app.py

"""Banco Ripley – Morosidad Predictor API.

FastAPI service that exposes the trained LightGBM inference pipeline.
Accepts raw financial / demographic features (same schema as data/raw/data.csv)
and returns a binary prediction plus probability of morosidad.

Target convention  →  1 = Malo/Moroso  |  0 = Bueno
"""

import os
import pickle
import time
from pathlib import Path
from typing import Any, Optional

import numpy as np
import pandas as pd
from fastapi import FastAPI, HTTPException
from fastapi.responses import JSONResponse
from pydantic import BaseModel, Field

# ── Paths ─────────────────────────────────────────────────────────────────────
BASE_DIR   = Path(__file__).parent
MODEL_PATH = BASE_DIR / "model.pkl"

# ── Load model once at startup ────────────────────────────────────────────────
print(f"[startup] Loading pipeline from {MODEL_PATH} …")
with open(MODEL_PATH, "rb") as f:
    PIPELINE = pickle.load(f)
print("[startup] Pipeline loaded ✓")

# ── FastAPI app ───────────────────────────────────────────────────────────────
app = FastAPI(
    title="Banco Ripley – Modelo Predictiva Early",
    description=(
        "API de inferencia para el modelo de morosidad temprana de Banco Ripley. "
        "El pipeline acepta datos financieros / demográficos crudos y devuelve "
        "la predicción binaria junto con la probabilidad de morosidad.\n\n"
        "**Convención del target:** `1 = Malo/Moroso` | `0 = Bueno`"
    ),
    version="1.0.0",
    docs_url="/docs",
    redoc_url="/redoc",
)

# ── Schema ────────────────────────────────────────────────────────────────────

class ClientRecord(BaseModel):
    """Raw feature record – misma estructura que data/raw/data.csv.

    Los campos ``serie`` y ``PERIODO`` son **ignorados** por el pipeline
    (el ColumnSelector los descarta antes del modelo); no es necesario enviarlos.
    """

    PERIODO: Optional[Any] = Field(None, description="Ignorado por el pipeline – no requerido")
    serie: Optional[Any] = Field(None, description="Ignorado por el pipeline – no requerido")
    MESES_ANT_RCC: Optional[Any] = None
    EDAD: Optional[Any] = None
    GENERO: Optional[str] = None
    FLAG_ENTIDAD_PRINCIPAL: Optional[Any] = None
    GRADO_INSTRUCCION: Optional[str] = None
    DEPARTAMENTO: Optional[str] = None
    PROVINCIA: Optional[str] = None
    DISTRITO: Optional[str] = None
    FLAG_TC_MODELOS: Optional[Any] = None
    FLAG_MES: Optional[Any] = None
    CONTAR_COMP: Optional[Any] = None
    MARCA_HP: Optional[Any] = None
    MARCA_SEG_VIDA: Optional[Any] = None
    MARCA_DIF: Optional[Any] = None
    MARCA_CONV: Optional[Any] = None
    DMAX_SIN_HP: Optional[Any] = None
    DOTROS_DTOTAL: Optional[Any] = None
    DMES_DTOTAL: Optional[Any] = None
    LINEA_BASE: Optional[Any] = None
    RATIO_CONS: Optional[Any] = None
    UTIL_TARJ: Optional[Any] = None
    UTIL_EFEC: Optional[Any] = None
    UTIL_COMP: Optional[Any] = None
    RATIO_PRES: Optional[Any] = None
    INT_SALDO: Optional[Any] = None
    GARAN_SALDO: Optional[Any] = None
    CONV_PREST: Optional[Any] = None
    CONV_DEUDA: Optional[Any] = None
    MAX_CALIF3: Optional[Any] = None
    MAX_ATRASO3: Optional[Any] = None
    MAX_KTOT_M3: Optional[Any] = None
    MAX_CONV_PREST3: Optional[Any] = None
    PROM_UTIL_TARJ3: Optional[Any] = None
    PROM_UTIL_COMP3: Optional[Any] = None
    PROM_RATIO_PRES3: Optional[Any] = None
    PROM_UTIL_EFEC3: Optional[Any] = None
    MAX_INT_SALDO3: Optional[Any] = None
    MARCA_DIF3: Optional[Any] = None
    MARCA_SEG_VIDA3: Optional[Any] = None
    MARCA_GAR3: Optional[Any] = None
    DIF_BUE_MAL3: Optional[Any] = None
    DIF_BUE_MAL100_3: Optional[Any] = None
    PROM_DOTROS_DTOTAL3: Optional[Any] = None
    MAX_DMES_DTOTAL3: Optional[Any] = None
    PROM_DMES_DTOTAL3: Optional[Any] = None
    MAX_DOTROS_DTOTAL3: Optional[Any] = None
    MAX_LINEA3: Optional[Any] = None
    PROM_LINEA3: Optional[Any] = None
    MAX_CALIF6: Optional[Any] = None
    MAX_KTOT_M6: Optional[Any] = None
    PROM_UTIL_TARJ6: Optional[Any] = None
    PROM_UTIL_COMP6: Optional[Any] = None
    PROM_RATIO_PRES6: Optional[Any] = None
    PROM_UTIL_EFEC6: Optional[Any] = None
    MAX_INT_SALDO6: Optional[Any] = None
    MARCA_DIF6: Optional[Any] = None
    MARCA_SEG_VIDA6: Optional[Any] = None
    MARCA_GAR6: Optional[Any] = None
    MARCA_HIP6: Optional[Any] = None
    DIF_BUE_MAL6: Optional[Any] = None
    DIF_BUE_MAL100_6: Optional[Any] = None
    PROM_DOTROS_DTOTAL6: Optional[Any] = None
    MAX_DMES_DTOTAL6: Optional[Any] = None
    PROM_DMES_DTOTAL6: Optional[Any] = None
    MAX_DOTROS_DTOTAL6: Optional[Any] = None
    MAX_LINEA6: Optional[Any] = None
    PROM_LINEA6: Optional[Any] = None
    MAX_KTOT_M12: Optional[Any] = None
    PROM_UTIL_TARJ12: Optional[Any] = None
    PROM_UTIL_COMP12: Optional[Any] = None
    PROM_RATIO_PRES12: Optional[Any] = None
    PROM_UTIL_EFEC12: Optional[Any] = None
    MAX_INT_SALDO12: Optional[Any] = None
    MARCA_DIF12: Optional[Any] = None
    MARCA_SEG_VIDA12: Optional[Any] = None
    MARCA_GAR12: Optional[Any] = None
    MARCA_HIP12: Optional[Any] = None
    DIF_BUE_MAL12: Optional[Any] = None
    DIF_BUE_MAL100_12: Optional[Any] = None
    PROM_DOTROS_DTOTAL12: Optional[Any] = None
    MAX_DMES_DTOTAL12: Optional[Any] = None
    PROM_DMES_DTOTAL12: Optional[Any] = None
    MAX_DOTROS_DTOTAL12: Optional[Any] = None
    MAX_LINEA12: Optional[Any] = None
    PROM_LINEA12: Optional[Any] = None
    D12_MAXD: Optional[Any] = None
    D12_MAXD6: Optional[Any] = None
    D12_MAXD3: Optional[Any] = None
    PROM3_PROM12: Optional[Any] = None
    PROM3_PROM6: Optional[Any] = None
    PROM6_PROM12: Optional[Any] = None
    Max_AumKP: Optional[Any] = None
    Max_AumMORA: Optional[Any] = None
    Max_AumKT: Optional[Any] = None
    Max_DismDTOTAL: Optional[Any] = None
    Max_AumMES: Optional[Any] = None
    Max_DismMES: Optional[Any] = None
    Max_DismCONS: Optional[Any] = None
    Max_AumCONS: Optional[Any] = None
    Max_AumDTOTAL: Optional[Any] = None
    Max_DismKP: Optional[Any] = None
    Max_DismKT: Optional[Any] = None
    NMES_UMORA: Optional[Any] = None
    PROMDIR_IND12: Optional[Any] = None
    PROMDIR_IND6: Optional[Any] = None
    PROMDIR_IND3: Optional[Any] = None
    RATIO_CONS12: Optional[Any] = None
    LINT_PROM12: Optional[Any] = None
    MAX_PORC_ACR12: Optional[Any] = None
    MAX_CALIF12: Optional[Any] = None
    MARCA_CONV3: Optional[Any] = None
    MAX_CONV_DEUDA3: Optional[Any] = None
    MAX_ENT_REP3: Optional[Any] = None
    MARCA_HIP3: Optional[Any] = None
    MAX_PORC_ACR3: Optional[Any] = None
    MAX_RATIO_PRES3: Optional[Any] = None
    MAX_UTIL_COMP3: Optional[Any] = None
    MAX_UTIL_EFEC3: Optional[Any] = None
    MAX_UTIL_TARJ3: Optional[Any] = None
    RATIOS_PROM_COMP_DEUDA3: Optional[Any] = None
    RATIOS_PROM_EFEC_DEUDA3: Optional[Any] = None
    MAX_ENT_ACR3: Optional[Any] = None
    MAX_GARAN_SALDO3: Optional[Any] = None
    MARCA_CONV6: Optional[Any] = None
    MAX_CONV_DEUDA6: Optional[Any] = None
    MAX_ENT_REP6: Optional[Any] = None
    MAX_PORC_ACR6: Optional[Any] = None
    MAX_RATIO_PRES6: Optional[Any] = None
    MAX_UTIL_COMP6: Optional[Any] = None
    MAX_UTIL_EFEC6: Optional[Any] = None
    MAX_UTIL_TARJ6: Optional[Any] = None
    RATIOS_PROM_COMP_DEUDA6: Optional[Any] = None
    RATIOS_PROM_EFEC_DEUDA6: Optional[Any] = None
    MAX_ENT_ACR6: Optional[Any] = None
    MAX_GARAN_SALDO6: Optional[Any] = None
    MAX_ATRASO6: Optional[Any] = None
    MAX_CONV_PREST6: Optional[Any] = None
    MARCA_CONV12: Optional[Any] = None
    MAX_CONV_DEUDA12: Optional[Any] = None
    MAX_ENT_REP12: Optional[Any] = None
    MAX_RATIO_PRES12: Optional[Any] = None
    MAX_UTIL_COMP12: Optional[Any] = None
    MAX_UTIL_EFEC12: Optional[Any] = None
    MAX_UTIL_TARJ12: Optional[Any] = None
    RATIOS_PROM_COMP_DEUDA12: Optional[Any] = None
    RATIOS_PROM_EFEC_DEUDA12: Optional[Any] = None
    MAX_ENT_ACR12: Optional[Any] = None
    MAX_GARAN_SALDO12: Optional[Any] = None
    MAX_ATRASO12: Optional[Any] = None
    MAX_CONV_PREST12: Optional[Any] = None
    MARCA_GAR: Optional[Any] = None
    FLAG_TENENCIA_VEHICULAR: Optional[Any] = None
    SITUACION_LABORAL: Optional[str] = None
    ESTADO_CIVIL: Optional[str] = None

    class Config:
        extra = "allow"


class PredictRequest(BaseModel):
    records: list[ClientRecord]


class PredictionResult(BaseModel):
    prediction: int         = Field(..., description="0=Bueno | 1=Malo/Moroso")
    probability_malo: float = Field(..., description="Probabilidad de morosidad (clase 1)")
    label: str              = Field(..., description="Etiqueta legible: Bueno | Malo")


class PredictResponse(BaseModel):
    predictions: list[PredictionResult]
    model_version: str = "lgbm-inference-pipeline-v1"
    run_id: str = "8967c637226f47209eeb1dba83e7519e"


# ── Helpers ───────────────────────────────────────────────────────────────────

def _predict(records: list[dict]) -> list[PredictionResult]:
    df = pd.DataFrame(records)
    probas  = PIPELINE.predict_proba(df)[:, 1]
    preds   = (probas >= 0.5).astype(int)
    return [
        PredictionResult(
            prediction=int(p),
            probability_malo=round(float(pr), 6),
            label="Malo" if p == 1 else "Bueno",
        )
        for p, pr in zip(preds, probas)
    ]


# ── Endpoints ─────────────────────────────────────────────────────────────────

@app.get("/", tags=["Info"])
def root():
    return {
        "service": "predictiva_early",
        "description": "Modelo de morosidad temprana – Banco Ripley",
        "target_convention": {"0": "Bueno", "1": "Malo/Moroso"},
        "docs": "/docs",
    }


@app.get("/health", tags=["Info"])
def health():
    return {"status": "ok"}


@app.post("/predict", response_model=PredictResponse, tags=["Inference"])
def predict_single(record: ClientRecord):
    """Predice morosidad para **un único cliente**.

    Envía un registro con la misma estructura que `data/raw/data.csv`.
    Devuelve la clase predicha (`0=Bueno`, `1=Malo/Moroso`) y la probabilidad.
    """
    try:
        results = _predict([record.model_dump()])
    except Exception as exc:
        raise HTTPException(status_code=422, detail=str(exc))
    return PredictResponse(predictions=results)


@app.post("/predict/batch", response_model=PredictResponse, tags=["Inference"])
def predict_batch(request: PredictRequest):
    """Predice morosidad para un **lote de clientes** (máximo 500 registros).

    Envía una lista de registros bajo la clave `records`.
    """
    if len(request.records) > 500:
        raise HTTPException(
            status_code=413,
            detail="Batch demasiado grande. Máximo 500 registros por llamada.",
        )
    try:
        results = _predict([r.model_dump() for r in request.records])
    except Exception as exc:
        raise HTTPException(status_code=422, detail=str(exc))
    return PredictResponse(predictions=results)