# magic-lamp
[](https://pypi.org/project/magic-lamp/)
Create magic LLM-powered Python functions that return anything you ask for. Many caveats.
## Quickstart
```bash
pip install magic-lamp
```
Define a function with a description and a set of examples.
```python
import magic_lamp
get_atoms = magic_lamp.Function(
"Break this molecule down into it's constituent atoms. Return as a set.",
examples=[
("water", {"hydrogen", "oxygen"}),
("glucose", {"carbon", "hydrogen", "oxygen"}),
],
)
print(get_atoms("ammonia")) # => {"nitrogen", "hydrogen"}
```
Functions can return any Python literal (strings, numbers, dicts, tuples lists, etc). No API keys are required, since by default `magic-lamp` downloads and runs a local LLM.
## Configuring the LLM
By default, `magic-lamp` downloads and runs a local LLM from Hugging Face. For more complex tasks, OpenAI models will perform better.
### Using OpenAI
`OPENAI_API_KEY` must be set in the environment. Pass in the name of a `gpt-*` model to the function constructor.
```python
import magic_lamp
format_number = magic_lamp.Function(
'Write this number out in words.',
examples=[
(1, "one"),
(35, "thirty-five"),
(15690, "fifteen thousand, six hundred ninety"),
],
model="gpt-4o-mini"
)
print(format_number(328745226793))
```
## Links
- https://github.com/jackmpcollins/magentic - A similar concept but using decorators.
- https://github.com/PrefectHQ/marvin - Antoher similar concept.
- https://github.com/abetlen/llama-cpp-python - Used by this library to run the local LLM.
- https://ai.meta.com/blog/meta-llama-3-1/ - Llama 3.1 8b is the default model.
- https://huggingface.co/bullerwins/Meta-Llama-3.1-8B-Instruct-GGUF - Uses these GGUFs by default.
Raw data
{
"_id": null,
"home_page": "https://github.com/rameshvarun/magic-lamp",
"name": "magic-lamp",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.9",
"maintainer_email": null,
"keywords": null,
"author": "Varun Ramesh",
"author_email": "varunramesh@live.com",
"download_url": "https://files.pythonhosted.org/packages/dd/b2/08345d4fddbb974bf8b111fdc2dd80a753e8979de1bd7c5607596ca0f77a/magic_lamp-0.4.1.tar.gz",
"platform": null,
"description": "# magic-lamp\n[](https://pypi.org/project/magic-lamp/)\n\nCreate magic LLM-powered Python functions that return anything you ask for. Many caveats.\n\n## Quickstart\n\n```bash\npip install magic-lamp\n```\n\nDefine a function with a description and a set of examples.\n\n```python\nimport magic_lamp\n\nget_atoms = magic_lamp.Function(\n \"Break this molecule down into it's constituent atoms. Return as a set.\",\n examples=[\n (\"water\", {\"hydrogen\", \"oxygen\"}),\n (\"glucose\", {\"carbon\", \"hydrogen\", \"oxygen\"}),\n ],\n)\n\nprint(get_atoms(\"ammonia\")) # => {\"nitrogen\", \"hydrogen\"}\n```\n\nFunctions can return any Python literal (strings, numbers, dicts, tuples lists, etc). No API keys are required, since by default `magic-lamp` downloads and runs a local LLM.\n\n## Configuring the LLM\n\nBy default, `magic-lamp` downloads and runs a local LLM from Hugging Face. For more complex tasks, OpenAI models will perform better.\n\n### Using OpenAI\n\n`OPENAI_API_KEY` must be set in the environment. Pass in the name of a `gpt-*` model to the function constructor.\n\n```python\nimport magic_lamp\n\nformat_number = magic_lamp.Function(\n 'Write this number out in words.',\n examples=[\n (1, \"one\"),\n (35, \"thirty-five\"),\n (15690, \"fifteen thousand, six hundred ninety\"),\n ],\n model=\"gpt-4o-mini\"\n)\n\nprint(format_number(328745226793))\n```\n\n## Links\n- https://github.com/jackmpcollins/magentic - A similar concept but using decorators.\n- https://github.com/PrefectHQ/marvin - Antoher similar concept.\n- https://github.com/abetlen/llama-cpp-python - Used by this library to run the local LLM.\n- https://ai.meta.com/blog/meta-llama-3-1/ - Llama 3.1 8b is the default model.\n- https://huggingface.co/bullerwins/Meta-Llama-3.1-8B-Instruct-GGUF - Uses these GGUFs by default.\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Easily integrate LLMs into Python code.",
"version": "0.4.1",
"project_urls": {
"Homepage": "https://github.com/rameshvarun/magic-lamp",
"Repository": "https://github.com/rameshvarun/magic-lamp"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "8d4f3ec727a54b982e065fa150e0d2f6c3c4e0ea4a659d36bcce8fc688709975",
"md5": "0ca2ad0cbba664289f27196586a01ace",
"sha256": "ecc215928b03b99b09f3722c215b938e51a0263d8f4f62b6427b7c78a3c87a70"
},
"downloads": -1,
"filename": "magic_lamp-0.4.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "0ca2ad0cbba664289f27196586a01ace",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.9",
"size": 4922,
"upload_time": "2024-08-04T05:35:13",
"upload_time_iso_8601": "2024-08-04T05:35:13.884309Z",
"url": "https://files.pythonhosted.org/packages/8d/4f/3ec727a54b982e065fa150e0d2f6c3c4e0ea4a659d36bcce8fc688709975/magic_lamp-0.4.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "ddb208345d4fddbb974bf8b111fdc2dd80a753e8979de1bd7c5607596ca0f77a",
"md5": "923ec0fb1f1ef8636a35e96ee4fdc521",
"sha256": "234f496ab958c3c221f8090b56a0f6e673c20bd8b3cae072440b5203a48b9f6d"
},
"downloads": -1,
"filename": "magic_lamp-0.4.1.tar.gz",
"has_sig": false,
"md5_digest": "923ec0fb1f1ef8636a35e96ee4fdc521",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.9",
"size": 4460,
"upload_time": "2024-08-04T05:35:15",
"upload_time_iso_8601": "2024-08-04T05:35:15.689149Z",
"url": "https://files.pythonhosted.org/packages/dd/b2/08345d4fddbb974bf8b111fdc2dd80a753e8979de1bd7c5607596ca0f77a/magic_lamp-0.4.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-08-04 05:35:15",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "rameshvarun",
"github_project": "magic-lamp",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "magic-lamp"
}
JSON
