manifest 0.1.2

Use an LLM to execute code

Manifest

man·i·fest [verb]

: to make something happen by imagining it and consciously thinking that it will happen

Manifest is an experiment in letting an LLM provide the return value for a function. It allows you to manifest powerful behaviors into existence merely by defining the function's facade.

Shout out to @dstufft for gifting me the PyPI repo name manifest 🙏

pip install manifest

Examples

Sentiment analysis

from manifest import ai

@ai
def is_pessimistic(text: str) -> bool:
    """ Determines if the text is pessimistic"""
    ...

assert is_pessimistic("This is a silly tool and nobody will ever use it.")

Translation

from manifest import ai

@ai
def translate(english_text: str, target_lang: str) -> str:
    """ Translates text from english into a target language """
    ...

assert translate("Hello", "fr") == "Bonjour"

Image analysis

You can pass in bytes to make use of a model's multimodal abilities. COMING SOON

import io
from manifest import ai

@ai
def breed_of_dog(image: io.BytesIO) -> str:
    """ Determines the breed of dog from a photo """
    ...

image = open("/path/to/terrier.jpg", "r")
print(breed_of_dog(image))

Complex objects

Your function can use fairly complex data structures.

from dataclasses import dataclass
from manifest import ai

@dataclass
class Actor:
    name: str
    character: str

@dataclass
class Movie:
    title: str
    director: str
    year: int
    top_cast: list[Actor]

@ai
def similar_movie(movie: str, before_year: int | None=None) -> Movie:
    """Discovers a similar movie, before a certain year, if the year is
    provided."""
    ...

like_inception = similar_movie("Inception")
print(like_inception)

How does it work?

Manifest relies heavily on runtime metadata, such as a function's name, docstring, arguments, and type hints. It uses all of these to compose a prompt behind the scenes, then sends the prompt to an LLM. The LLM "executes" the prompt, and returns a json-based format that we can safely parse back into the appropriate object.

To get the most out the @ai decorator:

• Name your function well.
• Add type hints to your function.
• Add a high-value docstring to your function.

Limitations

You can only pass in and return the following types:

• Dataclasses
• Enum subclasses
• primitives (str, int, bool, None, etc)
• basic container types (list, dict)
• unions
• Any combination of the above

The prompt templates are also a little fiddly sometimes. They can be improved.

Initialization

To make things super simple, manifest uses ambient LLM credentials, currently just OPENAI_API_KEY. If environment credentials are not found, you will be instructed to initialize the library yourself.