[project]
name = "misata"
version = "0.9.6.59"
description = "Outcome-conformant synthetic data: realistic multi-table test data that hits the revenue curves, rates, and aggregates you declare, exactly. From a sentence, YAML, or your database. No ML model, no real data. MCP server for AI agents."
authors = [
    {name = "Muhammed Rasin", email = "rasin@misata.studio"}
]
readme = "README.md"
license = {text = "MIT"}
requires-python = ">=3.10"
keywords = [
    "synthetic-data",
    "synthetic-data-python",
    "python-synthetic-data",
    "data-generation",
    "fake-data",
    "test-data-generator",
    "database-seeding",
    "mock-data",
    "tabular-data",
    "multi-table",
    "relational-data",
    "referential-integrity",
    "outcome-conformant",
    "data-conformance",
    "scenario-simulation",
    "pytest-fixtures",
    "dbt",
    "machine-learning",
    "testing",
    "mcp",
    "mcp-server",
    "llm",
    "ai",
    "ai-agents",
    "database",
]
classifiers = [
    "Development Status :: 5 - Production/Stable",
    "Intended Audience :: Developers",
    "Intended Audience :: Science/Research",
    "License :: OSI Approved :: MIT License",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Topic :: Software Development :: Libraries :: Python Modules",
    "Topic :: Software Development :: Testing",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Topic :: Scientific/Engineering :: Information Analysis",
    "Topic :: Database",
    "Topic :: Software Development :: Testing :: Mocking",
    "Topic :: Utilities",
    "Natural Language :: English",
]
dependencies = [
    "pandas>=2.0.0",
    "numpy>=1.24.0",
    "pydantic>=2.0.0",
    "click>=8.1.0",
    "pyyaml>=6.0",
    "rich>=13.0.0",
    "scipy>=1.10.0",
    "networkx>=3.0",
    "jinja2>=3.1.0",
    "faker>=20.0.0",
]

[project.urls]
Homepage = "https://misata.studio"
Studio = "https://misata.studio"
Documentation = "https://misata.studio/docs"
Repository = "https://github.com/rasinmuhammed/misata"
Issues = "https://github.com/rasinmuhammed/misata/issues"
Changelog = "https://github.com/rasinmuhammed/misata/blob/main/CHANGELOG.md"

[project.optional-dependencies]
dev = [
    "pytest>=7.4.0",
    "pytest-cov>=4.0.0",
    "pytest-benchmark>=4.0.0",
    "black>=23.0.0",
    "ruff>=0.1.0",
    "mypy>=1.5.0",
    "simpleeval>=0.9.0",   # formula columns used in HR domain tests
    "jsonschema>=4.0.0",   # YAML validation tests
    "scikit-learn>=1.0.0", # ML-efficacy (TSTR) path in fidelity_report
    "hypothesis>=6.100",   # property-based fuzzing of the dict-schema contract
    "duckdb>=0.9.0",       # evalpack verification tests
    # Upper bound is deliberate: the MCP server targets FastMCP as it exists
    # in the 1.x line. mcp 2.0.0 removed `mcp.server.fastmcp`, which took CI
    # down on all four interpreters the day it shipped. Lift the bound when
    # the server is ported, not before.
    "mcp>=1.0.0,<2.0.0",   # seed_database guardrail tests (the only tool that writes to a DB)
]
db = [
    "psycopg[binary]>=3.1.0",
]
orm = [
    "sqlalchemy>=2.0.0",
]
llm = [
    "groq>=0.4.0",
    "openai>=1.0.0",
]
bedrock = [
    # >=1.39 so the simple Bedrock API key (AWS_BEARER_TOKEN_BEDROCK) is honoured,
    # not just IAM access keys. Converse API (Claude), server-funded LLM path.
    "boto3>=1.39.0",
]
formulas = [
    "simpleeval>=0.9.0",
]
api = [
    "fastapi>=0.109.0",
    "uvicorn>=0.27.0",
    "python-multipart>=0.0.6",
]
studio = [
    "streamlit>=1.30.0",
    "plotly>=5.0.0",
    "openpyxl>=3.0.0",
]
# Advanced features for world-class generation
advanced = [
    "sdv>=1.0.0",           # Copula-based quality generation
    "langgraph>=0.2.0",     # Multi-agent pipeline
    "z3-solver>=4.12.0",    # Constraint satisfaction
]
documents = [
    "weasyprint>=60.0",       # HTML → PDF conversion
]
docs = [
    "mkdocs-material>=9.5.0",
]
mcp = [
    "mcp>=1.0.0,<2.0.0",      # Model Context Protocol, exposes Misata to AI agents (FastMCP is 1.x-only)
    "jsonschema>=4.0.0",      # YAML validation tool
]
kaggle = [
    "kaggle>=1.0.0",          # Kaggle API for vocabulary enrichment from real datasets
]
fidelity = [
    "scikit-learn>=1.0.0",    # ML-efficacy (TSTR) scoring in fidelity_report
]
spark = [
    "pyspark>=3.3.0",         # Apache Spark + Delta Lake integration (write_delta, from_catalog_schema, etc.)
]
evalpack = [
    "duckdb>=0.9.0",          # Independent SQL verification of evalpack answer keys
]
all = [
    "misata[dev,db,orm,llm,formulas,api,studio,advanced,documents,mcp,kaggle,spark,fidelity,evalpack]",
]

[project.scripts]
misata = "misata.cli:main"
misata-studio = "misata.studio:launch"
misata-mcp = "misata.mcp.server:main"

[build-system]
requires = ["setuptools>=68.0.0", "wheel"]
build-backend = "setuptools.build_meta"

[tool.setuptools.packages.find]
include = ["misata*"]
exclude = ["web*", "tests*"]

[tool.setuptools.package-data]
misata = ["py.typed", "_schemas/*.json", "capsules_registry/*.json"]

[tool.black]
line-length = 100
target-version = ['py310']

[tool.ruff]
line-length = 100
target-version = "py310"

# Honest configuration: the public API is typed and py.typed ships, but the
# internals do not pass disallow_untyped_defs today (~500 unannotated defs).
# Strictness is tightened module by module rather than claimed globally.
# See LIMITATIONS.md.
[tool.mypy]
python_version = "3.10"
warn_return_any = true
warn_unused_configs = true
disallow_untyped_defs = false
ignore_missing_imports = true

[tool.pytest.ini_options]
testpaths = ["tests"]
python_files = ["test_*.py"]
python_classes = ["Test*"]
python_functions = ["test_*"]
markers = [
    "slow: marks tests as slow (run with -m slow)",
]