mirror of
https://github.com/kennethreitz/instructor.git
synced 2026-06-05 22:50:18 +00:00
Jason Liu
268 lines
9.8 KiB
Markdown
268 lines
9.8 KiB
Markdown
# Instructor (openai_function_call)
|
|
|
|
!!! note "Renaming from openai_function_call"
|
|
This library used to be called `openai_function_call` simply change the import and you should be good to go!
|
|
|
|
```sh
|
|
find /path/to/dir -type f -exec sed -i 's/openai_function_call/instructor/g' {} \;
|
|
```
|
|
|
|
*Structured extraction in Python, powered by OpenAI's function calling api, designed for simplicity, transparency, and control.*
|
|
|
|
-----
|
|
|
|
This library is built to interact with openai's function call api from python code, with python structs / objects. It's designed to be intuitive, easy to use, but give great visibily in how we call openai.
|
|
|
|
The approach of combining a human prompt and a "response schema" is not necessarily unique; however, it shows great promise. As we have been concentrating on translating user intent into structured data, we have discovered that Python with Pydantic is exceptionally well-suited for this task.
|
|
|
|
**OpenAISchema** is based on Python type annotations, and powered by Pydantic.
|
|
|
|
The key features are:
|
|
|
|
* **Intuitive to write**: Great support for editors, completions. Spend less time debugging.
|
|
* **Writing prompts as code**: Collocate docstrings and descriptions as part of your prompting.
|
|
* **Extensible**: Bring your own kitchen sink without being weighted down by abstractions.
|
|
|
|
## Structured Extraction with `openai`
|
|
|
|
Welcome to the Quick Start Guide for OpenAI Function Call. This guide will walk you through the installation process and provide examples demonstrating the usage of function calls and schemas with OpenAI and Pydantic.
|
|
|
|
### Requirements
|
|
|
|
This library depends on **Pydantic** and **OpenAI** that's all.
|
|
|
|
### Installation
|
|
|
|
To get started with OpenAI Function Call, you need to install it using `pip`. Run the following command in your terminal:
|
|
|
|
!!! note Requirement
|
|
Ensure you have Python version 3.9 or above.
|
|
|
|
```sh
|
|
$ pip install instructor
|
|
```
|
|
|
|
## Quick Start with Patching ChatCompletion
|
|
|
|
To simplify your work with OpenAI models and streamline the extraction of Pydantic objects from prompts, we offer a patching mechanism for the `ChatCompletion`` class. Here's a step-by-step guide:
|
|
|
|
### Step 1: Import and Patch the Module
|
|
|
|
First, import the required libraries and apply the patch function to the OpenAI module. This exposes new functionality with the response_model parameter.
|
|
|
|
```python
|
|
import openai
|
|
import instructor
|
|
from pydantic import BaseModel
|
|
|
|
# This enables response_model keyword
|
|
# from openai.ChatCompletion.create
|
|
instructor.patch()
|
|
```
|
|
|
|
### Step 2: Define the Pydantic Model
|
|
|
|
Create a Pydantic model to define the structure of the data you want to extract. This model will map directly to the information in the prompt.
|
|
|
|
```python
|
|
class UserDetail(BaseModel):
|
|
name: str
|
|
age: int
|
|
```
|
|
|
|
### Step 3: Extract Data with ChatCompletion
|
|
|
|
Use the openai.ChatCompletion.create method to send a prompt and extract the data into the Pydantic object. The response_model parameter specifies the Pydantic model to use for extraction.
|
|
|
|
```python
|
|
user: UserDetail = openai.ChatCompletion.create(
|
|
model="gpt-3.5-turbo",
|
|
response_model=UserDetail,
|
|
messages=[
|
|
{"role": "user", "content": "Extract Jason is 25 years old"},
|
|
]
|
|
)
|
|
```
|
|
|
|
### Step 4: Validate the Extracted Data
|
|
|
|
You can then validate the extracted data by asserting the expected values. By adding the type things you also get a bunch of nice benefits with your IDE like spell check and auto complete!
|
|
|
|
```python
|
|
assert user.name == "Jason"
|
|
assert user.age == 25
|
|
```
|
|
|
|
## Introduction to `OpenAISchema`
|
|
|
|
If you want more control than just passing a single class we can use the `OpenAISchema` which extends `BaseModel`.
|
|
|
|
This quick start guide contains the follow sections:
|
|
|
|
1. Defining a schema
|
|
2. Adding Additional Prompting
|
|
3. Calling the ChatCompletion
|
|
4. Deserializing back to the instance
|
|
|
|
OpenAI Function Call allows you to leverage OpenAI's powerful language models for function calls and schema extraction. This guide provides a quick start for using OpenAI Function Call.
|
|
|
|
### Section 1: Defining a Schema
|
|
|
|
To begin, let's define a schema using OpenAI Function Call. A schema describes the structure of the input and output data for a function. In this example, we'll define a simple schema for a `User` object:
|
|
|
|
```python
|
|
from instructor import OpenAISchema
|
|
|
|
class UserDetails(OpenAISchema):
|
|
name: str
|
|
age: int
|
|
```
|
|
|
|
In this schema, we define a `UserDetails` class that extends `OpenAISchema`. We declare two fields, `name` and `age`, of type `str` and `int` respectively.
|
|
|
|
### Section 2: Adding Additional Prompting
|
|
|
|
To enhance the performance of the OpenAI language model, you can add additional prompting in the form of docstrings and field descriptions. They can provide context and guide the model on how to process the data.
|
|
|
|
!!! note Using `patch`
|
|
these docstrings and fields descriptions are powered by `pydantic.BaseModel` so they'll work via the patching approach as well.
|
|
|
|
```python hl_lines="5 6"
|
|
from instructor import OpenAISchema
|
|
from pydantic import Field
|
|
|
|
class UserDetails(OpenAISchema):
|
|
""""
|
|
Correctly extracted user information
|
|
:param age: age of the user
|
|
"""
|
|
name: str = Field(..., description="User's full name")
|
|
age: int
|
|
```
|
|
|
|
In this updated schema, we use the `Field` class from `pydantic` to add descriptions to the `name` field. Moreover, we use the docstring to add information for the parameter `age`.
|
|
In both cases, the description provides information about the fields, giving even more context to the language model.
|
|
Information from the docstring is extracted using [docstring-parser](https://github.com/rr-/docstring_parser) which supports different docstring styles.
|
|
Note that if the `Field` contains a description for a parameter as well as the docstring, the `Field`'s description is used.
|
|
|
|
!!! note "Code, schema, and prompt"
|
|
We can run `openai_schema` to see exactly what the API will see, notice how the docstrings, attributes, types, and parameter descriptions are now part of the schema. This describes on this library's core philosophies.
|
|
|
|
```python
|
|
class UserDetails(OpenAISchema):
|
|
"""
|
|
Correctly extracted user information
|
|
|
|
:param name: the user's full name
|
|
:param age: age of the user
|
|
"""
|
|
name: str = Field(..., description="User's full name")
|
|
age: int
|
|
|
|
UserDetails.openai_schema
|
|
```
|
|
|
|
```json hl_lines="3 9 13"
|
|
{
|
|
"name": "UserDetails",
|
|
"description": "Correctly extracted user information",
|
|
"parameters": {
|
|
"type": "object",
|
|
"properties": {
|
|
"name": {
|
|
"type": "string",
|
|
"description": "User's full name"
|
|
},
|
|
"age": {
|
|
"type": "integer"
|
|
"description": "age of the user"
|
|
}
|
|
},
|
|
"required": [
|
|
"age",
|
|
"name"
|
|
]
|
|
}
|
|
}
|
|
```
|
|
|
|
### Section 3: Calling the ChatCompletion
|
|
|
|
With the schema defined, let's proceed with calling the `ChatCompletion` API using the defined schema and messages.
|
|
|
|
```python hl_lines="11 12"
|
|
from instructor import OpenAISchema
|
|
from pydantic import Field
|
|
|
|
class UserDetails(OpenAISchema):
|
|
name: str = Field(..., description="User's full name")
|
|
age: int
|
|
|
|
|
|
completion = openai.ChatCompletion.create(
|
|
model="gpt-3.5-turbo-0613",
|
|
functions=[UserDetails.openai_schema],
|
|
function_call={"name": UserDetails.openai_schema["name"]},
|
|
messages=[
|
|
{"role": "system", "content": "Extract user details from my requests"},
|
|
{"role": "user", "content": "My name is John Doe and I'm 30 years old."},
|
|
],
|
|
)
|
|
```
|
|
|
|
In this example, we make a call to the `ChatCompletion` API by providing the model name (`gpt-3.5-turbo-0613`) and a list of messages. The messages consist of a system message and a user message. The system message sets the context by requesting user details, while the user message provides the input with the user's name and age.
|
|
|
|
Note that we have omitted the additional parameters that can be included in the API request, such as `temperature`, `max_tokens`, and `n`. These parameters can be customized according to your requirements.
|
|
|
|
### Section 4: Deserializing Back to the Instance
|
|
|
|
To deserialize the response from the `ChatCompletion` API back into an instance of the `UserDetails` class, we can use the `from_response` method.
|
|
|
|
```python
|
|
user = UserDetails.from_response(completion)
|
|
|
|
print(user.name) # Output: John Doe
|
|
print(user.age) # Output: 30
|
|
```
|
|
|
|
By calling `UserDetails.from_response`, we create an instance of the `UserDetails` class using the response from the API call. Subsequently, we can access the extracted user details through the `name` and `age` attributes of the `user` object.
|
|
|
|
## IDE Support
|
|
|
|
Everything is designed for you to get the best developer experience possible, with the best editor support.
|
|
|
|
Including **autocompletion**:
|
|
|
|
|
|
|
|
And even **inline errors**
|
|
|
|
|
|
|
|
## OpenAI Schema and Pydantic
|
|
|
|
This quick start guide provided you with a basic understanding of how to use OpenAI Function Call for schema extraction and function calls. You can now explore more advanced use cases and creative applications of this library.
|
|
|
|
Since `UserDetails` is a `OpenAISchems` and a `pydantic.BaseModel` you can use inheritance and nesting to create more complex emails while avoiding code duplication
|
|
|
|
```python
|
|
class UserDetails(OpenAISchema):
|
|
name: str = Field(..., description="User's full name")
|
|
age: int
|
|
|
|
class UserWithAddress(UserDetails):
|
|
address: str
|
|
|
|
class UserWithFriends(UserDetails):
|
|
best_friend: UserDetails
|
|
friends: List[UserDetails]
|
|
```
|
|
|
|
If you have any questions, feel free to leave an issue or reach out to the library's author on [Twitter](https://twitter.com/jxnlco). For a more comprehensive solution with additional features, consider checking out [MarvinAI](https://www.askmarvin.ai/).
|
|
|
|
To see more examples of how we can create interesting models check out some [examples.](examples/index.md)
|
|
|
|
## License
|
|
|
|
This project is licensed under ther terms of the MIT License.
|