[project]
name = "minirag-mcp"
version = "0.6.1"
description = "Local-first RAG MCP server: hybrid search over your documents, fully local"
readme = "README.md"
requires-python = ">=3.11"
license = "MIT"
dependencies = [
    "fastmcp>=3.4",
    "lancedb>=0.36",
    "fastembed>=0.8",
    "markitdown[pdf,docx,pptx,xlsx]>=0.1.7",
    "cyclopts>=4.0",
    "platformdirs>=4.0",
    # markitdown already pulls this in, but we import it directly: the URL fetch runs
    # through a requests session we build and guard ourselves.
    "requests>=2.32",
    # fastembed already pulls this in, but we import it directly: the chunker's token
    # counter is built from a copy of the model's own tokenizer with truncation off.
    "tokenizers>=0.15",
    # markitdown[pdf] already pulls this in, but we import it directly: OCR routing
    # counts a PDF's text per page with pdfminer's own layout pass, and that import
    # runs at startup on every install, extra or not. The bound matches the floor
    # markitdown[pdf] already imposes, so declaring it can never move the resolution.
    "pdfminer.six>=20251230",
]

[project.scripts]
minirag-mcp = "minirag_mcp.__main__:main"

[project.optional-dependencies]
ocr = [
    "rapidocr>=3.9",
    "onnxruntime>=1.19",
    "pypdfium2>=5",
    "rapid-orientation>=0.0.11",
    # rapidocr already pulls this in, but we import it directly to decode image files,
    # and its own floor (4.5.1.48) is below both of the floors that bind here. The API
    # floor is 4.7.0.68: `imdecodemulti` — the call that keeps every frame of a
    # multi-page TIFF — is absent from 4.5.5.62, 4.5.5.64 and 4.6.0.66, and no wheel
    # sits between 4.6.0.66 and 4.7.0.68. The floor declared is higher because it is
    # the numpy one: no wheel before 4.10.0.84 imports against numpy 2.x (4.7.0.68
    # through 4.10.0.82 all raise "numpy.core.multiarray failed to import"), and
    # nothing in this dependency graph permits numpy 1.x — fastembed, a core
    # dependency, requires numpy>=2.1 on 3.13. All four components are spelled out
    # because `>=4.10` would admit 4.10.0.82, which is one of the broken ones.
    "opencv-python>=4.10.0.84",
]

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

[tool.hatch.build.targets.wheel]
packages = ["src/minirag_mcp"]

[dependency-groups]
dev = ["pytest>=8", "pytest-asyncio>=1.0", "ruff>=0.6", "bump-my-version>=1.5.1"]

# `uv run bump-my-version bump <major|minor|patch>` does the whole release
# prologue: rewrite `version` above, commit it, tag it. It never pushes — the
# push is what makes a version number real on PyPI, and everything up to it is
# still reversible, so that stays a separate, deliberate command.
#
# `__version__` is derived from installed metadata, so `version` in [project]
# is the only place a number lives and the only thing this has to rewrite.
[tool.bumpversion]
current_version = "0.6.1"
commit = true
message = "chore: release v{new_version}"
tag = true
# release.yml strips one leading "v" from the tag before comparing it against
# the built wheel's version, so the tag has to be spelled exactly this way.
# Change one without the other and a release either fails that check or needs
# a retag after the fact.
tag_name = "v{new_version}"
tag_message = "v{new_version}"
sign_tags = false
# A bump has to be the only thing in its commit; refuse to sweep up whatever
# else happens to be in the tree.
allow_dirty = false

# bump-my-version 1.5.1 also rewrites `project.version` unprompted, but that is
# not documented configuration and a release now depends on that line being
# right, so it is pinned here too. The two agree; the line is simply written
# twice. Anchored to a whole line because `current_version = "0.1.0"` above
# ends in the same `version = "0.1.0"` text, which a bare search would match.
[[tool.bumpversion.files]]
filename = "pyproject.toml"
regex = true
search = '^version = "{current_version}"$'
replace = 'version = "{new_version}"'

# uv.lock pins this project's own version alongside its dependencies. Left out,
# it stays a release behind until the next `uv run` rewrites it — and that dirty
# file then blocks the *following* bump, because allow_dirty is false. Bumping
# it here keeps the lock in the release commit where it belongs. Anchored to the
# minirag-mcp package block, since other packages in the lock have their own
# `version = ` lines.
[[tool.bumpversion.files]]
filename = "uv.lock"
regex = true
search = 'name = "minirag-mcp"\nversion = "{current_version}"'
replace = 'name = "minirag-mcp"\nversion = "{new_version}"'

# server.json is the MCP Registry's copy of this release, and it spells the
# version twice: once for the server entry and once for the PyPI package it
# points at. Both must equal the tag — publish-mcp.yml refuses a tag that
# disagrees with either, and the registry itself rejects a re-publish of a
# version it already holds, so a stale server.json fails the publish of a tag
# whose wheel is already on PyPI, and the only way out of that is another
# release. The two lines differ only by indentation, so they get one entry
# each, anchored to whole lines; a single unanchored search would rewrite
# whichever bump-my-version happened to reach first and silently leave the
# other behind. tests/test_scaffold.py asserts both against pyproject.toml.
[[tool.bumpversion.files]]
filename = "server.json"
regex = true
search = '^  "version": "{current_version}",$'
replace = '  "version": "{new_version}",'

[[tool.bumpversion.files]]
filename = "server.json"
regex = true
search = '^      "version": "{current_version}",$'
replace = '      "version": "{new_version}",'

[tool.pytest.ini_options]
addopts = "-m 'not slow'"
markers = ["slow: needs the real embedding model (downloads ~220 MB)"]
asyncio_mode = "auto"
testpaths = ["tests"]

[tool.ruff]
line-length = 100
src = ["src", "tests"]

[tool.ruff.lint]
select = ["E", "F", "I", "UP", "B"]

[tool.ruff.lint.per-file-ignores]
# Verbatim excerpts from the real corpus. Rewrapping a markdown table row or a
# transcript turn would change the very shape the chunker is being tested against.
"tests/corpus_samples.py" = ["E501"]