {
  constructor(conn) {
    super("buildings", conn);
  }
  columns = {
    id: new Column({ name: "id", type: "text", primary: true }),
    config: new Column({ name: "config", type: "jsonb", nullable: true }),
  };
}
DB.register(Building);

const b = new Building();

await b.update({
  update: {
    config: JsonUpdateOperators.mergeJsonb("config", { version: 1 }),
  },
  where: {
    config: JsonWhereClauseOperators.equalsJsonb("config", "version", 1),
  },
});

Basic examples

const model = new User();
const data = await model.find() // multiple rows
const data = await model.findOne() // single row
const data = await model.create({email:'[email protected]',password:'12345678'});
const data = await model.createMany([{email:'[email protected]',password:'12345678'},{email:'[email protected]',password:'12345678'}])
const data = await model.update({where:{...},update:{email:'[email protected]'}});
const data = await model.delete({where: {...}});
const {count,max} = await model.aggregate({_count:true,max:{email:true},where:{...}}) // _count,max,min,sum,avg

Use Relations

const model = new User();
const usersWithRoles = await model.find({ include: { role: true } });

Use Relations With Filters

const model = new User();
const usersWithRoles = await model.find({ include: { role: {where:{...}} } });

Basic Filtering Example

const model = new User();
const data = await model.find({
  where: { email: { _eq: "[email protected]" } },
});

Basic Filtering Operators

const model = new User();
const data = await model.find({
  where: {
    password: { _in: ["12345678", "123456789"] },
    email: { _eq: "[email protected]" },
    _or: [{ password: { _is_not: null } }, { password: { _is: null } }],
    _and: [
      { grid_order: { _lte: 200 } },
      { grid_order: { _lt: 201 } },
      { grid_order: { _gte: 1 } },
      { grid_order: { _gt: 0 } },
      { password: { _nin: ["12345678", "123456789"] } },
      { email: { _ilike: "test" } },
      { email: { _neq: "[email protected]" } },
      {
        _or: [
          { password: { _is_not: null } },
          { password: { _is: null } },
          {
            _and: [
              { email: { _neq: "[email protected]" } },
              {
                _or: [
                  { password: { _is_not: null } },
                  { password: { _is: null } },
                ],
              },
            ],
          },
        ],
      },
    ],
  },
});

Basic Filtering Operators Are Applied to Relations too

const model = new User();
const data = await model.find({
  where: {
    password: { _eq_: "12345678" },
    email: { _eq: "[email protected]" },
    // this is the role related entity
    role: {
      name: { _eq: "admin" },
    },
  },
});

Other Operators

const model = new Role();
const data = await model.find({where:{...},distinct:[...],groupBy:[...],limit:100,offset:0,orderBy: {
    name: 'asc',
    users_aggregate:{
        _count:true
    }
}})

Using nested aggregations

const { Model, Column } = require("easy-psql");

class Role extends Model {
  constructor(connection) {
    super("roles", connection);
  }

  columns = {
    id: new Column({
      name: "id",
      type: "uuid",
      primary: true,
      defaultValue: "gen_random_uuid()",
    }),
    name: new Column({
      name: "name",
      type: "text",
      nullable: false,
    }),
  };

  relations = {
    users: new Relation({
      from_table: "roles",
      from_column: "id",
      to_table: "users",
      to_column: "role_id",
      alias: "users",
      type: "array", // object or array
    }),
  };
}

class User extends Model {
  constructor(connection) {
    super("users", connection);
  }

  columns = {
    id: new Column({
      name: "id",
      type: "uuid",
      primary: true,
      defaultValue: "gen_random_uuid()",
    }),
    role_id: new Column({
      name: "role_id",
      type: "uuid",
    }),
    email: new Column({
      name: "email",
      type: "text",
      unique: true,
    }),
    password: new Column({
      name: "password",
      type: "text",
      nullable: false,
    }),
    grid_order: new Column({
      name: "grid_order",
      type: "int",
      nullable: false,
      defaultValue: 1,
    }),
  };

  relations = {
    role: new Relation({
      schema: "public", // refers to the schema of the referenced table. In this example it is "roles"
      from_table: "users",
      from_column: "role_id",
      to_table: "roles",
      to_column: "id",
      alias: "role",
      type: "object", // object or array
    }),
  };
}

// Register the models in order to be able to use relations
DB.register(User);
DB.register(Role);

const model = new Role();
const role = await model.findOne({
  where: { name: { _eq: "admin" } },
  include: {
    users_aggregate: {
      // _count,max,min,avg,sum
      _count: true,
      max: { grid_order: true },
      min: { grid_order: true, email: true },
      // you can use where inside here
    },
  },
});

Using multiple nested relations

const model = new User();
const data = await model.find({
  include: {
    role: {
      include: {
        users: {
          include: {
            role: true, // .....
          },
        },
      },
    },
  },
});

Transactions Example 1

const model = new User();

const result = await model.withTransaction(async (tx) => {
  const newUser = await model.create({
    email: "[email protected]",
    password: "12345678",
    role_id: "...",
  });
  if (!newUser) {
    throw new Error("User not created");
  }

  return newUser;
});

// Errors inside withTransaction will use rollback; otherwise a commit is performed

// Check the result

if (result instanceof Error) {
  console.error(result);
} else {
  console.log(result);
}

Transactions Example 2

const model = new User();

const result = await model.withTransaction(async (tx) => {
  const newUser = await model.create({
    email: "[email protected]",
    password: "12345678",
    role_id: "...",
  });
  if (!newUser) {
    throw new Error("User not created");
  }

  const roleModel = new Role(tx); // pass the tx -> connection to postgres otherwise this operation is not atomic

  const roleData = await role.create({ name: "testRole" });
  if (!roleData) {
    throw new Error("Role not created");
  }
  return [newUser, roleData];
});

// Errors inside withTransaction will use rollback; otherwise a commit is performed

// Check the result

if (result instanceof Error) {
  console.error(result);
} else {
  console.log(result);
}

Transactions Example 3

const model = new Role();
const inputData = {
  name: "newRole",
  users: [
    { email: "[email protected]", password: "12345678" },
    { email: "[email protected]", password: "12345678" },
  ],
};

// Over here perform the createTX operation

const result = await model.createTX(inputData);

// or for multiple inserts use await model.createManyTX([inputData])   -> use an array of objects

// The roles.id generated from this query will be assigned to each users.role_id field

Register Effects

// The following events will be triggered only by using the model classes

// e.g : find and findOne  on model User -> will trigger onSelectAsync but DB.pool.query('select * from users'); will not!

DB.onSelectAsync("users", async (data, instance) => {
  // ...
});
DB.onInsertAsync("users", async (data, instance) => {
  // ...
});
DB.onUpdateAsync("users", async (data, instance) => {
  // ...
});
DB.onDeleteAsync("users", async (data, instance) => {
  // ...
});
DB.onErrorAsync("users", async (error, instance) => {
  // ...
});

More generic with actions

// The following actions will be triggered only by using the model classes

// e.g : find and findOne  on model User -> will trigger onSelectActionAsync but DB.pool.query('select * from users'); will not!

// Not like effects, you can register many actions of same type e.g Selection Actions and they are going to be executed using Promise.all in the end of each operation find,findOne,create,createTX,createMany,createManyTX,update,delete.

DB.onSelectActionAsync(async (data, instance) => {
  // ...
});
DB.onInsertActionAsync(async (data, instance) => {
  // ...
});
DB.onUpdateActionAsync(async (data, instance) => {
  // ...
});
DB.onDeleteActionAsync(async (data, instance) => {
  // ...
});
DB.onErrorActionAsync(async (error, instance) => {
  // ...
});

AutoDiscoverAPI

const { Postgres, Relation } = require('easy-psql');

const db = new Postgres({
  connectionConfig: {
    host: ...,
    port: ...,
    user: ...,
    password: ...,
    schema: ...,
    database: ...
  },
  relations: [new Relation({...}), {from_table:...,from_column:...,to_table:...,to_column:...,alias:...,type: 'object' | 'array'}], // not required
  options: {
      createFiles: false, // create files for the models
      skipIfDirectoryExists: true, // skip files' creation if the specified folder already exists
      dirname: "easy-psql-models", // folder name to create model files
      useESM: false, // true -> use exports / false -> require
      extension: "js", // can be js,ts or mjs
  }
});

const postgres = async () => {
  await db.init()
  return db
}


module.exports = postgres


// e.g in another file

const postgres = require('./path-of-the-file-above');

const run = async () => {
  const db = await postgres();
  // db.model(table_name,connection,schema)
  const data = await db.model('my-table').find({...}); //same api as the functions in the other examples

  return data; // ([{...},{...}])
}


run();