Golang Database Management and Code Generation
Branch: master
Clone or download
Latest commit 0a80fcf Feb 8, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.github Add CONTRIBUTORS file Apr 20, 2018
cmd Bump version Dec 20, 2018
doc/img
fake Simplify the routine provider May 16, 2018
integration Remove keep-schema flag May 10, 2018
script Add Build script Dec 20, 2018
sqlexec Split statements by GO separator May 24, 2018
sqlmigr
sqlmodel Use schema package to generate sql column types Dec 20, 2018
.gitignore Split statements by GO separator May 24, 2018
.travis.yml Migrate to Go modules Dec 18, 2018
CONTRIBUTORS Add CONTRIBUTORS file Apr 20, 2018
LICENSE Add License Mar 29, 2018
README.md Update README.md Feb 8, 2019
common.go Improve test coverage May 8, 2018
common_test.go Improve test coverage May 8, 2018
go.mod Migrate to Go modules Dec 18, 2018
go.sum Use schema package to generate sql column types Dec 20, 2018
suite_test.go Move Prana as separate package Apr 24, 2018

README.md

Prana

Documentation [License Build Status Coverage Go Report Card

Golang Database Manager

Prana

Overview

Prana is a package for rapid application development with relational databases in Golang. It has a command line interface that provides:

  • SQL Migrations
  • Embedded SQL Scripts
  • Model generation from SQL schema

Installation

GitHub

$ go get -u github.com/phogolabs/prana
$ go install github.com/phogolabs/prana/cmd/prana

Homebrew (for Mac OS X)

$ brew tap phogolabs/tap
$ brew install prana

Introduction

Note that Prana is in BETA. We may introduce breaking changes until we reach v1.0.

SQL Migrations

Each migration is a SQL script that contains two operations for upgrade and rollback. They are labelled with the following comments:

  • -- name: up for upgrade
  • -- name: down for revert

In order to prepare the project for migration, you have to set it up:

$ prana migration setup

Then you can create a migration with the following command:

$ prana migration create schema

The command will create the following migration file in /database/migration:

$ tree database

database/
└── migration
    ├── 00060524000000_setup.sql
    └── 20180329162010_schema.sql

The 20180329162010_schema.sql migration has similar to example below format:

-- Auto-generated at Thu Mar 29 16:20:10 CEST 2018
-- Please do not change the name attributes

-- name: up
CREATE TABLE users (
  id INT PRIMARY KEY NOT NULL,
  first_name TEXT NOT NULL,
  last_name TEXT
);

GO; -- split execution of the migration

INSERT INTO users (id, first_name, last_name) VALUES (1, 'John', 'Doe');

-- name: down
DROP TABLE IF EXISTS users;

You can run the migration with the following command:

$ prana migration run

If you want to rollback the migration you have to revert it:

$ prana migration revert

If you have an SQL script that is compatible with particular database, you can append the database's driver name suffix. For instance if you want to run part of a particular migration for MySQL, you should have the following directory tree:

$ tree database

database/
└── migration
    ├── 00060524000000_setup.sql
    ├── 20180406190015_users.sql
    ├── 20180406190015_users_mysql.sql
    └── 20180406190015_users_sqlite3.sql

Prana will execute the following migrations with users suffix, when MySQL driver is used:

  • 20180406190015_users.sql
  • 20180406190015_users_mysql.sql

Presently the following suffixes are supported:

  • sqlite3
  • mysql
  • postgres

SQL Schema and Code Generation

Let's assume that we want to generate a mode for the users table.

You can use the prana command line interface to generate a package that contains Golang structs, which represents each table from the desired schema.

For that purpose you should call the following subcommand:

$ prana model sync

By default the command will place the generated code in single schema.go file in $PWD/database/model package for the default database schema. Any other schemas will be placed in the same package but in separate files. You can control the behavior by passing --keep-schema flag which will cause each schema to be generated in own package under the /$PWD/database/model package.

You can print the source code without generating a package by executing the following command:

$ prana model print

Note that you can specify the desired schema or tables by providing the correct arguments.

If you pass --extra-tag argument, you can specify which tag to be included in your final result. Supported extra tags are:

The model representation of the users table is:

package model

import "github.com/phogolabs/schema"

// User represents a data base table 'users'
type User struct {
	// ID represents a database column 'id' of type 'INT PRIMARY KEY NOT NULL'
	ID int `db:"id,primary_key,not_null" json:"id" xml:"id" validate:"required"`

	// FirstName represents a database column 'first_name' of type 'TEXT NOT NULL'
	FirstName string `db:"first_name,not_null" json:"first_name" xml:"first_name" validate:"required"`

	// LastName represents a database column 'last_name' of type 'TEXT NULL'
	LastName schema.NullString `db:"last_name,null" json:"last_name" xml:"last_name" validate:"-"`
}

Note that the code generation depends on two packages. In order to produce a source code that compiles you should have in your $GOPATH/src directory installed:

The generated db tag is recognized by orm.Gateway as well as sqlx.

If you wan to generate models for gorm, you should pass --orm-tag gorm. Note that constraints like unique or indexes are not included for now.

$ prana model --orm-tag gorm -e json -e xml -e validate sync

The command above will produce the following model:

package model

import "github.com/phogolabs/schema"

// User represents a data base table 'users'
type User struct {
	// ID represents a database column 'id' of type 'INT PRIMARY KEY NOT NULL'
	ID int `gorm:"column:id;type:int;primary_key;not null" json:"id" xml:"id" validate:"required"`

	// FirstName represents a database column 'first_name' of type 'TEXT NOT NULL'
	FirstName string `gorm:"column:first_name;type:text;not null" json:"first_name" xml:"first_name" validate:"required"`

	// LastName represents a database column 'last_name' of type 'TEXT NULL'
	LastName schema.NullString `gorm:"column:last_name;type:text;null" json:"last_name" xml:"last_name" validate:"-"`
}

SQL Scripts and Commands

Also, it provides a way to work with embeddable SQL scripts which can be executed easily by ORM as SQL Routines. You can see the ORM example to understand more about that. First of all you have create a script that contains your SQL statements.

The easies way to generate a SQL script with correct format is by using prana command line interface:

$ prana routine create show-sqlite-master

The command above will generate a script in your $PWD/database/routine;

$ tree database/

database/
└── routine
    └── 20180328184257.sql

You can enable the script for particular type of database by adding the driver name as suffix: 20180328184257_slite3.sql.

It has the following contents:

-- Auto-generated at Wed Mar 28 18:42:57 CEST 2018
-- name: show-sqlite-master
SELECT type,name,rootpage FROM sqlite_master;

The -- name: show-sqlite-master comment define the name of the command in your SQL script. The SQL statement afterwards is considered as the command body. Note that the command must have only one statement.

Then you can use the prana command line interface to execute the command:

$ prana script run show-sqlite-master

Running command 'show-sqlite-master' from '$PWD/database/script'
+-------+-------------------------------+----------+
| TYPE  |             NAME              | ROOTPAGE |
+-------+-------------------------------+----------+
| table | migrations                    |        2 |
| index | sqlite_autoindex_migrations_1 |        3 |
+-------+-------------------------------+----------+

You can also generate all CRUD operations for given table. The command below will generate a SQL script that contains SQL queries for each table in the default schema:

$ prana routine sync

It will produce the following script in $PWD/database/rotuine:

-- name: select-all-users
SELECT * FROM users;

-- name: select-user
SELECT * FROM users
WHERE id = ?;

-- name: insert-user
INSERT INTO users (id, first_name, last_name)
VALUES (?, ?, ?);

-- name: update-user
UPDATE users
SET first_name = ?, last_name = ?
WHERE id = ?;

-- name: delete-user
DELETE FROM users
WHERE id = ?;

Command Line Interface Advance Usage

By default the CLI work with sqlite3 database called prana.db at your current directory.

Prana supports:

  • PostgreSQL
  • MySQL
  • SQLite

If you want to change the default connection, you can pass it via command line argument:

$ prana --database-url [driver-name]://[connection-string] [command]

prana uses a URL schema to determines the right database driver. If you want to pass the connection string via environment variable, you should export PRANA_DB_URL.

Help

For more information, how you can change the default behavior you can read the help documentation by executing:

$ prana -h

NAME:
   prana - Golang Database Manager

USAGE:
   prana [global options]

VERSION:
   1.0

COMMANDS:
     migration  A group of commands for generating, running, and reverting migrations
     model      A group of commands for generating object model from database schema
     routine    A group of commands for generating, running, and removing SQL commands
     help, h    Shows a list of commands or help for one command

GLOBAL OPTIONS:
   --database-url value  Database URL (default: "sqlite3://prana.db") [$PRANA_DB_URL]
   --log-format value    format of the logs [$PRANA_LOG_FORMAT]
   --log-level value     level of logging (default: "info") [$PRANA_LOG_LEVEL]
   --help, -h            show help
   --version, -v         print the version

Contributing

We are welcome to any contributions. Just fork the project.

logo made by Free Pik