egutierrez
268a76602a
Funciones Python para interactuar con Jupyter Lab programáticamente: descubrir instancias, leer/escribir celdas, ejecutar código y gestionar kernels. Reemplazan MCP jupyter con API REST + WebSocket directa. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
137 lines
4.6 KiB
Markdown
137 lines
4.6 KiB
Markdown
---
|
|
name: jupyter_kernel
|
|
kind: function
|
|
lang: py
|
|
domain: notebook
|
|
version: "1.0.0"
|
|
purity: impure
|
|
signature: "def jupyter_kernel_list(server_url: str = \"http://localhost:8888\", token: str = \"\") -> list[dict]"
|
|
description: "CRUD completo de kernels Jupyter via REST API. Expone seis operaciones: list, start, restart, interrupt, shutdown y sessions. Usa solo stdlib (urllib, json), sin dependencias externas."
|
|
tags: [jupyter, notebook, kernel, api, http, rest, sessions, crud]
|
|
uses_functions: []
|
|
uses_types: []
|
|
returns: []
|
|
returns_optional: false
|
|
error_type: "error_go_core"
|
|
imports: [json, urllib.error, urllib.request]
|
|
tested: false
|
|
tests: []
|
|
test_file_path: ""
|
|
file_path: "python/functions/notebook/jupyter_kernel.py"
|
|
---
|
|
|
|
## Funciones expuestas
|
|
|
|
| Funcion | Endpoint | Descripcion |
|
|
|---------|----------|-------------|
|
|
| `jupyter_kernel_list(server_url, token)` | `GET /api/kernels` | Lista kernels activos |
|
|
| `jupyter_kernel_start(server_url, token, name)` | `POST /api/kernels` | Inicia un kernel nuevo |
|
|
| `jupyter_kernel_restart(server_url, token, kernel_id)` | `POST /api/kernels/{id}/restart` | Reinicia un kernel |
|
|
| `jupyter_kernel_interrupt(server_url, token, kernel_id)` | `POST /api/kernels/{id}/interrupt` | Interrumpe ejecucion |
|
|
| `jupyter_kernel_shutdown(server_url, token, kernel_id)` | `DELETE /api/kernels/{id}` | Apaga y elimina un kernel |
|
|
| `jupyter_kernel_sessions(server_url, token)` | `GET /api/sessions` | Lista sesiones activas |
|
|
|
|
## Ejemplo
|
|
|
|
```python
|
|
from notebook.jupyter_kernel import (
|
|
jupyter_kernel_list,
|
|
jupyter_kernel_start,
|
|
jupyter_kernel_restart,
|
|
jupyter_kernel_interrupt,
|
|
jupyter_kernel_shutdown,
|
|
jupyter_kernel_sessions,
|
|
)
|
|
|
|
# Listar kernels activos
|
|
kernels = jupyter_kernel_list("http://localhost:8888", token="mi_token")
|
|
# [{"id": "abc123", "name": "python3", "execution_state": "idle",
|
|
# "last_activity": "2026-04-01T10:00:00.000Z", "connections": 1}]
|
|
|
|
# Iniciar un kernel nuevo
|
|
kernel = jupyter_kernel_start("http://localhost:8888", token="mi_token", name="python3")
|
|
# {"id": "def456", "name": "python3", "execution_state": "starting"}
|
|
|
|
# Reiniciar
|
|
jupyter_kernel_restart("http://localhost:8888", "mi_token", kernel["id"])
|
|
|
|
# Interrumpir ejecucion en curso
|
|
jupyter_kernel_interrupt("http://localhost:8888", "mi_token", kernel["id"])
|
|
|
|
# Apagar
|
|
jupyter_kernel_shutdown("http://localhost:8888", "mi_token", kernel["id"])
|
|
|
|
# Listar sesiones (mapeo notebook <-> kernel)
|
|
sessions = jupyter_kernel_sessions("http://localhost:8888", token="mi_token")
|
|
# [{"id": "s1", "notebook": "notebooks/01_analisis.ipynb",
|
|
# "kernel_id": "abc123", "kernel_state": "idle", "type": "notebook", "name": "01_analisis"}]
|
|
```
|
|
|
|
## CLI
|
|
|
|
```bash
|
|
# Listar kernels
|
|
python python/functions/notebook/jupyter_kernel.py list --server http://localhost:8888 --token TK
|
|
|
|
# Iniciar kernel (por defecto python3)
|
|
python python/functions/notebook/jupyter_kernel.py start --name python3
|
|
|
|
# Reiniciar
|
|
python python/functions/notebook/jupyter_kernel.py restart abc123-...
|
|
|
|
# Interrumpir
|
|
python python/functions/notebook/jupyter_kernel.py interrupt abc123-...
|
|
|
|
# Apagar
|
|
python python/functions/notebook/jupyter_kernel.py shutdown abc123-...
|
|
|
|
# Listar sesiones
|
|
python python/functions/notebook/jupyter_kernel.py sessions
|
|
```
|
|
|
|
Todos los subcomandos aceptan `--server` y `--token`. El output es siempre JSON.
|
|
|
|
## Estructura de los dicts retornados
|
|
|
|
**`jupyter_kernel_list`** — cada elemento:
|
|
```python
|
|
{
|
|
"id": "abc123-...",
|
|
"name": "python3",
|
|
"execution_state": "idle", # idle | busy | starting
|
|
"last_activity": "2026-04-01T10:00:00.000Z",
|
|
"connections": 1
|
|
}
|
|
```
|
|
|
|
**`jupyter_kernel_start`** — elemento retornado:
|
|
```python
|
|
{
|
|
"id": "def456-...",
|
|
"name": "python3",
|
|
"execution_state": "starting"
|
|
}
|
|
```
|
|
|
|
**`jupyter_kernel_sessions`** — cada elemento:
|
|
```python
|
|
{
|
|
"id": "session-id",
|
|
"notebook": "notebooks/01_analisis.ipynb", # path relativo al root del servidor
|
|
"kernel_id": "abc123-...",
|
|
"kernel_state": "idle",
|
|
"type": "notebook",
|
|
"name": "01_analisis"
|
|
}
|
|
```
|
|
|
|
## Notas
|
|
|
|
Solo usa stdlib: `urllib.request`, `urllib.error`, `json`. No requiere `requests`, `jupyter_client` ni ningun paquete externo.
|
|
|
|
La funcion interna `_make_request` centraliza la construccion de headers y serializacion JSON. El header `Authorization: token {token}` se omite si `token` esta vacio, permitiendo conectar a servidores sin autenticacion.
|
|
|
|
`jupyter_kernel_interrupt` y `jupyter_kernel_shutdown` retornan `None` porque la API de Jupyter devuelve 204 No Content en esos casos.
|
|
|
|
Compatible con Jupyter Lab 3.x, 4.x y Jupyter Notebook 6.x/7.x — todos exponen la misma REST API en `/api/kernels` y `/api/sessions`.
|