pyproject.toml

[build-system]
requires = ["setuptools>=65.0", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "lobster-ai"
dynamic = ["version"]
authors = [
    {name = "Omics-OS", email = "info@omics-os.com"}
]
description = "Multi-Agent Bioinformatics Analysis System powered by LangGraph"
readme = "README.md"
requires-python = ">=3.12,<3.14"
license = "AGPL-3.0-or-later"
license-files = ["LICENSE"]
classifiers = [
    "Development Status :: 5 - Production/Stable",
    "Intended Audience :: Science/Research",
    "Intended Audience :: Healthcare Industry",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: 3.13",
    "Topic :: Scientific/Engineering :: Bio-Informatics",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Environment :: Console",
    "Framework :: FastAPI",
]
keywords = ["bioinformatics", "RNA-seq", "single-cell", "AI", "machine-learning", "data-analysis", "genomics"]
dependencies = [
    # Core data
    "pandas>=1.5.0",
    "numpy>=1.23.0,<2.4",
    "scipy>=1.10.0",
    "scikit-learn>=1.3.0",
    "anndata>=0.9.0",
    "mudata>=0.2.0",
    "h5py>=3.9.0",
    # Agent framework (only langchain-core, provider packages are extras)
    "langchain-core>=0.3.79",
    "langgraph>=1.0.5",
    # CLI
    "rich>=12.0.0",
    "typer>=0.7.0",
    "python-dotenv>=1.0.0",
    "prompt-toolkit>=3.0.52",
    # Visualization
    "plotly>=5.0.0",
    "matplotlib>=3.7.0",
    "kaleido>=0.2.0",
    # Bioinformatics
    "biopython>=1.81",
    "statsmodels>=0.14.0",
    "gseapy>=1.1.0",
    "GEOparse",
    # Data I/O
    "openpyxl>=3.1.0",
    "pyreadr>=0.4.0",
    "xmltodict>=0.13.0",
    "lxml>=4.9.0",
    # HTTP
    "requests>=2.31.0",
    "httpx>=0.27.0",
    "cloudscraper>=1.2.71",
    # SRA dataset discovery
    "pysradb>=2.5.1",
    # Jupyter notebook support for pipeline replay
    "nbformat>=5.9.0",
    "papermill>=2.4.0",
    # Template rendering (scaffold)
    "jinja2>=3.1",
    # Utilities
    "packaging>=25.0",
    "tabulate>=0.9.0",
    "tomli_w>=1.0",
    "psutil>=7.0.0",
    # Go TUI (default chat UI on supported platforms)
    'lobster-ai-tui~=1.1.0; platform_system != "Windows"',
]

[project.optional-dependencies]
# =============================================================================
# LLM Providers (one required, installed via `lobster init`)
# =============================================================================
anthropic = ["langchain-anthropic>=0.3.20"]
bedrock = ["langchain-aws>=0.1.0", "boto3>=1.26.0"]
ollama = ["langchain-ollama>=0.3.10", "ollama>=0.6.0"]
gemini = ["langchain-google-genai>=4.1.2"]
openai = ["langchain-openai>=0.3.0"]
all-providers = ["lobster-ai[anthropic,bedrock,ollama,gemini,azure,openai]"]

# =============================================================================
# Document Intelligence (optional, installed via `lobster init`)
# =============================================================================
docling = ["docling>=2.60.0", "docling-core>=2.50.0"]

# =============================================================================
# Legacy HDF5 support via PyTables (pandas read_hdf)
# =============================================================================
hdf5 = ["tables>=3.8.0"]

# =============================================================================
# Server mode
# =============================================================================
server = ["fastapi>=0.100.0", "uvicorn>=0.23.0", "python-multipart>=0.0.20"]

# =============================================================================
# Terminal UI (classic Rich/Textual fallback — Go TUI is in core dependencies)
# =============================================================================
classic-tui = ["textual>=6.7.1"]

# =============================================================================
# Semantic Vector Search (optional, installed via `lobster init`)
# =============================================================================
vector-search = [
    # Backend runtime currently lives in lobster-metadata (development package).
    # Keep public extra installable by shipping only dependency prerequisites.
    "chromadb>=1.0.0",
    "openai>=1.0.0",
]

# =============================================================================
# Observability
# =============================================================================
observability = ["langfuse>=3.2.6", "redis>=6.4.0"]

# =============================================================================
# Extended data access
# =============================================================================
extended-data = [
    "polars>=1.32.3",
    "rispy>=0.10.0",
]

# =============================================================================
# Domain extras — each is a complete workflow: agent + deps + research + viz.
#   pip install lobster-ai[transcriptomics]   # scRNA-seq + bulk RNA-seq
#   pip install lobster-ai[genomics]           # VCF, GWAS, variant annotation
#   pip install lobster-ai[proteomics]         # DDA/DIA mass spec
#   pip install lobster-ai[machine-learning]   # Feature selection, survival, deep learning
#   pip install lobster-ai[research]           # Literature search + visualization only
#   pip install lobster-ai[full]               # Everything
# =============================================================================
transcriptomics = [
    "lobster-transcriptomics[batch-integration]~=1.1.0",
    "lobster-research~=1.1.0",
    "lobster-visualization~=1.1.0",
]
# Doublet detection requires scrublet → annoy (C++ compilation).
# Install separately: pip install lobster-transcriptomics[doublet-detection]
doublet-detection = [
    "lobster-transcriptomics[doublet-detection]~=1.1.0",
]
genomics = [
    "lobster-genomics~=1.1.0",
    "lobster-research~=1.1.0",
    "lobster-visualization~=1.1.0",
]
proteomics = [
    "lobster-proteomics[network-analysis]~=1.1.0",
    "lobster-research~=1.1.0",
    "lobster-visualization~=1.1.0",
]
research = [
    "lobster-research~=1.1.0",
    "lobster-visualization~=1.1.0",
]
plink = [
    "bed-reader>=0.2.0",
]

# =============================================================================
# Infrastructure extras
# =============================================================================
azure = [
    "langchain-azure-ai>=0.1.0",
    "azure-ai-agents>=1.0.0b1",
    "azure-ai-inference>=1.0.0b1",
    "azure-ai-projects>=1.0.0b1",
    "azure-identity>=1.25.0",
    "azure-search-documents>=11.6.0",
    "azure-storage-blob>=12.28.0",
]

# =============================================================================
# Convenience meta-extras
# =============================================================================
# What `lobster init` typically installs for a researcher
recommended = ["lobster-ai[anthropic,transcriptomics,classic-tui,extended-data]"]
# Everything available via public lobster-ai extras.
# Excludes metadata_assistant/vector backend until lobster-metadata is published.
full = [
    "lobster-ai[transcriptomics,genomics,proteomics]",
    "lobster-ai[all-providers,classic-tui,extended-data,hdf5,docling]",
]

# =============================================================================
# Development extras
# =============================================================================
dev = [
    # Core testing framework
    "pytest>=7.0.0",
    "pytest-cov>=4.0.0",
    "pytest-xdist>=3.0.0",
    "pytest-asyncio>=0.21.0",
    "pytest-mock>=3.10.0",
    "pytest-benchmark>=4.0.0",
    "pytest-timeout>=2.1.0",
    "pytest-html>=3.2.0",
    "pytest-json-report>=1.5.0",

    # Mock and test data generation
    "factory-boy>=3.2.0",
    "responses>=0.25.8",
    "httpretty>=1.1.4",
    "freezegun>=1.2.0",
    "faker>=19.0.0",

    # Mock services for testing
    "moto>=4.1.0",
    "fakeredis>=2.10.0",
    "ftputil>=5.0.4",

    # Performance and memory profiling
    "memory-profiler>=0.60.0",
    "psutil>=5.9.0",

    # Code quality and formatting
    "black>=23.0.0",
    "isort>=5.12.0",
    "flake8>=6.0.0",
    "pylint>=2.17.0",
    "mypy>=1.0.0",
    "bandit>=1.7.0",
    "ruff>=0.0.300",

    # Development tools
    "pre-commit>=3.0.0",
    "bumpversion>=0.6.0",
    "twine>=4.0.0",
    "build>=0.10.0",

    # Documentation
    "mkdocs>=1.5.0",
    "mkdocs-material>=9.0.0",

    # Additional utilities
    "langfuse>=2.0.0",
    "tabulate>=0.9.0",

    # Architecture validation (Kraken Phase 1)
    "import-linter>=2.1",
    "pipdeptree>=2.25",
]
all = [
    # Complete installation: all domain extras + dev tools
    "lobster-ai[transcriptomics,bulk-rnaseq,genomics,ml,azure,batteries,dev]",
]

[project.urls]
"Homepage" = "https://www.omics-os.com"
"Bug Tracker" = "https://github.com/the-omics-os/lobster/issues"
"Documentation" = "https://github.com/the-omics-os/lobster.wiki"
"Source Code" = "https://github.com/the-omics-os/lobster"

[project.scripts]
lobster = "lobster.cli:app"

# =============================================================================
# Entry Points - Service Discovery (PEP 420 compliant)
# =============================================================================
# IMPORTANT: Only register services that are PUBLIC (in allowlist without '!')
# Premium services (sdk_delegation, extraction_cache, metadata services) should
# be registered in their respective custom package's pyproject.toml
[project.entry-points."lobster.services"]
# Core orchestration services (public as of v2.x - see allowlist line 420)
publication_processing = "lobster.services.orchestration.publication_processing_service:PublicationProcessingService"

# Core queue data structures (public as of v2.x - see allowlist lines 427-429)
publication_queue = "lobster.core.publication_queue:PublicationQueue"

# =============================================================================
# Entry Points - Queue Preparers (PLUG-03, PLUG-04)
# =============================================================================
# All 5 supported databases discoverable via lobster.queue_preparers group.
# Entry-point name MUST match the first value returned by supported_databases().
[project.entry-points."lobster.queue_preparers"]
geo          = "lobster.services.data_access.geo_queue_preparer:GEOQueuePreparer"
sra          = "lobster.services.data_access.sra_queue_preparer:SRAQueuePreparer"
pride        = "lobster.services.data_access.pride_queue_preparer:PRIDEQueuePreparer"
massive      = "lobster.services.data_access.massive_queue_preparer:MassIVEQueuePreparer"
metabolights = "lobster.services.data_access.metabolights_queue_preparer:MetaboLightsQueuePreparer"

# =============================================================================
# Entry Points - Download Services (PLUG-03, PLUG-04)
# =============================================================================
# All 5 supported databases discoverable via lobster.download_services group.
# Entry-point name MUST match supported_databases()[0] for the class.
[project.entry-points."lobster.download_services"]
geo          = "lobster.services.data_access.geo_download_service:GEODownloadService"
sra          = "lobster.services.data_access.sra_download_service:SRADownloadService"
pride        = "lobster.services.data_access.pride_download_service:PRIDEDownloadService"
massive      = "lobster.services.data_access.massive_download_service:MassIVEDownloadService"
metabolights = "lobster.services.data_access.metabolights_download_service:MetaboLightsDownloadService"

# =============================================================================
# Agent Entry Points - Now in Separate Packages
# =============================================================================
# Agent entry points have been moved to their respective packages:
# - lobster-transcriptomics (transcriptomics_expert, annotation_expert, de_analysis_expert)
# - lobster-research (research_agent, data_expert_agent)
# - lobster-visualization (visualization_expert_agent)
# - lobster-metadata (metadata_assistant)
# - lobster-structural-viz (protein_structure_visualization_expert)
# - lobster-genomics (genomics_expert)
# - lobster-proteomics (proteomics_expert)
# - lobster-ml (machine_learning_expert)
#
# Install: pip install lobster-ai[full] for all agents

# =============================================================================
# uv Workspace Configuration (Kraken Phase 1-6)
# =============================================================================
# Enables monorepo package management for modular agent packages.
# Premium packages (Phase 6 extraction):
#   - lobster-genomics: VCF/PLINK, GWAS, variant annotation
#   - lobster-proteomics: MS (DDA/DIA), affinity platforms (Olink, SomaScan)
#   - lobster-ml: Feature engineering, data splitting, scVI embeddings
[tool.uv.workspace]
# Explicit list — private packages (metadata, ml, structural-viz) are .gitignore'd
# and absent from CI checkouts, so wildcard "packages/*" breaks uv build.
members = [
    "packages/lobster-ai-tui",
    "packages/lobster-drug-discovery",
    "packages/lobster-genomics",
    "packages/lobster-metabolomics",
    "packages/lobster-proteomics",
    "packages/lobster-research",
    "packages/lobster-transcriptomics",
    "packages/lobster-visualization",
]

[tool.uv.sources]
# Declare workspace members for [full] extra dependencies
lobster-ai = { workspace = true }
lobster-transcriptomics = { workspace = true }
lobster-research = { workspace = true }
lobster-visualization = { workspace = true }
lobster-genomics = { workspace = true }
lobster-proteomics = { workspace = true }
lobster-ai-tui = { workspace = true }

[tool.setuptools.packages.find]
# Core lobster-ai package discovery with extracted agent exclusions
# Agents have been moved to separate packages in packages/
where = ["."]
include = ["lobster*"]
namespaces = true
exclude = [
    # Extracted FREE tier agent packages (now in packages/)
    "lobster.agents.transcriptomics*",
    "lobster.agents.research*",
    "lobster.agents.data_expert*",
    "lobster.agents.visualization_expert*",
    "lobster.agents.metadata_assistant*",
    "lobster.agents.protein_structure_visualization_expert*",
    # Extracted PREMIUM tier agent packages
    "lobster.agents.genomics*",
    "lobster.agents.proteomics*",
    "lobster.agents.machine_learning_expert*",
    "lobster.agents.drug_discovery*",

    # Extracted services (Phase 11.1)
    # Transcriptomics services
    "lobster.services.analysis.enhanced_singlecell_service",
    "lobster.services.analysis.differential_formula_service",
    "lobster.services.analysis.pseudobulk_service",
    "lobster.services.quality.preprocessing_service",
    "lobster.services.quality.quality_service",
    "lobster.services.metadata.manual_annotation_service",
    "lobster.services.visualization.bulk_visualization_service",
    "lobster.services.templates.annotation_templates",

    # Proteomics services
    "lobster.services.analysis.proteomics_analysis_service",
    "lobster.services.analysis.proteomics_differential_service",
    "lobster.services.analysis.proteomics_network_service",
    "lobster.services.analysis.proteomics_survival_service",
    "lobster.services.quality.proteomics_preprocessing_service",
    "lobster.services.quality.proteomics_quality_service",
    "lobster.services.visualization.proteomics_visualization_service",
    "lobster.services.data_access.parsers.spectronaut_parser",
    "lobster.services.data_access.parsers.diann_parser",
    "lobster.services.data_access.parsers.maxquant_parser",
    "lobster.services.data_access.parsers.olink_parser",

    # Genomics services
    "lobster.services.analysis.gwas_service",
    "lobster.services.analysis.variant_annotation_service",
    "lobster.services.quality.genomics_quality_service",

    # ML services
    "lobster.services.ml.ml_preparation_service",
    "lobster.services.ml.ml_transcriptomics_service_ALPHA",
    "lobster.services.ml.ml_proteomics_service_ALPHA",
    "lobster.services.analysis.scvi_embedding_service",

    # Drug discovery services
    "lobster.services.drug_discovery*",

    # Structural-viz services
    "lobster.services.analysis.structure_analysis_service",
    "lobster.services.visualization.pymol_visualization_service",
    "lobster.services.visualization.chimerax_visualization_service_ALPHA",
    "lobster.services.data_access.protein_structure_fetch_service",

    # Visualization services
    "lobster.services.visualization.visualization_service",

    # Metadata services
    "lobster.services.metadata.disease_ontology_service",
    "lobster.services.metadata.disease_standardization_service",
    "lobster.services.metadata.metadata_filtering_service",
    "lobster.services.metadata.metadata_standardization_service",
    "lobster.services.metadata.microbiome_filtering_service",
    "lobster.services.metadata.sample_mapping_service",
    "lobster.services.metadata.sample_grouping_service",
    "lobster.services.metadata.clinical_metadata_service",

    # Research services
    "lobster.services.data_management.modality_detection_service",

    # Vector search (moved to lobster-metadata)
    "lobster.services.vector*",
]
# Core keeps: supervisor, graph.py, state, services, tools, config, prompts

[tool.setuptools.dynamic]
version = {attr = "lobster.version.__version__"}

[tool.setuptools.package-data]
lobster = ["**/*.json", "**/*.yaml", "**/*.yml"]
"lobster.scaffold" = ["templates/*.j2"]

# Development tools configurations
[tool.black]
line-length = 88
target-version = ["py311"]

[tool.isort]
profile = "black"
line_length = 88

[tool.pytest.ini_options]
testpaths = ["tests"]
python_files = "test_*.py"
python_classes = "Test*"
python_functions = "test_*"
addopts = [
    "-v",
    "--strict-markers",
    "--strict-config",
    "--cov=lobster",
    "--cov-report=term-missing",
    "--cov-report=html:htmlcov",
    "--cov-report=xml:coverage.xml",
    "--cov-fail-under=0",
    "--durations=10",
    "--tb=short",
    "--import-mode=importlib",
]
markers = [
    "unit: Unit tests for individual components",
    "integration: Integration tests for multi-component workflows",
    "contract: Contract tests validating plugin API and AQUADIF taxonomy compliance",
    "system: System tests for complete workflows",
    "performance: Performance and benchmarking tests",
    "slow: Tests that take more than 30 seconds",
    "requires_api: Tests that require external API access",
    "real_api: Tests that make real API calls to external services (PubMed, GEO, etc.)",
    "requires_gpu: Tests that require GPU resources",
    "flaky: Flaky tests that may fail intermittently",
    "stress: Stress tests",
    "ux: User experience tests",
    "manual: Manual tests not for CI",
    "external: Tests requiring external services",
    "very_slow: Tests that take more than 5 minutes",
    "nightly: Tests for nightly CI only",
]
filterwarnings = [
    "ignore::DeprecationWarning:pkg_resources.*",
    "ignore::PendingDeprecationWarning",
    "ignore::UserWarning:anndata.*",
]
minversion = "7.0"
timeout = 300
collect_ignore = ["setup.py", "build/", "dist/", ".tox/", ".eggs/", "venv/", ".venv/", "env/", ".env/", "tests/archive/"]
collect_ignore_glob = ["**/.*", "**/__pycache__/**", "**/node_modules/**", "**/.DS_Store", "tests/archive/**", "kevin_notes/**", "kevin_notes_completed/**", "tests/manual/**"]
norecursedirs = [".git", ".tox", ".eggs", "venv", ".venv", "build", "dist", "node_modules", "__pycache__", "tests/archive", "tests/manual", "kevin_notes", "kevin_notes_completed"]

[tool.mypy]
python_version = "3.11"
warn_return_any = true
warn_unused_configs = true
disallow_untyped_defs = true
disallow_incomplete_defs = true

[tool.bumpversion]
current_version = "0.3.3"
commit = true
tag = true

[dependency-groups]
dev = [
    "psutil>=7.0.0",
    "pytest-asyncio>=1.1.0",
    "pytest-cov>=6.2.1",
]

[tool.ruff]
line-length = 88

[tool.ruff.lint]
ignore = [
    "E402",  # module-import-not-at-top-of-file (intentional lazy loading)
]