Movine is a simple database migration manager that aims to be compatible with real-world migration work. Many migration managers get confused with complicated development strategies for migrations. Oftentimes migration managers do not warn you if the SQL saved in git differs from what was actually run on the database. Movine solves this issue by keeping track of the unique hashes for the up.sql and down.sql for each migration, and provides tools for fixing issues. This allows users to easily keep track of whether their local migration history matches the one on the database.
This project is currently in early stages.
Movine does not aim to be an ORM. Consider diesel instead if you want an ORM.
Movine keeps track of four different states of migrations on the database. There are the basic ones:
Then there are the more complicated ones, which Movine was specifically designed to handle:
A 3.5 minute video showcasing the various tools Movine provides.
The first step to get started with Movine is to set the configuration. Configuration can be supplied either through a movile.toml file or environment variables:
If Movine finds a config file named movine.toml it will use the parameters specified.
```toml [postgres] host = {host} database = {db} user = {username} password = {pass} port = {port} sslrootcert = {cert filename}
[sqlite] file={file}
databaseurl={urlstring} ```
Note: SSLRootCert currently does not work when supplying a database_url. Note: You should only specify connection details for one database type, or Movine will implicitly choose one
You can configure the PostgreSQL adaptor using the environment variables described in the PostgreSQL documentation. Specifically PGHOST, PGPORT, PGDATABASE, PGUSER, and PGPASSWORD and PGSSLROOTCERT are supported.
You can configure the SQLite adaptor using an SQLITE_FILE environment variable.
Finally, you can also supply a DATABASE_URL environment variable.
Note: SSLRootCert does not work when using a database URL.
Movine supports .env files as a source of configuration.
Next, you can run the init command to set everything up, the generate command to create your first migration, and once those are written you can run up to apply them.
```
$ movine init
$ tree migrations/
migrations/
└── 1970-01-01-000000movineinit
├── down.sql
└── up.sql
1 directory, 2 files $ movine generate createnewtable $ tree migrations/ migrations/ ├── 1970-01-01-000000movineinit │ ├── down.sql │ └── up.sql └── 2019-03-17-163451createnew_table ├── down.sql └── up.sql
2 directories, 4 files $ movine up $ movine status 2019-03-17 16:34:51 UTC - Applied 2019-03-17-163451createnewtable 1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000movine_init ```
There are a few commands that Movine uses, and all of them can be listed by using --help on the command line.
The init command will run the initialization routine for Movine, which will create a table on the database to keep track of migrations and create a local migrations folder.
```
$ movine init
$ ls
migrations/ movine.toml
$ tree migrations/
migrations/
└── 1970-01-01-000000movineinit
├── down.sql
└── up.sql
1 directory, 2 files $ psql $PARAMS -c "\d" List of relations Schema | Name | Type | Owner --------+--------------------------+----------+-------- public | movinemigrations | table | movine public | movinemigrationsidseq | sequence | movine ```
The generate command will generate a folder with the current date and the given name in the migrations/ directory with blank up.sql and down.sql files.
```
$ movine generate createnewtable
$ tree migrations/
migrations/
├── 1970-01-01-000000movineinit
│ ├── down.sql
│ └── up.sql
└── 2019-03-17-163451createnew_table
├── down.sql
└── up.sql
2 directories, 4 files ```
The status command will tell you the current state of all migrations, both local and on the database.
$ movine status
2019-03-17 16:34:51 UTC - Pending 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
The up command will run all pending migrations. You can also run with the -p flag to show the migration plan without running it. This is true for all commands that modify the database and is useful for seeing if Movine will do what you expect.
$ movine up -p
1. Up - 2019-03-17-163451_create_new_table
$ movine status
2019-03-17 16:34:51 UTC - Pending 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
$ movine up
$ movine status
2019-03-17 16:34:51 UTC - Applied 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
The down command will rollback the most recent migration.
$ movine down
$ movine status
2019-03-17 16:34:51 UTC - Pending 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
The redo command will rollback and then re-apply the most recent applied migration or variant migration.
Note: If the latest migration is divergent then redo will simply skip it. Be careful, and run fix if you want to fix divergent migrations.
$ movine status
2019-03-17 16:34:51 UTC - Variant 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
$ movine redo
$ movine status
2019-03-17 16:34:51 UTC - Applied 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
The fix command will rollback everything until there are no divergent or variant migrations, and then apply all migrations except the migrations that were pending at the start.
$ movine status
2019-03-17 16:41:07 UTC - Pending 2019-03-17-164107_create_another_table
2019-03-17 16:40:59 UTC - Divergent 2019-03-17-164059_modify_table
2019-03-17 16:34:51 UTC - Variant 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
$ movine fix
$ movine status
2019-03-17 16:41:07 UTC - Pending 2019-03-17-164107_create_another_table
2019-03-17 16:34:51 UTC - Applied 2019-03-17-163451_create_new_table
1970-01-01 00:00:00 UTC - Applied 1970-01-01-000000_movine_init
The custom command will allow you to specify your own migration strategy (in case Movine is not smart enough). Note: this is currently not implemented
Note: While the Movine implementation is stable at this point, the config API may be in flux (specifically the helper functions). Please let me know any feedback!
Movine can be used as a library like so (using helper functions to load the database connection): ```rust use movine::{Movine, Config}; use movine::errors::Error;
fn main() -> Result<(), Error> { let config = Config::load(&"movine.toml")?; let mut conn = config.intosqliteconn(); let mut movine = Movine::new(&mut conn); movine.up()?; Ok(()) } ```
Or if you already have a connection: ```rust use movine::{Movine, Config}; use movine::errors::Error;
fn main() -> Result<(), Error> { // Same concept with a postgres connection! let mut conn = rusqlite::Connection::open("file.db")?; let mut movine = Movine::new(&mut conn); movine.up()?; Ok(()) } ```