Documentation Index
Fetch the complete documentation index at: https://mintlify.com/JoseOlivares19/Proyecto-PC3-JavaScript-Avanzado/llms.txt
Use this file to discover all available pages before exploring further.
The SmartStock360 AI service is a self-contained Python application. All five classifiers train from synthetic data every time the server starts, so there are no model files to download or manage — just install the dependencies and run.
Prerequisites
- Python 3.9 or later — verify with
python --version
- pip — verify with
pip --version
Dependencies
The following pinned versions are required (from requirements.txt):
fastapi==0.115.0
uvicorn[standard]==0.30.6
scikit-learn==1.5.2
numpy==2.1.1
pydantic==2.9.2
Setup Steps
Navigate to the service directory
Open a terminal and change into the AI service folder:cd update/Kit_Python_Modelos_IA_JavaScript_Avanzado_S15_UTP
Create a virtual environment
Isolate the project dependencies from your global Python installation: Activate the virtual environment
The activation command differs by operating system:Windows:macOS / Linux:Your terminal prompt should now be prefixed with (venv). Install dependencies
Install all pinned packages from the requirements file:pip install -r requirements.txt
Start the server
Launch Uvicorn with hot-reload enabled on port 8001:uvicorn app:app --reload --port 8001
Application startup complete. Verify the service is healthy
In a second terminal (with the virtual environment activated), run:curl http://localhost:8001/health
{
"status": "ok",
"modelos_cargados": [
"utp-risk",
"fraud-shield",
"cyber-sentinel",
"smart-stock",
"talent-match"
]
}
Startup Training Code
All five models are registered and trained inside a single startup event handler. The handler is called once, synchronously, before the server accepts any requests:
@app.on_event("startup")
def startup_train_models():
X, y = generate_utp_risk_data()
MODELS["utp-risk"] = train_pack(
["promedio_actual", "asistencia_pct", "tareas_entregadas_pct",
"participacion_pct", "horas_estudio_semana", "nota_pc_anterior"], X, y
)
X, y = generate_fraud_data()
MODELS["fraud-shield"] = train_pack(
["monto", "hora_24", "intentos_previos", "antiguedad_cliente_meses",
"dispositivo_nuevo", "pais_riesgo", "compras_ultima_hora"], X, y
)
X, y = generate_cyber_data()
MODELS["cyber-sentinel"] = train_pack(
["intentos_login_fallidos", "puertos_abiertos", "vulnerabilidades_criticas",
"trafico_anomalo_pct", "equipos_afectados", "parcheado_pct"], X, y
)
X, y = generate_stock_data()
MODELS["smart-stock"] = train_pack(
["precio", "stock_actual", "ventas_7d", "descuento_pct",
"temporada", "dias_sin_reabastecer", "rating_producto"], X, y
)
X, y = generate_talent_data()
MODELS["talent-match"] = train_pack(
["javascript", "react", "spring_boot", "python_datos",
"sql", "experiencia_proyectos", "preferencia"], X, y
)
train_pack() call fits a LabelEncoder on the label array, then trains a RandomForestClassifier(n_estimators=180, max_depth=7, random_state=42, class_weight="balanced_subsample") and wraps both inside a ModelPack.
All five models train synchronously at startup before any request is served. Training typically completes in a few seconds on a standard development machine. Do not send prediction requests until Uvicorn logs Application startup complete.
Interactive API Documentation
Once the server is running, browse the auto-generated Swagger UI at:
http://localhost:8001/docs
The /metadata endpoint returns the feature list and label list for every loaded model — useful for programmatically validating request payloads before sending them.