Metadata-Version: 2.1
Name: turbo-chat
Version: 0.1.13
Summary: Idiomatic way to build chatgpt apps using async generators in python
License: MIT
Author: Diwank Singh Tomer
Author-email: singh@diwank.name
Requires-Python: >=3.8,<3.11
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Requires-Dist: aiohttp (>=3.8.4,<4.0.0)
Requires-Dist: openai (>=0.27.0,<0.28.0)
Requires-Dist: pydantic (>=1.10.5,<2.0.0)
Requires-Dist: tenacity (>=8.2.2,<9.0.0)
Description-Content-Type: text/markdown

# turbo-chat

> Idiomatic way to build chatgpt apps using async generators in python

The [ChatGPT API](https://openai.com/blog/introducing-chatgpt-and-whisper-apis) uses a new input format called [chatml](https://github.com/openai/openai-python/blob/main/chatml.md). In openai's [python client](https://github.com/openai/openai-python/blob/main/chatml.md), the format is used something like this:

```python
messages = [
    {"role": "system", "content": "Greet the user!"},
    {"role": "user", "content": "Hello world!"},
]
```

The idea here is to incrementally build the messages using an async generator and then use that to generate completions. [Async generators](https://superfastpython.com/asynchronous-generators-in-python/) are incredibly versatile and simple abstraction for doing this kind of stuff. They can also be composed together very easily.

```python
# Equivalent turbo-chat generator
async def example():
    yield System(content="Greet the user!")
    yield User(content="Hello World!")

    # To run generation, just yield Generate(),
    # the lib will take care of correctly running the app, and
    # return the value back here.
    output = yield Generate()
    print(output.content)
```

See more detailed example below.

## Installation

```bash
pip install turbo-chat
```

## Example

```python
from typing import AsyncGenerator, Union

from turbo_chat import (
    turbo,
    System,
    User,
    Assistant,
    GetUserInput,
    Generate,
    run,
)

# Get user
async def get_user(id):
    return {"zodiac": "pisces"}

# Set user zodiac mixin
# Notice that no `@turbo()` decorator used here
async def set_user_zodiac(context: dict):

    user_id: int = context["user_id"]
    user_data: dict = await get_user(user_id)
    zodiac: str = user_data["zodiac"]

    yield User(content=f"My zodiac sign is {zodiac}")


# Horoscope app
@turbo(temperature=0.0)
async def horoscope(context: dict):

    yield System(content="You are a fortune teller")

    # Yield from mixin
    async for output in set_user_zodiac(context):
        yield output

    # Prompt runner to ask for user input
    input = yield GetUserInput(message="What do you want to know?")

    # Yield the input
    yield User(content=input)

    # Generate (overriding the temperature)
    value = yield Generate(settings={"temperature": 0.9})


# Let's run this
app: AsyncGenerator[Union[Assistant, GetUserInput], str] = horoscope({"user_id": 1})

_input = None
while response := await run(app, _input):
    result, done = response

    if isinstance(result, GetUserInput):
        _input = input(result.message)
        continue

    if isinstance(result, Assistant):
        print(result.content)

    if done:
        break

# Output
# >>> What do you want to know? Tell me my fortune
# >>> As an AI language model, I cannot predict the future or provide supernatural fortune-telling. However, I can offer guidance and advice based on your current situation and past experiences. Is there anything specific you would like me to help you with?
#

```

### Custom memory

You can also customize how the messages are persisted in-between the executions.

```python
from turbo_chat import turbo, BaseMemory

class RedisMemory(BaseMemory):
    """Implement BaseMemory methods here"""

    async def init(self, context) -> None:
        ...

    async def append(self, item) -> None:
        ...

    async def clear(self) -> None:
        ...


# Now use the memory in a turbo_chat app
@turbo(memory_class=RedisMemory)
async def app(context):
    ...
```

### Get access to memory object directly (just declare an additional param)

```python
@turbo()
async def app(context: dict, memory: BaseMemory):

    messages = await memory.get()
    ...
```

### Generate a response to use internally but don't yield downstream

```python
@turbo()
async def example(context):
    yield System(content="You are a good guy named John")
    yield User(content="What is your name?")
    result = yield Generate(yield_downstream=False)

    yield User(content="How are you doing?")
    result = yield Generate()

b = example()
results = [output async for output in b]

assert len(results) == 1
```

### Add a simple in-memory cache

You can also subclass the `BaseCache` class to create a custom cache.

```python
cache = SimpleCache()

@turbo(cache=cache)
async def example(context={}):
    yield System(content="You are a good guy named John")
    yield User(content="What is your name?")
    result = yield Generate()

b = example()
results = [output async for output in b]

assert len(cache.cache) == 1

```

---

![turbo](https://user-images.githubusercontent.com/931887/222912628-8662fad0-091f-4cb8-92f3-6cce287716e9.jpg)

