Schema Reference

Aprisma uses the standard Prisma schema language. This page covers the features Aprisma supports and any Aprisma-specific generator options.

Generator block

generator client {
  provider  = "aprisma"
  output    = "./prisma"      # where to write the generated client
  interface = "asyncio"       # only asyncio is supported
}

Option	Values	Description
`provider`	`"aprisma"`	Required — selects the Aprisma generator
`output`	path	Directory for the generated `client.py`
`interface`	`"asyncio"`	Client interface style

Datasource block

datasource db {
  provider = "sqlite"   # or "postgresql"
}

No url is set in the schema — the connection URL is passed to db.connect() at runtime. The provider field is used by the Prisma CLI for migrations.

Scalar types

Prisma type	Python type	Notes
`String`	`str`
`Int`	`int`
`BigInt`	`int`	Python `int` handles arbitrary precision
`Float`	`float`
`Decimal`	`Decimal`	`decimal.Decimal`
`Boolean`	`bool`
`DateTime`	`datetime`	`datetime.datetime` (naive UTC)
`Bytes`	`bytes`
`Json`	`Any`	Stored as text; round-trips as Python object

Append ? to make any field nullable (String?, Int?, etc.). A nullable field is None in Python when no value is stored.

Default values

id        Int      @id @default(autoincrement())
createdAt DateTime @default(now())
isActive  Boolean  @default(true)
score     Float    @default(0.0)

Field attributes

Attribute	Description
`@id`	Marks the primary key
`@unique`	Adds a unique constraint; field can be used in `find_unique`
`@default(...)`	Sets a default value
`@map("col_name")`	Maps the Prisma field to a different column name
`@updatedAt`	Automatically set to the current time on every update

Model attributes

Attribute	Description
`@@id([fieldA, fieldB])`	Composite primary key
`@@unique([fieldA, fieldB])`	Composite unique constraint
`@@index([fieldA, fieldB])`	Creates an index (handled by Prisma CLI)
`@@map("table_name")`	Maps the model to a different table name
`@@ignore`	Excludes the model from the generated client

Relations

One-to-many

model User {
  id    Int    @id @default(autoincrement())
  posts Post[]
}

model Post {
  id       Int  @id @default(autoincrement())
  authorId Int
  author   User @relation(fields: [authorId], references: [id])
}

Post.authorId is the foreign key. User.posts is the back-reference list.

One-to-one

model User {
  id      Int      @id @default(autoincrement())
  profile Profile?
}

model Profile {
  id     Int  @id @default(autoincrement())
  userId Int  @unique
  user   User @relation(fields: [userId], references: [id])
}

Many-to-many (explicit join table)

model Post {
  id   Int       @id @default(autoincrement())
  tags PostTag[]
}

model Tag {
  id    Int       @id @default(autoincrement())
  name  String    @unique
  posts PostTag[]
}

model PostTag {
  postId Int  @map("post_id")
  tagId  Int  @map("tag_id")
  post   Post @relation(fields: [postId], references: [id])
  tag    Tag  @relation(fields: [tagId], references: [id])

  @@id([postId, tagId])
}

Enums

enum Role {
  USER
  ADMIN
  MODERATOR
}

model User {
  id   Int    @id @default(autoincrement())
  role Role   @default(USER)
}

Enums are stored as strings. The generated Python client exposes them as str values matching the Prisma enum member names.

Unsupported types

Fields declared with Unsupported("...") are excluded from the generated client. Use raw queries (query_raw, execute_raw) to interact with those columns.