[build-system]
requires = ["poetry-core>=1.0.0"]
build-backend = "poetry.core.masonry.api"

[tool.poetry]
name = "airbyte-cdk"
version = "5.17.0"
description = "A framework for writing Airbyte Connectors."
authors = ["Airbyte <contact@airbyte.io>"]
license = "MIT"
readme = "README.md"
homepage = "https://github.com/airbytehq/airbyte"
repository = "https://github.com/airbytehq/airbyte"
documentation = "https://docs.airbyte.io/"
classifiers = [
    "Development Status :: 3 - Alpha",
    "Intended Audience :: Developers",
    "Topic :: Scientific/Engineering",
    "Topic :: Software Development :: Libraries :: Python Modules",
    "License :: OSI Approved :: MIT License",
    "Programming Language :: Python :: 3.10",
]
keywords = ["airbyte", "connector-development-kit", "cdk"]


[tool.poetry.dependencies]
python = "^3.10"
airbyte-protocol-models-dataclasses = "^0.13"
backoff = "*"
cachetools = "*"
Deprecated = "~1.2"
dpath = "^2.1.6"
genson = "1.2.2"
isodate = "~0.6.1"
Jinja2 = "~3.1.2"
jsonref = "~0.2"
jsonschema = "~3.2.0"
pandas = "2.2.2"
pendulum = "<3.0.0"
pydantic = "^2.7"
pyrate-limiter = "~3.1.0"
python-dateutil = "*"
PyYAML = "^6.0.1"
requests = "*"
requests_cache = "*"
wcmatch = "8.4"
# Extras depedencies
avro = { version = "~1.11.2", optional = true }
cohere = { version = "4.21", optional = true }
fastavro = { version = "~1.8.0", optional = true }
langchain = { version = "0.1.16", optional = true }
langchain_core = { version = "0.1.42", optional = true }
markdown = { version = "*", optional = true }
openai = { version = "0.27.9", extras = ["embeddings"], optional = true }
pdf2image = { version = "1.16.3", optional = true }
"pdfminer.six" = { version = "20221105", optional = true }
pyarrow = { version = "~15.0.0", optional = true }
pytesseract = { version = "0.3.10", optional = true }
python-calamine = { version = "0.2.3", optional = true }
python-snappy = { version = "0.7.3", optional = true }
Sphinx = { version = "~4.2", optional = true }
sphinx-rtd-theme = { version = "~1.0", optional = true }
tiktoken = { version = "0.4.0", optional = true }
nltk = { version = "3.8.1", optional = true }
unstructured = { version = "0.10.27", extras = ["docx", "pptx"], optional = true }
"unstructured.pytesseract" = { version = ">=0.3.12", optional = true }
pyjwt = "^2.8.0"
cryptography = "^42.0.5"
pytz = "2024.1"
orjson = "^3.10.7"
serpyco-rs = "^1.10.2"
sqlalchemy = {version = "^2.0,!=2.0.36", optional = true }
xmltodict = "^0.13.0"

[tool.poetry.group.dev.dependencies]
freezegun = "*"
mypy = "*"
asyncio = "3.4.3"
poethepoet = "^0.24.2"
pyproject-flake8 = "^6.1.0"
pytest = "^7"
pytest-memray = "^1.6.0"
pytest-cov = "*"
pytest-httpserver = "*"
pytest-mock = "*"
requests-mock = "*"

[tool.poetry.extras]
file-based = ["avro", "fastavro", "pyarrow", "unstructured", "pdf2image", "pdfminer.six", "unstructured.pytesseract", "pytesseract", "markdown", "python-calamine", "python-snappy"]
sphinx-docs = ["Sphinx", "sphinx-rtd-theme"]
vector-db-based = ["langchain", "openai", "cohere", "tiktoken"]
sql = ["sqlalchemy"]

[tool.ruff]
# Setting python version to at least 3.10 avoids `from __future__ import annotations`.
target-version = "py310"
# This is consistent with airbytehq/airbyte root pyproject.toml Black rule defined.
line-length = 140

[tool.poe.tasks]
# Build tasks
assemble = {cmd = "bin/generate-component-manifest-dagger.sh", help = "Generate component manifest files."}
build-package = {cmd = "poetry build", help = "Build the python package: source and wheels archives."}
build = {sequence = ["assemble", "build-package"], help = "Run all tasks to build the package."}

# Check tasks
lint =  {cmd = "pflake8 --config ../../pyproject.toml ./", help = "Lint with flake8."}
type-check = {cmd = "bin/run-mypy-on-modified-files.sh", help = "Type check modified files with mypy."}
unit-test-with-cov = {cmd = "pytest -s unit_tests -c pytest.ini --cov=airbyte_cdk --cov-report=term --cov-config ../../pyproject.toml", help = "Run unit tests and create a coverage report."}
# TODO: find a version of the modified mypy check that works both locally and in CI.
check-lockfile = {cmd = "poetry check", help = "Check the poetry lock file."}
check-local = {sequence = ["lint", "type-check", "check-lockfile", "unit-test-with-cov"], help = "Lint all code, type-check modified files, and run unit tests."}
check-ci = {sequence = ["check-lockfile", "build", "lint", "unit-test-with-cov"], help = "Build the package, lint and run unit tests. Does not include type-checking."}

# Build and check
pre-push = {sequence = ["build", "check-local"], help = "Run all build and check tasks."}

[tool.airbyte_ci]
python_versions = ["3.10", "3.11"]
optional_poetry_groups = ["dev"]
poetry_extras = ["file-based", "sphinx-docs", "vector-db-based"]
poe_tasks = ["check-ci"]
mount_docker_socket = true