"""FastAPI app serving the trained model to the web front end.

    cd ml
    uvicorn api.main:app --reload --port 8000
"""

import os

from dotenv import load_dotenv
from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse
from pydantic import BaseModel, Field

from api.service import (
    EmptyAfterCleaning,
    analyze,
    describe_model,
    load_metrics,
    plot_path,
)

load_dotenv()


def allowed_origins():
    """Browser origins permitted to call this API."""
    raw = os.getenv("ASA_CORS_ORIGINS", "http://localhost:3000")
    return [origin.strip() for origin in raw.split(",") if origin.strip()]


app = FastAPI(
    title="Arabic Sentiment Analysis API",
    description="Serves the trained TF-IDF + classifier pipeline.",
    version="1.0.0",
)

app.add_middleware(
    CORSMiddleware,
    allow_origins=allowed_origins(),
    allow_credentials=False,
    allow_methods=["GET", "POST"],
    allow_headers=["*"],
)


class PredictRequest(BaseModel):
    """One piece of raw Arabic text to classify."""
    text: str = Field(min_length=1, max_length=5000)
    convert_emojis: bool = True


@app.get("/api/health")
def health():
    try:
        model = describe_model()
    except FileNotFoundError:
        raise HTTPException(
            status_code=503,
            detail="No saved model found. Run `python run_pipeline.py` first.",
        )
    return {"status": "ok", "model": model["name"]}


@app.get("/api/model")
def model_info():
    """What the served pipeline is, and which score types it can produce."""
    try:
        return describe_model()
    except FileNotFoundError:
        raise HTTPException(
            status_code=503,
            detail="No saved model found. Run `python run_pipeline.py` first.",
        )


@app.get("/api/metrics")
def metrics():
    """Test-set scores for all four candidate models, plus the chart manifest."""
    return load_metrics()


@app.get("/api/plots/{name}")
def plot(name: str):
    """Serve one of the pipeline's chart PNGs by file name."""
    path = plot_path(name)
    if path is None:
        raise HTTPException(status_code=404, detail=f"No such chart: {name}")
    return FileResponse(path, media_type="image/png")


@app.post("/api/predict")
def predict(request: PredictRequest):
    """Clean and classify one piece of text."""
    try:
        return analyze(request.text, convert_emojis=request.convert_emojis)
    except EmptyAfterCleaning as error:
        raise HTTPException(status_code=422, detail=str(error))
    except FileNotFoundError:
        raise HTTPException(
            status_code=503,
            detail="No saved model found. Run `python run_pipeline.py` first.",
        )