Project Architecture Patterns

Purpose

Establish consistent architecture patterns for the rRNA-Phylo project, covering backend design, service organization, API conventions, and testing strategies.

When to Use

This skill activates when:

• Setting up new services or modules
• Designing API endpoints
• Implementing background tasks
• Working with configuration or dependencies
• Writing tests
• Integrating external tools (HMMER, BLAST, alignment tools)

Tech Stack

Core Backend

• FastAPI: Modern async web framework
• Pydantic: Data validation and settings
• SQLAlchemy: ORM for database operations
• Celery: Distributed task queue for long-running jobs
• Redis: Message broker and cache
• PostgreSQL: Primary database (SQLite for dev)

Scientific Computing

• Biopython: Sequence parsing and analysis
• NumPy/Pandas: Numerical computing and data manipulation
• scikit-learn: ML baseline models
• PyTorch: Deep learning (Phase 2+)

Testing & Quality

• pytest: Testing framework
• pytest-asyncio: Async test support
• pytest-cov: Coverage reporting
• httpx: Async HTTP client for API tests
• ruff: Fast linter
• black: Code formatter

---

Project Structure

backend/
├── app/
│   ├── __init__.py
│   ├── main.py                    # FastAPI application
│   ├── config.py                  # Settings (via Pydantic)
│   │
│   ├── models/                    # SQLAlchemy ORM models
│   │   ├── __init__.py
│   │   ├── base.py               # Base model class
│   │   ├── job.py                # Job model
│   │   ├── sequence.py           # Sequence model
│   │   └── tree.py               # Phylogenetic tree model
│   │
│   ├── schemas/                   # Pydantic schemas (API contracts)
│   │   ├── __init__.py
│   │   ├── common.py             # Shared schemas
│   │   ├── rrna.py               # rRNA detection schemas
│   │   ├── phylo.py              # Phylogenetics schemas
│   │   └── job.py                # Job status schemas
│   │
│   ├── api/                       # API routes
│   │   ├── __init__.py
│   │   ├── deps.py               # Dependency injection
│   │   └── v1/
│   │       ├── __init__.py
│   │       ├── router.py         # Main router
│   │       ├── rrna.py           # rRNA endpoints
│   │       ├── phylo.py          # Phylogenetics endpoints
│   │       └── jobs.py           # Job management
│   │
│   ├── services/                  # Business logic layer
│   │   ├── rrna/                 # rRNA detection service
│   │   ├── phylo/                # Phylogenetics service
│   │   └── sequences/            # Sequence processing
│   │
│   ├── workers/                   # Celery tasks
│   │   ├── __init__.py
│   │   ├── celery_app.py         # Celery config
│   │   └── tasks.py              # Task definitions
│   │
│   ├── core/                      # Core utilities
│   │   ├── __init__.py
│   │   ├── errors.py             # Custom exceptions
│   │   └── logging.py            # Logging setup
│   │
│   └── db/                        # Database utilities
│       ├── __init__.py
│       ├── session.py            # DB session management
│       └── migrations/           # Alembic migrations
│
├── tests/
│   ├── conftest.py               # Pytest fixtures
│   ├── unit/                     # Unit tests
│   ├── integration/              # Integration tests
│   └── fixtures/                 # Test data
│
├── requirements.txt
├── pyproject.toml
└── Dockerfile

---

Core Patterns

1. Configuration Management

Pattern: Use Pydantic Settings for type-safe configuration.

# app/config.py
from pydantic_settings import BaseSettings
from functools import lru_cache

class Settings(BaseSettings):
    """Application settings."""

    # API Settings
    api_v1_prefix: str = "/api/v1"
    project_name: str = "rRNA-Phylo"
    debug: bool = False

    # Database
    database_url: str = "sqlite:///./rrna_phylo.db"

    # Celery
    celery_broker_url: str = "redis://localhost:6379/0"
    celery_result_backend: str = "redis://localhost:6379/0"

    # External Tools
    hmmer_path: str = "/usr/local/bin/hmmsearch"
    blast_path: str = "/usr/local/bin/blastn"

    # Data Paths
    hmm_profiles_dir: str = "./data/hmm_profiles"
    blast_db_dir: str = "./data/blast_dbs"

    # ML Models
    ml_models_dir: str = "./models"

    class Config:
        env_file = ".env"
        case_sensitive = False

@lru_cache()
def get_settings() -> Settings:
    """Get cached settings instance."""
    return Settings()

Usage in dependencies:

# app/api/deps.py
from fastapi import Depends
from app.config import Settings, get_settings

async def get_config() -> Settings:
    return get_settings()

2. Service Layer Pattern

Pattern: Encapsulate business logic in service classes.

# app/services/rrna/detector.py
from typing import List, Optional
from app.schemas.rrna import RRNADetectionResult, RRNAType
from app.config import Settings

class RRNADetectorService:
    """Service for detecting rRNA in sequences."""

    def __init__(self, settings: Settings):
        self.settings = settings
        self.hmm_detector = HMMDetector(settings.hmm_profiles_dir)
        self.pattern_detector = PatternDetector()

    async def detect(
        self,
        sequence: str,
        methods: List[str] = ["hmm", "pattern"],
        min_confidence: float = 0.8
    ) -> List[RRNADetectionResult]:
        """
        Detect rRNA in sequence using multiple methods.

        Args:
            sequence: DNA/RNA sequence string
            methods: Detection methods to use
            min_confidence: Minimum confidence threshold

        Returns:
            List of detection results
        """
        results = []

        if "hmm" in methods:
            hmm_results = await self.hmm_detector.detect(sequence)
            results.extend(hmm_results)

        if "pattern" in methods:
            pattern_results = await self.pattern_detector.detect(sequence)
            results.extend(pattern_results)

        # Filter by confidence
        results = [r for r in results if r.confidence >= min_confidence]

        # Deduplicate and merge
        return self._merge_results(results)

    def _merge_results(
        self,
        results: List[RRNADetectionResult]
    ) -> List[RRNADetectionResult]:
        """Merge overlapping detections."""
        # Implementation here
        pass

Usage in API:

# app/api/v1/rrna.py
from fastapi import APIRouter, Depends
from app.services.rrna.detector import RRNADetectorService
from app.api.deps import get_rrna_detector

router = APIRouter()

@router.post("/detect")
async def detect_rrna(
    sequence: str,
    detector: RRNADetectorService = Depends(get_rrna_detector)
):
    """Detect rRNA in sequence."""
    results = await detector.detect(sequence)
    return {"results": results}

3. Pydantic Schemas (API Contracts)

Pattern: Define clear input/output schemas with validation.

# app/schemas/rrna.py
from pydantic import BaseModel, Field, validator
from typing import List, Optional
from enum import Enum

class RRNAType(str, Enum):
    """Supported rRNA types."""
    SSU_16S = "16S"
    SSU_18S = "18S"
    LSU_23S = "23S"
    LSU_28S = "28S"
    FIVE_S = "5S"
    FIVE_EIGHT_S = "5.8S"

class DetectionMethod(str, Enum):
    """Detection methods."""
    HMM = "hmm"
    BLAST = "blast"
    PATTERN = "pattern"
    ML = "ml"

class RRNADetectionRequest(BaseModel):
    """Request schema for rRNA detection."""

    sequence: str = Field(..., min_length=100, max_length=100000)
    methods: List[DetectionMethod] = Field(
        default=[DetectionMethod.HMM, DetectionMethod.PATTERN]
    )
    min_confidence: float = Field(default=0.8, ge=0.0, le=1.0)

    @validator("sequence")
    def validate_sequence(cls, v):
        """Ensure sequence contains only valid nucleotides."""
        valid_chars = set("ACGTUNacgtun-")
        if not set(v).issubset(valid_chars):
            raise ValueError("Sequence contains invalid characters")
        return v.upper()

class ConservedRegion(BaseModel):
    """Conserved region within detected rRNA."""
    name: str
    start: int
    end: int
    sequence: str
    confidence: float

class RRNADetectionResult(BaseModel):
    """Result schema for rRNA detection."""

    rrna_type: RRNAType
    start: int
    end: int
    length: int
    confidence: float
    method: DetectionMethod
    score: float

    # Quality metrics
    completeness: float = Field(..., ge=0.0, le=1.0)
    quality: str = Field(..., pattern="^(high|medium|low|very_low)$")

    # Optional details
    conserved_regions: Optional[List[ConservedRegion]] = None
    secondary_structure: Optional[str] = None

    class Config:
        json_schema_extra = {
            "example": {
                "rrna_type": "16S",
                "start": 0,
                "end": 1542,
                "length": 1542,
                "confidence": 0.95,
                "method": "hmm",
                "score": 1250.5,
                "completeness": 0.98,
                "quality": "high"
            }
        }

4. Dependency Injection

Pattern: Use FastAPI's dependency injection for services.

# app/api/deps.py
from typing import Generator
from fastapi import Depends
from sqlalchemy.orm import Session

from app.db.session import SessionLocal
from app.config import Settings, get_settings
from app.services.rrna.detector import RRNADetectorService

# Database dependency
def get_db() -> Generator:
    """Get database session."""
    db = SessionLocal()
    try:
        yield db
    finally:
        db.close()

# Service dependencies
def get_rrna_detector(
    settings: Settings = Depends(get_settings)
) -> RRNADetectorService:
    """Get rRNA detector service."""
    return RRNADetectorService(settings)

def get_phylo_service(
    settings: Settings = Depends(get_settings),
    db: Session = Depends(get_db)
):
    """Get phylogenetic analysis service."""
    from app.services.phylo.tree_builder import PhyloService
    return PhyloService(settings, db)

5. Error Handling

Pattern: Use custom exceptions and FastAPI exception handlers.

# app/core/errors.py
class RRNAPhyloException(Exception):
    """Base exception for rRNA-Phylo."""
    pass

class SequenceValidationError(RRNAPhyloException):
    """Raised when sequence validation fails."""
    pass

class DetectionError(RRNAPhyloException):
    """Raised when rRNA detection fails."""
    pass

class AlignmentError(RRNAPhyloException):
    """Raised when sequence alignment fails."""
    pass

class TreeBuildingError(RRNAPhyloException):
    """Raised when tree building fails."""
    pass

# app/main.py
from fastapi import FastAPI, Request
from fastapi.responses import JSONResponse
from app.core.errors import RRNAPhyloException

app = FastAPI()

@app.exception_handler(RRNAPhyloException)
async def rrna_phylo_exception_handler(
    request: Request,
    exc: RRNAPhyloException
):
    """Handle custom exceptions."""
    return JSONResponse(
        status_code=400,
        content={
            "error": exc.__class__.__name__,
            "message": str(exc)
        }
    )

6. Async Task Processing (Celery)

Pattern: Use Celery for long-running tasks.

# app/workers/celery_app.py
from celery import Celery
from app.config import get_settings

settings = get_settings()

celery_app = Celery(
    "rrna_phylo",
    broker=settings.celery_broker_url,
    backend=settings.celery_result_backend
)

celery_app.conf.update(
    task_serializer="json",
    accept_content=["json"],
    result_serializer="json",
    timezone="UTC",
    enable_utc=True,
    task_track_started=True,
    task_time_limit=3600,  # 1 hour max
)

# app/workers/tasks.py
from app.workers.celery_app import celery_app
from app.services.phylo.tree_builder import PhyloService

@celery_app.task(bind=True)
def build_phylogenetic_tree(
    self,
    sequences: List[dict],
    method: str,
    parameters: dict
):
    """
    Build phylogenetic tree (long-running task).

    Args:
        self: Task instance (for progress updates)
        sequences: List of sequences
        method: Tree building method
        parameters: Method parameters
    """
    try:
        # Update progress
        self.update_state(state="PROGRESS", meta={"stage": "alignment"})

        service = PhyloService()

        # Align sequences
        alignment = service.align_sequences(sequences)

        self.update_state(state="PROGRESS", meta={"stage": "tree_building"})

        # Build tree
        tree = service.build_tree(alignment, method, parameters)

        return {"tree": tree, "status": "completed"}

    except Exception as e:
        self.update_state(state="FAILURE", meta={"error": str(e)})
        raise

API integration:

# app/api/v1/phylo.py
from fastapi import APIRouter, BackgroundTasks
from app.workers.tasks import build_phylogenetic_tree

router = APIRouter()

@router.post("/tree")
async def create_tree(request: TreeBuildRequest):
    """Submit tree building job."""

    # Submit Celery task
    task = build_phylogenetic_tree.delay(
        sequences=request.sequences,
        method=request.method,
        parameters=request.parameters
    )

    return {
        "job_id": task.id,
        "status": "submitted"
    }

@router.get("/tree/{job_id}")
async def get_tree_status(job_id: str):
    """Get tree building job status."""
    task = build_phylogenetic_tree.AsyncResult(job_id)

    return {
        "job_id": job_id,
        "status": task.state,
        "result": task.result if task.ready() else None
    }

---

Testing Patterns

1. Test Structure

# tests/conftest.py
import pytest
from fastapi.testclient import TestClient
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker

from app.main import app
from app.db.session import Base, get_db

# Test database
SQLALCHEMY_TEST_URL = "sqlite:///./test.db"

@pytest.fixture(scope="session")
def test_engine():
    """Create test database engine."""
    engine = create_engine(SQLALCHEMY_TEST_URL)
    Base.metadata.create_all(bind=engine)
    yield engine
    Base.metadata.drop_all(bind=engine)

@pytest.fixture(scope="function")
def test_db(test_engine):
    """Create test database session."""
    TestSessionLocal = sessionmaker(bind=test_engine)
    db = TestSessionLocal()
    try:
        yield db
    finally:
        db.close()

@pytest.fixture
def client(test_db):
    """Create test client."""
    def override_get_db():
        yield test_db

    app.dependency_overrides[get_db] = override_get_db
    yield TestClient(app)
    app.dependency_overrides.clear()

@pytest.fixture
def sample_16s_sequence():
    """Sample 16S rRNA sequence for testing."""
    return "AGAGTTTGATCCTGGCTCAG..."  # Truncated

2. Unit Test Example

# tests/unit/test_rrna_detector.py
import pytest
from app.services.rrna.detector import RRNADetectorService

@pytest.mark.asyncio
async def test_detect_16s_rrna(sample_16s_sequence, get_settings):
    """Test 16S rRNA detection."""
    detector = RRNADetectorService(get_settings())

    results = await detector.detect(sample_16s_sequence)

    assert len(results) > 0
    assert results[0].rrna_type == "16S"
    assert results[0].confidence >= 0.8

3. Integration Test Example

# tests/integration/test_api.py
def test_detect_rrna_endpoint(client, sample_16s_sequence):
    """Test rRNA detection API endpoint."""
    response = client.post(
        "/api/v1/rrna/detect",
        json={
            "sequence": sample_16s_sequence,
            "methods": ["hmm", "pattern"],
            "min_confidence": 0.8
        }
    )

    assert response.status_code == 200
    data = response.json()
    assert "results" in data
    assert len(data["results"]) > 0

---

API Design Conventions

1. RESTful Endpoints

# rRNA Detection
POST   /api/v1/rrna/detect           # Detect rRNA
GET    /api/v1/rrna/types             # List supported types

# Phylogenetics
POST   /api/v1/phylo/align            # Align sequences
POST   /api/v1/phylo/tree             # Build tree
POST   /api/v1/phylo/bootstrap        # Bootstrap analysis
POST   /api/v1/phylo/consensus        # Consensus tree

# Jobs
GET    /api/v1/jobs                   # List jobs
GET    /api/v1/jobs/{id}              # Get job status
DELETE /api/v1/jobs/{id}              # Cancel job

# Health
GET    /health                        # Health check
GET    /metrics                       # Prometheus metrics

2. Response Format

# Success response
{
    "success": true,
    "data": { ... },
    "metadata": {
        "timestamp": "2025-11-20T12:00:00Z",
        "version": "1.0.0"
    }
}

# Error response
{
    "success": false,
    "error": {
        "code": "VALIDATION_ERROR",
        "message": "Invalid sequence format",
        "details": { ... }
    }
}

---

Best Practices

✅ DO

• Use type hints everywhere
• Write docstrings for public APIs
• Use Pydantic for validation
• Implement proper logging
• Write tests for new features
• Use async/await for I/O operations
• Separate concerns (routes, services, models)
• Use dependency injection
• Handle errors gracefully
• Document API with examples

❌ DON'T

• Mix business logic with API routes
• Use synchronous I/O in async functions
• Hardcode configuration values
• Skip input validation
• Ignore exceptions
• Write god classes
• Couple services tightly
• Skip tests
• Use global state
• Block the event loop

---

External Tool Integration Pattern

# app/services/rrna/hmm.py
import subprocess
from pathlib import Path
from typing import Optional

class HMMDetector:
    """HMM-based rRNA detection using HMMER."""

    def __init__(self, profiles_dir: str, hmmsearch_path: str = "hmmsearch"):
        self.profiles_dir = Path(profiles_dir)
        self.hmmsearch_path = hmmsearch_path

    async def detect(self, sequence: str, rrna_type: str) -> dict:
        """Run HMMER search."""

        # Create temp files
        with tempfile.NamedTemporaryFile(mode='w', suffix='.fasta') as seq_file:
            seq_file.write(f">query\n{sequence}\n")
            seq_file.flush()

            profile = self.profiles_dir / f"{rrna_type}.hmm"

            # Run HMMER
            result = subprocess.run(
                [
                    self.hmmsearch_path,
                    "--tblout", "/dev/stdout",
                    str(profile),
                    seq_file.name
                ],
                capture_output=True,
                text=True,
                timeout=300
            )

            if result.returncode != 0:
                raise DetectionError(f"HMMER failed: {result.stderr}")

            # Parse output
            return self._parse_hmmer_output(result.stdout)

---

Related Skills: rRNA-prediction-patterns, ml-integration-patterns

Line Count: < 500 lines ✅