pyproject.toml

[project]
name = "edgecodedpo"
description = "A toolkit for generating and fine-tuning code models using Direct Preference Optimization (DPO)"
authors = [{ name = "Simon", email = "dubailsimon@gmail.com" }]
readme = "README.md"
classifiers = [
  "Intended Audience :: Developers",
  "Natural Language :: English",
  "Programming Language :: Python",
  "Programming Language :: Python :: 3", 
  "Programming Language :: Python :: 3.12",
  "Topic :: Software Development",
]
requires-python = ">=3.10"
dynamic = ["version"]

scripts = { edgecodedpo = "edgecodedpo:__main__.app" }

dependencies = [
    # Core dependencies
    "numpy",
    "matplotlib",
    "typer",
    "tqdm",
    "pandas",
    "pydantic_settings",
    "aiohttp>=3.8.0",
    "pathlib",

    # Dataset generation
    "datasets>=2.14.0",
    "openai>=1.0.0",
    "huggingface-hub>=0.19.0",
    "seaborn",
    
    # DPO training
    "torch>=2.0.0",
    "transformers>=4.38.0",
    "trl>=0.8.1",
    "accelerate>=0.27.0",
    "bitsandbytes>=0.41.0",
    "peft>=0.7.0",
    "safetensors>=0.4.0",
    "tensorboard>=2.15.0",
    "rich>=13.5.0",
    "scikit-learn==1.6.1",
    "radon==6.0.1",
    "pylint==3.3.5",

]

[project.optional-dependencies]
dev = [
  "ruff==0.5.*",
  "mypy==1.11.*",
  "pytest"
]
doc = [
]

# Optional dependencies for specific training setups
# training = [
#   "wandb",            # For tracking experiments
#   "deepspeed>=0.12.0", # For distributed training
#   "unsloth>=0.5.0",   # For faster LoRA training
# ]

all = ["edgecodedpo[dev]", "edgecodedpo[doc]", "edgecodedpo[training]"]

[tool.ruff]
target-version = "py310"
unsafe-fixes = true

[tool.ruff.lint]
select = [
  "E",    # pycodestyle
  "F",    # pyflakes
  "UP",   # pyupgrade
  "I",    # isort
  "PERF", # Perflint
  "S",    # flake8-bandit
  "B",    # flake8-bugbear
  "C4",   # flake8-comprehensions
  "DTZ",  # flake8-datetimez
  "TID",  # flake8-tidy-imports
  "ICN",  # flake8-import-conventions
  # "TCH",  # flake8-type-checking
  "RUF", # Ruff-specific rules
]

ignore = [
  "E501",   # pycodestyle line too long
  "B010",   # flake8-bugbear - do not call setattr with a constant attribute value
  "RUF012", # Ruff-specific rule - annotated with classvar
  "ISC001", # Ruff formatter incompatible
  "S608",   # flake8-bandit - Possible SQL injection vector through string-based query construction
  "B008",   # flake8-bugbear - contradict FastAPI
]

[tool.ruff.lint.flake8-import-conventions.aliases]
numpy = "np"
pandas = "pd"
torch = "torch"

[tool.ruff.lint.flake8-tidy-imports]
ban-relative-imports = "all"                 # Disallow all relative imports.
banned-module-level-imports = ["matplotlib"] # require a lazy import

[tool.ruff.lint.isort]
known-first-party = ["edgecodedpo", "tests"]

[tool.ruff.lint.per-file-ignores]
"tests/**/*.*" = [
  "A",
  "ARG",
  "B",
  "BLE",
  "C901",
  "D",
  "DTZ",
  "EM",
  "FBT",
  "G",
  "N",
  "PGH",
  "PIE",
  "PLR",
  "PLW",
  "PTH",
  "RSE",
  "S",
  "S101",
  "SIM",
  # "TCH",
  "TRY",
]

[tool.mypy]
# plugins = ["pydantic.mypy"]
python_version = "3.12" 

# Rules can be found at:
# https://mypy.readthedocs.io/en/latest/config_file.html

strict = true
warn_return_any = false
show_error_codes = true

# permissives rules:
implicit_reexport = true      # no need to specify __all__ in modules
ignore_missing_imports = true # ignore missing stubs
# follow_imports = "silent"  # does not work when using dmypy https://github.com/python/mypy/issues/9475


[[tool.mypy.overrides]]
module = ["tests.*", "docs.*"]
disallow_untyped_defs = false
disallow_untyped_calls = false
disallow_incomplete_defs = false
disallow_untyped_decorators = false

[tool.pydantic-mypy]
init_forbid_extra = true
init_typed = true
warn_required_dynamic_aliases = true

# [tool.pyright]

[tool.cruft]
skip = ["edgecodedpo/__init__.py", "tests", "playground"]