[project]
name = "synthetic-dataset-generator"
version = "0.2.0"
description = "Build datasets using natural language"

authors = [
    { name = "davidberenstein1957", email = "david.m.berenstein@gmail.com" },
]

keywords = [
    "gradio",
    "synthetic-data",
    "huggingface",
    "argilla",
    "generative-ai",
    "ai",
]

requires-python = ">=3.10,<3.13"

readme = "README.md"

license = { text = "Apache-2.0" }

dependencies = [
    "argilla>=2.4.0,<3.0.0",

    # IMPORTANT: this package must also support Gradio 6
    "distilabel[argilla,hf-inference-endpoints,hf-transformers,instructor,llama-cpp,ollama,openai,outlines,vllm,vision]>=1.5.0",

    "gradio[oauth]==6.20.0",

    "gradio-huggingfacehub-search>=0.0.12",

    "huggingface-hub>=0.26.0",

    "model2vec>=0.2.4",

    "nltk>=3.9.1",

    "pydantic>=2.10.5",

    "sentence-transformers>=3.2.0",

    "transformers>=4.44.2",

    "unstructured[md,pdf,docx]>=0.16.3",

    "setuptools",
]

[build-system]
requires = ["pdm-backend"]
build-backend = "pdm.backend"

[tool.pdm]
distribution = true