# OpenAI Decorator
The OpenAI Decorator is a Python package that enriches your Python functions with the capabilities of OpenAI's API. It leverages the function's signature and docstring to generate specifications comprehensible by OpenAI's API, subsequently making a request to the API.
## Table of Contents
- [Key Features](#key-features)
- [Installation](#installation)
- [Usage Guide](#usage-guide)
- [Basic Usage](#basic-usage)
- [Type Annotations & Optionals](#type-annotations-optionals)
- [Passing a Callable Prompt](#passing-callable-prompts)
- [Using the Decorator as a Regular Function](#using-the-decorator-as-a-regular-function)
- [Contributing](#contributing)
- [Setup](#setup)
- [Future Enhancements](#future-enhancements)
- [License](#license)
## Key Features
- Automatic generation of specs from the function's signature and docstring.
- Integration with OpenAI's API using the generated specs and a customizable prompt.
- Inference of function arguments from the OpenAI API response, negating the need for manual input.
- Handling of various edge cases related to function signatures, including default and optional values.
## Installation
Install the package using pip:
```sh
pip install openai-decorator
```
Save your OpenAI key in an environment variable named `OPENAI_API_KEY`:
```sh
export OPENAI_API_KEY=<your key>
```
## Usage Guide
### Basic Usage
Apply the `openai_func` decorator to a function with type annotations and a well-formatted Google Style docstring. Then, it will seamlessly interact with OpenAI's API to generate the parameters for your function based on the prompt.
```python
from openai_decorator import openai_func
@openai_func(prompt="Your OpenAI prompt here")
def example_function(arg1: str, arg2: int) -> str:
"""
This is an example function.
Args:
arg1 (str): Description for arg1
arg2 (int): Description for arg2
Returns:
str: Description for return value
"""
return f"Your output here: {arg1} and {arg2}"
result = example_function()
print(result)
```
### Type Annotations & Optionals
The decorator efficiently handles complex function signatures:
- **Default Values:** Parameters with a default value or `Optional` type annotation are marked as optional in the JSON specification.
- **Type Annotations:** Supports primitive types (int, str, bool, float) and complex types (List, Dict, Tuple, Set, Optional, custom classes).
```python
from openai_decorator import openai_func
@openai_func(prompt="Perform mathematical operations")
def example_function(a: int, b: int, c: float = 0.0, d: List[int] = []):
"""
This is an example function.
Args:
a: The first integer.
b: The second integer.
c: An optional float.
d: An optional list of integers.
"""
# Function body here
```
### Passing Callable Prompts
In addition to static strings, you can also pass a callable to the `prompt` parameter in the `openai_func` decorator to dynamically generate prompts. This callable should take no arguments and return a string. Here's an example:
```python
from openai_decorator import openai_func
def generate_prompt():
# Dynamically generate a prompt.
# In a real-world scenario, this could depend on various factors,
# such as the current time or the latest trending topics.
return "Your OpenAI prompt here"
@openai_func(prompt=generate_prompt)
def example_function(arg1: str, arg2: int) -> str:
"""
This is an example function.
Args:
arg1 (str): Description for arg1
arg2 (int): Description for arg2
Returns:
str: Description for return value
"""
return f"Your output here: {arg1} and {arg2}"
result = example_function()
print(result)
```
In this example, every time `example_function` is called, `generate_prompt
### Using the Decorator as a Regular Function
`openai_func` can also be used as a regular function. First, call `openai_func` with the prompt and any other arguments to get the actual decorator. Then, apply this decorator to your function.
```python
from openai_decorator import openai_func
def example_function(arg1: str, arg2: int) -> str:
"""
This is an example function.
Args:
arg1 (str): Description for arg1
arg2 (int): Description for arg2
Returns:
str: Description for return value
"""
return f"Your output here: {arg1} and {arg2}"
# Use openai_func as a regular function
prompt = "Your OpenAI prompt here"
example_func_with_generated_params = openai_func(prompt=prompt)(example_function)
result = example_func_with_generated_params()
print(result)
```
This approach gives you more flexibility in determining the prompt at runtime.
## Contributing
Contributions are welcome! Feel free to submit issues and pull requests.
### Setup
1. Install Python `3.11` and the latest version of [poetry](https://python-poetry.org/docs/#installing-with-pipx)
- `pyenv` can help manage multiple Python versions.
2. Clone the repository: `git clone`.
3. Set your OpenAI key as an environment variable:
```shell
export OPENAI_API_KEY=<insert your openai key>
```
4. Install dependencies: `poetry install --no-root`
## Future Enhancements
- [x] Handle optional parameters.
- [x] Handle parameters with default values.
- [ ] Add test to ensure default values are used if OpenAI doesn't return parameters.
- [ ] Publish the package to PyPI.
- [ ] Expand test coverage.
- [ ] Show tests passing & coverage as github badges.
- [X] Fix CI/CD -- Address the issue where the pre-release deployment to PyPI fails due to the need for version update.
- [ ] Add docstrings.
- [ ] Generate docs (from docstrings).
## License
This project is under the MIT License. See the [LICENSE](LICENSE) file for more details.
Raw data
{
"_id": null,
"home_page": "https://github.com/shruti222patel/openai-decorator",
"name": "openai-decorator",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.8,<4.0",
"maintainer_email": "",
"keywords": "openai,decorator,functions,openai-functions,openai-decorator,openai-api,openai-api-functions,openai-api-decorator,openai",
"author": "Shruti Patel",
"author_email": "shruti222patel@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/f1/da/9aa7db7476662661c4854ec047ae41b75a3bf723e95377b82bf2a4ecee9d/openai_decorator-0.1.3.tar.gz",
"platform": null,
"description": "# OpenAI Decorator\n\nThe OpenAI Decorator is a Python package that enriches your Python functions with the capabilities of OpenAI's API. It leverages the function's signature and docstring to generate specifications comprehensible by OpenAI's API, subsequently making a request to the API.\n\n## Table of Contents\n\n- [Key Features](#key-features)\n- [Installation](#installation)\n- [Usage Guide](#usage-guide)\n - [Basic Usage](#basic-usage)\n - [Type Annotations & Optionals](#type-annotations-optionals)\n - [Passing a Callable Prompt](#passing-callable-prompts)\n - [Using the Decorator as a Regular Function](#using-the-decorator-as-a-regular-function)\n- [Contributing](#contributing)\n - [Setup](#setup)\n- [Future Enhancements](#future-enhancements)\n- [License](#license)\n\n## Key Features\n\n- Automatic generation of specs from the function's signature and docstring.\n- Integration with OpenAI's API using the generated specs and a customizable prompt.\n- Inference of function arguments from the OpenAI API response, negating the need for manual input.\n- Handling of various edge cases related to function signatures, including default and optional values.\n\n## Installation\n\nInstall the package using pip:\n\n```sh\npip install openai-decorator\n```\nSave your OpenAI key in an environment variable named `OPENAI_API_KEY`:\n```sh\nexport OPENAI_API_KEY=<your key>\n```\n\n## Usage Guide\n\n### Basic Usage\nApply the `openai_func` decorator to a function with type annotations and a well-formatted Google Style docstring. Then, it will seamlessly interact with OpenAI's API to generate the parameters for your function based on the prompt.\n\n```python\nfrom openai_decorator import openai_func\n\n@openai_func(prompt=\"Your OpenAI prompt here\")\ndef example_function(arg1: str, arg2: int) -> str:\n \"\"\"\n This is an example function.\n\n Args:\n arg1 (str): Description for arg1\n arg2 (int): Description for arg2\n\n Returns:\n str: Description for return value\n \"\"\"\n return f\"Your output here: {arg1} and {arg2}\"\n\nresult = example_function()\nprint(result)\n```\n\n### Type Annotations & Optionals\n\nThe decorator efficiently handles complex function signatures:\n\n- **Default Values:** Parameters with a default value or `Optional` type annotation are marked as optional in the JSON specification.\n- **Type Annotations:** Supports primitive types (int, str, bool, float) and complex types (List, Dict, Tuple, Set, Optional, custom classes).\n\n```python\nfrom openai_decorator import openai_func\n\n@openai_func(prompt=\"Perform mathematical operations\")\ndef example_function(a: int, b: int, c: float = 0.0, d: List[int] = []):\n \"\"\"\n This is an example function.\n\n Args:\n a: The first integer.\n b: The second integer.\n c: An optional float.\n d: An optional list of integers.\n \"\"\"\n # Function body here\n```\n\n### Passing Callable Prompts\n\nIn addition to static strings, you can also pass a callable to the `prompt` parameter in the `openai_func` decorator to dynamically generate prompts. This callable should take no arguments and return a string. Here's an example:\n\n```python\nfrom openai_decorator import openai_func\n\ndef generate_prompt():\n # Dynamically generate a prompt.\n # In a real-world scenario, this could depend on various factors,\n # such as the current time or the latest trending topics.\n return \"Your OpenAI prompt here\"\n\n@openai_func(prompt=generate_prompt)\ndef example_function(arg1: str, arg2: int) -> str:\n \"\"\"\n This is an example function.\n\n Args:\n arg1 (str): Description for arg1\n arg2 (int): Description for arg2\n\n Returns:\n str: Description for return value\n \"\"\"\n return f\"Your output here: {arg1} and {arg2}\"\n\nresult = example_function()\nprint(result)\n```\n\nIn this example, every time `example_function` is called, `generate_prompt\n\n### Using the Decorator as a Regular Function\n\n`openai_func` can also be used as a regular function. First, call `openai_func` with the prompt and any other arguments to get the actual decorator. Then, apply this decorator to your function.\n\n```python\nfrom openai_decorator import openai_func\n\ndef example_function(arg1: str, arg2: int) -> str:\n \"\"\"\n This is an example function.\n\n Args:\n arg1 (str): Description for arg1\n arg2 (int): Description for arg2\n\n Returns:\n str: Description for return value\n \"\"\"\n return f\"Your output here: {arg1} and {arg2}\"\n\n# Use openai_func as a regular function\nprompt = \"Your OpenAI prompt here\"\n\nexample_func_with_generated_params = openai_func(prompt=prompt)(example_function)\n\nresult = example_func_with_generated_params()\nprint(result)\n```\nThis approach gives you more flexibility in determining the prompt at runtime.\n\n## Contributing\n\nContributions are welcome! Feel free to submit issues and pull requests.\n\n### Setup\n\n1. Install Python `3.11` and the latest version of [poetry](https://python-poetry.org/docs/#installing-with-pipx)\n - `pyenv` can help manage multiple Python versions.\n2. Clone the repository: `git clone`.\n3. Set your OpenAI key as an environment variable:\n```shell\nexport OPENAI_API_KEY=<insert your openai key>\n```\n4. Install dependencies: `poetry install --no-root`\n\n## Future Enhancements\n\n- [x] Handle optional parameters.\n- [x] Handle parameters with default values.\n- [ ] Add test to ensure default values are used if OpenAI doesn't return parameters.\n- [ ] Publish the package to PyPI.\n- [ ] Expand test coverage.\n- [ ] Show tests passing & coverage as github badges.\n- [X] Fix CI/CD -- Address the issue where the pre-release deployment to PyPI fails due to the need for version update.\n- [ ] Add docstrings.\n- [ ] Generate docs (from docstrings).\n\n## License\n\nThis project is under the MIT License. See the [LICENSE](LICENSE) file for more details.\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "This package allows you to genrate function paramater values by integrating OpenAI's API function capabilities through signature and docstring analysis.",
"version": "0.1.3",
"project_urls": {
"Homepage": "https://github.com/shruti222patel/openai-decorator",
"Repository": "https://github.com/shruti222patel/openai-decorator"
},
"split_keywords": [
"openai",
"decorator",
"functions",
"openai-functions",
"openai-decorator",
"openai-api",
"openai-api-functions",
"openai-api-decorator",
"openai"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "c46a4279fd650e3f6a2ba54f528f4af847381098b1dd74911888ea096044ab91",
"md5": "7768cb2794235cc05cf1bb9b619a8e99",
"sha256": "c2b2305c15feaf89870b57a6cb3640f85c209714dbdf27adff613094297db9b3"
},
"downloads": -1,
"filename": "openai_decorator-0.1.3-py3-none-any.whl",
"has_sig": false,
"md5_digest": "7768cb2794235cc05cf1bb9b619a8e99",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.8,<4.0",
"size": 7009,
"upload_time": "2023-06-18T21:03:47",
"upload_time_iso_8601": "2023-06-18T21:03:47.152029Z",
"url": "https://files.pythonhosted.org/packages/c4/6a/4279fd650e3f6a2ba54f528f4af847381098b1dd74911888ea096044ab91/openai_decorator-0.1.3-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "f1da9aa7db7476662661c4854ec047ae41b75a3bf723e95377b82bf2a4ecee9d",
"md5": "50d906a61d81f2217117541c4d8c9c15",
"sha256": "10b72404340c7cdbe7e449a4ab5c14c5839df7cf7cee661bf11d0393a5130b74"
},
"downloads": -1,
"filename": "openai_decorator-0.1.3.tar.gz",
"has_sig": false,
"md5_digest": "50d906a61d81f2217117541c4d8c9c15",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.8,<4.0",
"size": 6271,
"upload_time": "2023-06-18T21:03:48",
"upload_time_iso_8601": "2023-06-18T21:03:48.922645Z",
"url": "https://files.pythonhosted.org/packages/f1/da/9aa7db7476662661c4854ec047ae41b75a3bf723e95377b82bf2a4ecee9d/openai_decorator-0.1.3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-06-18 21:03:48",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "shruti222patel",
"github_project": "openai-decorator",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "openai-decorator"
}
JSON
