-
Notifications
You must be signed in to change notification settings - Fork 349
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #759 from PrefectHQ/2.0
2.0
- Loading branch information
Showing
191 changed files
with
7,525 additions
and
6,157 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,136 +1,248 @@ | ||
| <p align="center"> | ||
| <img src="docs/img/heroes/it_hates_me_hero.png" style="width: 95%; height: auto;"/> | ||
| <img src="docs/assets/images/heroes/it_hates_me_hero.png" style="width: 95%; height: auto;"/> | ||
| </p> | ||
|
|
||
| # Marvin | ||
| [](https://badge.fury.io/py/marvin) | ||
| [](https://twitter.com/AskMarvinAI) | ||
| [](https://www.askmarvin.ai) | ||
| ### An AI engineering framework | ||
| ... made with 💙 by the team at [Prefect](https://www.prefect.io/). | ||
| [](https://twitter.com/AskMarvinAI) | ||
|
|
||
| # Marvin | ||
|
|
||
| ### The AI engineering toolkit | ||
|
|
||
| Marvin is a lightweight AI toolkit for building natural language interfaces that are reliable, scalable, and easy to trust. | ||
|
|
||
| Each of Marvin's tools is simple and self-documenting, using AI to solve common but complex challenges like entity extraction, classification, and generating synthetic data. Each tool is independent and incrementally adoptable, so you can use them on their own or in combination with any other library. Marvin is also multi-modal, supporting both image and audio generation as well using images as inputs for extraction and classification. | ||
|
|
||
| Marvin is for developers who care more about _using_ AI than _building_ AI, and we are focused on creating an exceptional developer experience. Marvin users should feel empowered to bring tightly-scoped "AI magic" into any traditional software project with just a few extra lines of code. | ||
|
|
||
| Marvin aims to merge the best practices for building dependable, observable software with the best practices for building with generative AI into a single, easy-to-use library. It's a serious tool, but we hope you have fun with it. | ||
|
|
||
| Marvin is open-source, free to use, and made with 💙 by the team at [Prefect](https://www.prefect.io/). | ||
|
|
||
| ## Installation | ||
|
|
||
| Install the latest version with `pip`: | ||
|
|
||
| ```bash | ||
| pip install marvin | ||
| pip install marvin -U | ||
| ``` | ||
| Getting started? Head over to our [setup guide](https://www.askmarvin.ai/welcome/installation/). | ||
|
|
||
| --- | ||
| To verify your installation, run `marvin version` in your terminal. | ||
|
|
||
| ## Tools | ||
|
|
||
| Marvin consists of a variety of useful tools, all designed to be used independently. Each one represents a common LLM use case, and is designed to package that power into a simple, self-documenting interface. | ||
|
|
||
| ### General | ||
|
|
||
| 🦾 [Write custom AI-powered functions](https://askmarvin.ai/docs/text/functions) without source code | ||
|
|
||
| ### Text | ||
|
|
||
| 🏷️ [Classify text](https://askmarvin.ai/docs/text/classification) into categories | ||
|
|
||
| 🔍 [Extract structured entities](https://askmarvin.ai/docs/text/extraction) from text | ||
|
|
||
| 🪄 [Transform text](https://askmarvin.ai/docs/text/transforming) into structured data | ||
|
|
||
| ✨ [Generate synthetic data](https://askmarvin.ai/docs/text/generation) from a schema | ||
|
|
||
| ⚠️ Marvin is a work in progress, and we'd love your [feedback](https://github.com/PrefectHQ/marvin/discussions)! ⚠️ | ||
| ### Images | ||
|
|
||
| > [Looking for info on Marvin 1.x?](/docs/help/legacy_docs.md) | ||
| 🖼️ [Create images](https://askmarvin.ai/docs/images/generation) from text or functions | ||
|
|
||
| --- | ||
| 📝 [Describe images](https://askmarvin.ai/docs/vision/captioning) with natural language | ||
|
|
||
| ## Offerings | ||
| 🏷️ [Classify images](https://askmarvin.ai/docs/vision/classification) into categories | ||
|
|
||
| Marvin's high-level abstractions are familiar Python interfaces that make it easy to leverage AI in your application. These interfaces aim to be simple and self-documenting, adding a touch of AI magic to everyday objects. | ||
| 🔍 [Extract structured entities](https://askmarvin.ai/docs/vision/extraction) from images | ||
|
|
||
| 🪄 [**AI Functions**](https://www.askmarvin.ai/components/ai_function/) for complex business logic and transformations | ||
| 🪄 [Transform images](https://askmarvin.ai/docs/vision/transforming) into structured data | ||
|
|
||
| 🧩 [**AI Models**](https://www.askmarvin.ai/components/ai_model/) for structuring text into type-safe schemas | ||
| ### Audio | ||
|
|
||
| 🤖 (*beta*) [**Assistants**](/src/marvin/beta/assistants/README.md) for building stateful natural language interfaces | ||
| ___ | ||
| 🎙️ [Generate speech](https://askmarvin.ai/docs/audio/speech) from text or functions | ||
|
|
||
| ### 🪄 AI Functions | ||
| AI Functions look like regular functions, but have no source code. Instead, an AI interprets their description and inputs to generate their outputs, making them ideal for general NLP applications like sentiment analysis. | ||
| ### Interaction | ||
|
|
||
| You can learn more about AI Functions [here](https://www.askmarvin.ai/components/ai_function/). | ||
| 🤖 [Chat with assistants](https://askmarvin.ai/docs/interactive/assistants) and use custom tools | ||
|
|
||
| 🧭 [Build applications](https://askmarvin.ai/docs/interactive/applications) that manage persistent state | ||
|
|
||
| ## Quickstart | ||
|
|
||
| Here's a whirlwind tour of a few of Marvin's main features. For more information, [check the docs](https://askmarvin.ai/welcome/what_is_marvin/)! | ||
|
|
||
| ### 🏷️ Classify text | ||
|
|
||
| Marvin can `classify` text using a set of labels: | ||
|
|
||
| ```python | ||
| from marvin import ai_fn | ||
| import marvin | ||
|
|
||
| @ai_fn | ||
| def sentiment(text: str) -> float: | ||
| """Given `text`, returns a number between 1 (positive) and -1 (negative) | ||
| indicating its sentiment score. | ||
| """ | ||
| marvin.classify( | ||
| "Marvin is so easy to use!", | ||
| labels=["positive", "negative"], | ||
| ) | ||
|
|
||
| # "positive" | ||
| ``` | ||
|
|
||
| sentiment("I love working with Marvin!") # 0.8 | ||
| sentiment("These examples could use some work...") # -0.2 | ||
| Learn more about classification [here](https://askmarvin.ai/docs/text/classification). | ||
|
|
||
| ### 🔍 Extract structured entities | ||
|
|
||
| Marvin can `extract` structured entities from text: | ||
|
|
||
| ```python | ||
| from pydantic import BaseModel | ||
|
|
||
| class Location(BaseModel): | ||
| city: str | ||
| state: str | ||
|
|
||
| marvin.extract("I moved from NY to CHI", target=Location) | ||
|
|
||
| # [ | ||
| # Location(city="New York", state="New York"), | ||
| # Location(city="Chcago", state="Illinois") | ||
| # ] | ||
| ``` | ||
|
|
||
| 🎬 You can define your own types for AI Functions to return, using things like: | ||
| - Pydantic [BaseModel](https://pydantic-docs.helpmanual.io/usage/models/) | ||
| - [TypedDict](https://docs.python.org/3/library/typing.html#typing.TypedDict) | ||
| - [`Literal`](https://docs.python.org/3/library/typing.html#typing.Literal) types | ||
| Almost all Marvin functions can be given `instructions` for more control. Here we extract only monetary values: | ||
|
|
||
| ```python | ||
| from typing_extensions import TypedDict | ||
| marvin.extract( | ||
| "I paid $10 for 3 tacos and got a dollar and 25 cents back.", | ||
| target=float, | ||
| instructions="Only extract money" | ||
| ) | ||
|
|
||
| class DetailedSentiment(TypedDict): | ||
| """A detailed sentiment analysis result. | ||
| # [10.0, 1.25] | ||
| ``` | ||
|
|
||
| - `sentiment_score` is a number between 1 (positive) and -1 (negative) | ||
| - `summary_in_a_word` is a one-word summary of the general sentiment, | ||
| use any apt word that captures the nuance of the sentiment | ||
| """ | ||
| sentiment_score: float | ||
| summary_in_a_word: str | ||
|
|
||
| @ai_fn | ||
| def detailed_sentiment(text: str) -> DetailedSentiment: | ||
| """What do you think the sentiment of `text` is? | ||
| Use your theory of mind to put yourself in the shoes of its author. | ||
| """ | ||
| Learn more about entity extraction [here](https://askmarvin.ai/docs/text/extraction). | ||
|
|
||
| ### ✨ Generate data | ||
|
|
||
| Marvin can `generate` synthetic data for you, following instructions and an optional schema: | ||
|
|
||
| detailed_sentiment("I'ma Mario, and I'ma gonna wiiiiin!") | ||
| # {'sentiment_score': 0.8, 'summary_in_a_word': 'energetic'} | ||
| ```python | ||
| from pydantic import BaseModel | ||
|
|
||
| class Location(BaseModel): | ||
| city: str | ||
| state: str | ||
|
|
||
| marvin.generate( | ||
| n=4, | ||
| target=Location, | ||
| instructions="cities in the United States named after presidents" | ||
| ) | ||
|
|
||
| # [ | ||
| # Location(city='Washington', state='District of Columbia'), | ||
| # Location(city='Jackson', state='Mississippi'), | ||
| # Location(city='Cleveland', state='Ohio'), | ||
| # Location(city='Lincoln', state='Nebraska'), | ||
| # ] | ||
| ``` | ||
|
|
||
| ### 🧩 AI Models | ||
| AI models are based on Pydantic's [BaseModel](https://pydantic-docs.helpmanual.io/usage/models/), but with a twist: they are instantiated with plain text, and will use an LLM to infer their values. | ||
| Learn more about data generation [here](https://askmarvin.ai/docs/text/generation). | ||
|
|
||
| ### 🪄 Standardize text by casting to types | ||
|
|
||
| You can learn more about AI models [here](https://www.askmarvin.ai/components/ai_model/). | ||
| Marvin can `cast` arbitrary text to any Python type: | ||
|
|
||
| ```python | ||
| from marvin import ai_model | ||
| from pydantic import BaseModel, Field | ||
| marvin.cast("one two three", list[int]) | ||
|
|
||
| # [1, 2, 3] | ||
| ``` | ||
|
|
||
| This is useful for standardizing text inputs or matching natural language to a schema: | ||
|
|
||
| ```python | ||
| from pydantic import BaseModel | ||
|
|
||
| @ai_model | ||
| class Location(BaseModel): | ||
| """A city in the United States""" | ||
| city: str | ||
| state: str = Field(..., description="The two-letter state abbreviation") | ||
| state: str | ||
|
|
||
| marvin.cast("The Big Apple", Location) | ||
|
|
||
| # Location(city="New York", state="New York") | ||
| ``` | ||
|
|
||
| For a class-based approach, Marvin's `@model` decorator can be applied to any Pydantic model to let it be instantiated from text: | ||
|
|
||
| ```python | ||
| @marvin.model | ||
| class Location(BaseModel): | ||
| city: str | ||
| state: str | ||
|
|
||
| Location("The Big Apple") | ||
| # Location(city='New York', state='NY') | ||
|
|
||
| # Location(city="New York", state="New York") | ||
| ``` | ||
|
|
||
| ## Assistants (Beta) | ||
| Based on OpenAI's Assistant API, Marvin's Assistants are the easiest way to build a stateful natural language interface equipped with familiar tools (i.e. python functions). | ||
| Learn more about casting to types [here](https://askmarvin.ai/docs/text/transformation). | ||
|
|
||
| ### 🦾 Build AI-powered functions | ||
|
|
||
| Marvin functions let you combine any inputs, instructions, and output types to create custom AI-powered behaviors... without source code. These functions can can go well beyond the capabilities of `extract` or `classify`, and are ideal for complex natural language processing or mapping combinations of inputs to outputs. | ||
|
|
||
| ```python | ||
| from marvin.beta.assistants import Assistant, Thread | ||
| @marvin.fn | ||
| def sentiment(text: str) -> float: | ||
| """ | ||
| Returns a sentiment score for `text` | ||
| between -1 (negative) and 1 (positive). | ||
| """ | ||
|
|
||
| def multiply(x: float, y: float) -> float: | ||
| return x * y | ||
| sentiment("I love working with Marvin!") # 0.8 | ||
| sentiment("These examples could use some work...") # -0.2 | ||
| ``` | ||
|
|
||
| Marvin functions look exactly like regular Python functions, except that you don't have to write any source code. When these functions are called, an AI interprets their description and inputs and generates the output. | ||
|
|
||
| Note that Marvin does NOT work by generating or executing source code, which would be unsafe for most use cases. Instead, it uses the LLM itself as a "runtime" to predict function outputs. That's actually the source of its power: Marvin functions can handle complex use cases that would be difficult or impossible to express as code. | ||
|
|
||
| def divide(x: float, y: float) -> float: | ||
| return x / y | ||
| You can learn more about functions [here](https://www.askmarvin.ai/docs/text/functions/). | ||
|
|
||
| ### 🖼️ Generate images from text | ||
|
|
||
| with Assistant(tools=[multiply, divide]) as assistant: | ||
| thread = Thread() | ||
| while True: | ||
| message = input("You: ") | ||
| if message.lower() in ["exit", ":q", "bye"]: | ||
| break | ||
| thread.add(message) | ||
| thread.run(assistant) | ||
| print("\n\n".join(m.content[0].text.value for m in thread.get_messages())) | ||
| # what is the speed of light (m/s) times the number of days in a year? | ||
| Marvin can `paint` images from text: | ||
|
|
||
| # what is that number divided by 42? | ||
| ```python | ||
| marvin.paint("a simple cup of coffee, still warm") | ||
| ``` | ||
|
|
||
| Read more about [our SDK](/src/marvin/beta/assistants/README.md) and/or the [OpenAI docs](https://platform.openai.com/docs/assistants/overview). | ||
| <p align="center"> | ||
| <img src="docs/assets/images/docs/images/coffee.png" style="width: 50%; height: auto;"/> | ||
| </p> | ||
|
|
||
| Learn more about image generation [here](https://askmarvin.ai/docs/images/generation). | ||
|
|
||
| ### 🔍 Classify images (beta) | ||
|
|
||
| In addition to text, Marvin has beta support for captioning, classifying, transforming, and extracting entities from images using the GPT-4 vision model: | ||
|
|
||
| ```python | ||
| marvin.beta.classify( | ||
| marvin.Image("docs/images/coffee.png"), | ||
| labels=["drink", "food"], | ||
| ) | ||
|
|
||
| # "drink" | ||
| ``` | ||
|
|
||
| ## Get in touch! | ||
|
|
||
| 💡 **Feature idea?** share it in the `#development` channel in [our Discord](https://discord.com/invite/Kgw4HpcuYG). | ||
|
|
||
| ## Reach out! | ||
| 💡 **Have an idea for a feature?** toss it in `#development` in [our Discord](https://discord.com/invite/Kgw4HpcuYG) | ||
| 🐛 **Found a bug?** feel free to [open an issue](https://github.com/PrefectHQ/marvin/issues/new/choose). | ||
|
|
||
| 🐛 **found a bug?** feel free to [open an issue](https://github.com/PrefectHQ/marvin/issues/new/choose) | ||
| 👷 **Feedback?** Marvin is under active development, and we'd love to [hear it](https://github.com/PrefectHQ/marvin/discussions). |
Oops, something went wrong.