🦜 Parrot / type-safe SQL in gleam

How it works

This library makes use of a sqlc-gen-json plugin, which it then converts into gleam code.

So in a simplified manner, the pipeline looks like this:

graph LR
    SQL[SQL Queries] -- sqlc-gen-json --> JSON[JSON Schema]
    JSON -- parrot --> Gleam[Gleam Code]

An Example

Let’s take the following SQL queries as an example:

-- name: GetAuthor :one
SELECT
  *
FROM
  authors
WHERE
  id = ?
LIMIT
  1;

-- name: ListAuthors :many
SELECT
  *
FROM
  authors
ORDER BY
  name;

The --name: GetAuthor :one comment is part of sqlc and will be used to generate the name and return type of the wrapper.

Given the queries above, the following code will be generated:

//// file: src/gen/parrot.gleam

import gleam/option.{type Option}
import gleam/dynamic/decode
import gleam/time/timestamp.{type Timestamp}

// Custom decoders
fn datetime_decoder() -> decode.Decoder(Timestamp) {
  decode.string
  |> decode.then(fn(datetime_str) {
    case timestamp.parse_rfc3339(datetime_str) {
      Ok(ts) -> decode.success(ts)
      Error(_) -> decode.failure(timestamp.from_unix_seconds(0), "Invalid datetime format")
    }
  })
}

pub type GetAuthor {
  GetAuthor(
    id: Int,
    created_at: Timestamp,
    name: String,
    bio: Option(String)
  )
}

pub fn get_author(id: Int){
  let sql = "SELECT
  id, created_at, name, bio
FROM
  authors
WHERE
  id = ?
LIMIT
  1"
  #(sql, [sql.ParamInt(age)])
}

pub fn get_author_decoder() -> decode.Decoder(GetAuthor) {
  use id <- decode.field("id", decode.int)
  use created_at <- decode.field("created_at", datetime_decoder())
  use name <- decode.field("name", decode.string)
  use bio <- decode.field("bio", decode.optional(decode.string))
  decode.success(GetAuthor(id: , created_at: , name: , bio: ))
}

pub type ListAuthors {
  ListAuthors(
    id: Int,
    created_at: Timestamp,
    name: String,
    bio: Option(String)
  )
}

pub fn list_authors(){
  let sql = "SELECT
  id, created_at, name, bio
FROM
  authors
ORDER BY
  name"
  #(sql, Nil)
}

pub fn list_authors_decoder() -> decode.Decoder(ListAuthors) {
  use id <- decode.field("id", decode.int)
  use created_at <- decode.field("created_at", datetime_decoder())
  use name <- decode.field("name", decode.string)
  use bio <- decode.field("bio", decode.optional(decode.string))
  decode.success(ListAuthors(id: , created_at: , name: , bio: ))
}

Every SQL statement wrapper follows the schema of #(String, List(Params)). The first element is the raw SQL that can be executed by your database driver and the second element is a tuple of all of the parameters that you need for this query.

The query parameters are wrapped in a custom type that you can use to map them to your database driver’s types. Here is an example for lpil/sqlight:

import gleam/list
import gleam/time/calendar
import gleam/time/timestamp
import lpil_sqlight/sql
import parrot/sql as parrot
import sqlight

/// Generated sql module by parrot
import parrots/sql

pub fn params_to_sqlight(args: List(parrot.Param)) -> List(sqlight.Value) {
  list.map(args, fn(arg) {
    case arg {
      parrot.ParamInt(a) -> sqlight.int(a)
      parrot.ParamBool(a) -> sqlight.bool(a)
      parrot.ParamFloat(a) -> sqlight.float(a)
      parrot.ParamString(a) -> sqlight.text(a)
      parrot.ParamTimestamp(ts) -> {
        ts
        |> timestamp.to_rfc3339(calendar.utc_offset)
        |> sqlight.text
      }
    }
  })
}

pub fn main() {
  // ... setup database ...

  let #(raw_sql, args) = sql.get_author(7)
  let _ =
    echo sqlight.query(
      raw_sql,
      on: conn,
      with: params_to_sqlight(args),
      expecting: sql.get_author_decoder(),
    )
}

Edge Cases

“any” types

Usage

For parrot to work you need to make sure you have sqlc installed (guide) installed.

If you use MySQL, you also need mysqldump (comes per default if you have mysql installed)
If you use PostgreSQL, you also need pg_dump (comes per default if you have postgresql installed)

Install

$ gleam add parrot

Define your queries

Parrot will look for all *.sql files in any sql directory under your project’s src directory.
Each *.sql file can contain as many SQL queries as you want.
All of the queries will compile into a single src/parrots/sql.gleam module.

Here are some links to help you start out, if you are unfamiliar with the sqlc annotation syntax:

Generate Gleam code

$ gleam run -m parrot gen sqlite file.db
$ gleam run -m parrot gen mysql "mysql://user:password@127.0.0.1:3309/db"
$ gleam run -m parrot gen psql "mysql://user:password@127.0.0.1:3309/db"

Development

Database

There are scripts to spawn a MySQL or PostgreSQL docker container:

For example:

$ ./bin/mysql/docker.sh
# or
$ ./bin/psql/docker.sh

Running the project and tests

$ gleam run   # Run the project
$ gleam test  # Run the tests

FAQ

What flavour of SQL does parrot support?

This plugin supports everything that sqlc supports. As the time of this writing that would be MySQL, PostgreSQL and SQlite.

You can read more on language & SQL support here: https://docs.sqlc.dev/en/stable/reference/language-support.html

What sqlc features are not supported?

embeddeding structs (https://docs.sqlc.dev/en/stable/howto/embedding.html)

Acknowledgements

This project was heavily inspired by squirrel (Hex, GitHub). Thank you @giacomocavalieri!
Thank you to sqlc (GitHub, Website)
Thank you to dedent (GitHub, Hex)

License

Apache License, Version 2.0