Dynamic Types

Sometimes you have a output schemas that change at runtime — for example if you have a list of Categories that you need to classify that come from a database, or your schema is user-provided.

Dynamic types are types that can be modified at runtime, which means you can change the output schema of a function at runtime.

Here are the steps to make this work:

Add @@dynamic to the class or enum definition to mark it as dynamic

baml

1 enum Category {
2   VALUE1 // normal static enum values that don't change
3   VALUE2
4   @@dynamic // this enum can have more values added at runtime
5 } 
6 
7 function DynamicCategorizer(input: string) -> Category {
8   client GPT4
9   prompt #"
10     Given a string, classify it into a category
11     {{ input }}
12      
13     {{ ctx.output_format }}
14   "#
15 }

Create a TypeBuilder and modify the existing type. All dynamic types you define in BAML will be available as properties of TypeBuilder. Think of the typebuilder as a registry of modified runtime types that the baml function will read from when building the output schema in the prompt.

1 from baml_client.type_builder import TypeBuilder
2 from baml_client import b
3 
4 async def run():
5   tb = TypeBuilder()
6   tb.Category.add_value('VALUE3')
7   tb.Category.add_value('VALUE4')
8   # Pass the typebuilder in the baml_options argument -- the last argument of the function.
9   res = await b.DynamicCategorizer("some input", { "tb": tb })
10   # Now res can be VALUE1, VALUE2, VALUE3, or VALUE4
11   print(res)

Dynamic BAML Classes

Existing BAML classes marked with @@dynamic will be available as properties of TypeBuilder.

BAML

1 class User {
2   name string
3   age int
4   @@dynamic
5 }
6 
7 function DynamicUserCreator(user_info: string) -> User {
8   client GPT4
9   prompt #"
10     Extract the information from this chunk of text:
11     "{{ user_info }}"
12      
13     {{ ctx.output_format }}
14   "#
15 }

Modify the User schema at runtime:

1 from baml_client.type_builder import TypeBuilder
2 from baml_client import b
3 
4 async def run():
5   tb = TypeBuilder()
6   tb.User.add_property('email', tb.string())
7   tb.User.add_property('address', tb.string())
8   res = await b.DynamicUserCreator("some user info", { "tb": tb })
9   # Now res can have email and address fields
10   print(res)

Creating new dynamic classes or enums not in BAML

Here we create a new Hobbies enum, and a new class called Address.

1 from baml_client.type_builder import TypeBuilder
2 from baml_client import b
3 
4 async def run():
5   tb = TypeBuilder()
6   const hobbiesEnum = tb.add_enum('Hobbies')
7   hobbiesEnum.add_value('Soccer')
8   hobbiesEnum.add_value('Reading')
9 
10   address_class = tb.add_class('Address')
11   address_class.add_property('street', tb.string())
12 
13   tb.User.add_property('hobby', hobbiesEnum.type().optional())
14   tb.User.add_property('address', addressClass.type().optional())
15   res = await b.DynamicUserCreator("some user info", { "tb": tb })
16   # Now res might have the hobby property, which can be Soccer or Reading
17   print(res)

Adding descriptions to dynamic types

1 tb = TypeBuilder()
2 tb.User.add_property("email", tb.string()).description("The user's email")

Building dynamic types from JSON schema

We have a working implementation of this, but are waiting for a concrete use case to merge it. Please chime in on the GitHub issue if this is something you’d like to use.

1 import pydantic
2 from baml_client import b
3 
4 class Person(pydantic.BaseModel):
5     last_name: list[str]
6     height: Optional[float] = pydantic.Field(description="Height in meters")
7 
8 tb = TypeBuilder()
9 tb.unstable_features.add_json_schema(Person.model_json_schema())
10 
11 res = await b.ExtractPeople(
12     "My name is Harrison. My hair is black and I'm 6 feet tall. I'm pretty good around the hoop. I like giraffes.",
13     {"tb": tb},
14 )

1	enum Category {
2	VALUE1 // normal static enum values that don't change
3	VALUE2
4	@@dynamic // this enum can have more values added at runtime
5	}
6
7	function DynamicCategorizer(input: string) -> Category {
8	client GPT4
9	prompt #"
10	Given a string, classify it into a category
11	{{ input }}
12
13	{{ ctx.output_format }}
14	"#
15	}

1	from baml_client.type_builder import TypeBuilder
2	from baml_client import b
3
4	async def run():
5	tb = TypeBuilder()
6	tb.Category.add_value('VALUE3')
7	tb.Category.add_value('VALUE4')
8	# Pass the typebuilder in the baml_options argument -- the last argument of the function.
9	res = await b.DynamicCategorizer("some input", { "tb": tb })
10	# Now res can be VALUE1, VALUE2, VALUE3, or VALUE4
11	print(res)

1	class User {
2	name string
3	age int
4	@@dynamic
5	}
6
7	function DynamicUserCreator(user_info: string) -> User {
8	client GPT4
9	prompt #"
10	Extract the information from this chunk of text:
11	"{{ user_info }}"
12
13	{{ ctx.output_format }}
14	"#
15	}

1	tb = TypeBuilder()
2	tb.User.add_property("email", tb.string()).description("The user's email")

1	import pydantic
2	from baml_client import b
3
4	class Person(pydantic.BaseModel):
5	last_name: list[str]
6	height: Optional[float] = pydantic.Field(description="Height in meters")
7
8	tb = TypeBuilder()
9	tb.unstable_features.add_json_schema(Person.model_json_schema())
10
11	res = await b.ExtractPeople(
12	"My name is Harrison. My hair is black and I'm 6 feet tall. I'm pretty good around the hoop. I like giraffes.",
13	{"tb": tb},
14	)