A trivial set of API bindings for AI models, because I'd like them to be easy to use.
Project description
TrivialAI
(A set of requests-based, trivial bindings for AI models)
Basics
$ pip install pytrivialai
$ python
>>> from trivialai import claude, gcp, ollama, chatgpt
>>>
Basic model usage
Ollama
>>> client = ollama.Ollama("gemma2:2b", "http://localhost:11434/")
# or ollama.Ollama("deepseek-coder-v2:latest", "http://localhost:11434/")
# or ollama.Ollama("mannix/llama3.1-8b-abliterated:latest", "http://localhost:11434/")
>>> client.generate("This is a test message. Use the word 'platypus' in your response.", "Hello there! :D").content
'Hey! Did you know platypuses lay eggs and have webbed feet? Pretty cool, huh? 😁'
>>> client.generate_json("This is a test message. Use the word 'platypus' in your response.", "Generate a list of animal names in JSON format. Return [Name] and no other commentary").content
[{'name': 'Platypus'}, {'name': 'Eagle'}, {'name': 'Elephant'}, {'name': 'Giraffe'}, {'name': 'Lion'}, {'name': 'Tiger'}, {'name': 'Zebra'}, {'name': 'Dolphin'}, {'name': 'Whale'}, {'name': 'Bear'}, {'name': 'Wolf'}, {'name': 'Dog'}, {'name': 'Cat'}, {'name': 'Monkey'}]
>>>
Claude
>>> client = claude.Claude("claude-3-5-sonnet-20240620", os.environ["ANTHROPIC_API_KEY"])
>>> client.generate("This is a test message. Use the word 'platypus' in your response.", "Hello there! :D").content
"Hello! It's nice to meet you. I hope you're having a fantastic day. Since you mentioned using a specific word, I'll incorporate it here: Did you know that the platypus is one of the few mammals that can produce venom? It's quite an unusual and fascinating creature!"
>>>
GCP
>>> client = gcp.GCP("gemini-1.5-flash-001", "/path/to/your/gcp_creds.json", "us-central1")
>>> client.generate("This is a test message. Use the word 'platypus' in your response.", "Hello there! :D").content
"Hello! :D It's great to hear from you. Did you know platypuses are one of the few mammals that lay eggs? 🥚 They are truly fascinating creatures! What can I help you with today? 😊"
>>>
ChatGPT
>>> client = chatgpt.ChatGPT("gpt-3.5-turbo", os.environ["OPENAI_API_KEY"])
>>> client.generate("This is a test message. Use the word 'platypus' in your response.", "Hello there! :D").content
'Hello! How are you today? By the way, did you know the platypus is one of the few mammals that lays eggs?'
>>>
Basic Tool Use
>>> from src.trivialai import tools
>>> client = ollama.Ollama("deepseek-v2:16b", "http://localhost:11434/")
>>> tls = tools.Tools()
>>> from typing import Optional, List
>>> def _screenshot(url: str, selectors: Optional[List[str]] = None) -> None:
"Takes a url and an optional list of selectors. Takes a screenshot"
print(f"GOT {url}, {selectors}!")
... ... ...
>>> tls.define(_screenshot)
True
## You could also equivalently use
## >>> @tls.define()
## >>> def _screenshot(url: str, selectors: Optional[List[str]] = None) -> None:
## "Takes a url and an optional list of selectors. Takes a screenshot"
## print(f"GOT {url}, {selectors}!")
>>> tls.list()
[{'name': '_screenshot', 'type': {'url': <class 'str'>, 'selectors': typing.Optional[typing.List[str]]}, 'description': 'Takes a url and an optional list of selectors. Takes a screenshot'}]
>>> res = client.generate_tool_call(tls, "This space intentionally left blank.", "Take a screenshot of the Google website and highlight the search box")
>>> res.content
{'functionName': '_screenshot', 'args': {'url': 'https://www.google.com', 'selectors': ['#search']}}
>>> tls.call(res.content)
GOT https://www.google.com, ['#search']!
>>>
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
trivialai-0.1.15.tar.gz
(12.3 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file trivialai-0.1.15.tar.gz.
File metadata
- Download URL: trivialai-0.1.15.tar.gz
- Upload date:
- Size: 12.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.9.22
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
9157c27c3ae2f055bde13b74f67779cb2966de94991ffc18a527f4109421b1ce
|
|
| MD5 |
8e45a509e11a695d83c4e019243e9b9f
|
|
| BLAKE2b-256 |
d0a0a9db0fc144e448a32bb29feb461fb576736a41949041b78332dc693a8f86
|
File details
Details for the file trivialai-0.1.15-py3-none-any.whl.
File metadata
- Download URL: trivialai-0.1.15-py3-none-any.whl
- Upload date:
- Size: 11.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.9.22
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
93f2a2c03135db40fa32afd521937172c3bdcf39c4ed131ce19e4746e2677d14
|
|
| MD5 |
73ea6c5de0338b1d700cfea342aa09e6
|
|
| BLAKE2b-256 |
b846bcd2dbc554d1a3e83d0900595bb697c0f8ca0d151311ca58bf6b11c81d65
|