Spaces:
Sleeping
Sleeping
documentacion añadida
Browse files- api_ner.py +258 -68
api_ner.py
CHANGED
|
@@ -4,66 +4,77 @@
|
|
| 4 |
FastAPI endpoint para el proyecto lanc-ner
|
| 5 |
Recibe un archivo .txt y devuelve las anotaciones NER en formato JSON.
|
| 6 |
|
| 7 |
-
|
| 8 |
-
uvicorn api_ner:app --host 0.0.0.0 --port 8000 --reload
|
| 9 |
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
- Respuesta: JSON con anotaciones NER
|
| 14 |
"""
|
| 15 |
|
| 16 |
-
from fastapi import FastAPI, File, UploadFile, HTTPException
|
| 17 |
-
from fastapi.responses import JSONResponse
|
|
|
|
|
|
|
| 18 |
import spacy
|
| 19 |
from spacy import displacy
|
| 20 |
import os
|
| 21 |
import logging
|
| 22 |
-
from typing import Optional
|
| 23 |
|
| 24 |
-
# ── Logging ──────────────────────────────────────────────────────────────────
|
| 25 |
logging.basicConfig(level=logging.INFO)
|
| 26 |
logger = logging.getLogger(__name__)
|
| 27 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
# ── App ───────────────────────────────────────────────────────────────────────
|
| 29 |
app = FastAPI(
|
| 30 |
title="Lancaster NER API",
|
| 31 |
-
description=
|
| 32 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
)
|
| 34 |
|
| 35 |
-
# ──
|
| 36 |
-
# Ajusta el nombre del modelo al que uses en tu proyecto (lanc-ner).
|
| 37 |
-
# Si tienes un modelo entrenado propio en ./models/, cámbialo aquí:
|
| 38 |
-
# NLP_MODEL_PATH = "./models/mi_modelo"
|
| 39 |
-
NLP_MODEL_PATH = os.getenv("NER_MODEL", "en_core_web_sm")
|
| 40 |
-
|
| 41 |
try:
|
| 42 |
nlp = spacy.load(NLP_MODEL_PATH)
|
| 43 |
logger.info(f"Modelo cargado: {NLP_MODEL_PATH}")
|
| 44 |
except OSError:
|
| 45 |
-
logger.error(
|
| 46 |
-
f"No se encontró el modelo '{NLP_MODEL_PATH}'. "
|
| 47 |
-
"Ejecuta: python -m spacy download en_core_web_sm"
|
| 48 |
-
)
|
| 49 |
nlp = None
|
| 50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
|
| 52 |
-
|
|
|
|
| 53 |
def run_ner(text: str) -> dict:
|
| 54 |
-
"""
|
| 55 |
-
Aplica el pipeline NER de spaCy sobre el texto y devuelve
|
| 56 |
-
un diccionario con:
|
| 57 |
-
- entities : lista de entidades detectadas
|
| 58 |
-
- tokens : lista de todos los tokens con sus etiquetas IOB
|
| 59 |
-
- html : visualización HTML (displaCy) — opcional
|
| 60 |
-
"""
|
| 61 |
if nlp is None:
|
| 62 |
raise RuntimeError("El modelo NLP no está disponible.")
|
| 63 |
|
| 64 |
doc = nlp(text)
|
| 65 |
|
| 66 |
-
# Entidades detectadas
|
| 67 |
entities = [
|
| 68 |
{
|
| 69 |
"text": ent.text,
|
|
@@ -75,7 +86,6 @@ def run_ner(text: str) -> dict:
|
|
| 75 |
for ent in doc.ents
|
| 76 |
]
|
| 77 |
|
| 78 |
-
# Tokens con etiquetas IOB
|
| 79 |
tokens = [
|
| 80 |
{
|
| 81 |
"token": token.text,
|
|
@@ -88,7 +98,6 @@ def run_ner(text: str) -> dict:
|
|
| 88 |
for token in doc
|
| 89 |
]
|
| 90 |
|
| 91 |
-
# Visualización HTML de displaCy (útil para depuración/front)
|
| 92 |
html = displacy.render(doc, style="ent", page=False, minify=True)
|
| 93 |
|
| 94 |
return {
|
|
@@ -101,71 +110,98 @@ def run_ner(text: str) -> dict:
|
|
| 101 |
|
| 102 |
# ── Endpoints ─────────────────────────────────────────────────────────────────
|
| 103 |
|
| 104 |
-
@app.get(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
def health_check():
|
| 106 |
-
"""
|
| 107 |
-
|
| 108 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
|
| 110 |
|
| 111 |
-
@app.post(
|
| 112 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 113 |
"""
|
| 114 |
-
Recibe un archivo
|
| 115 |
-
NER producidas por el modelo spaCy configurado en el proyecto.
|
| 116 |
|
| 117 |
-
**
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 118 |
```json
|
| 119 |
{
|
| 120 |
"filename": "documento.txt",
|
| 121 |
-
"num_entities":
|
| 122 |
"entities": [
|
| 123 |
-
{
|
| 124 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 125 |
],
|
| 126 |
"tokens": [...],
|
| 127 |
"html_visualization": "<div>...</div>"
|
| 128 |
}
|
| 129 |
```
|
| 130 |
"""
|
| 131 |
-
# Validación del tipo de archivo
|
| 132 |
if not file.filename.lower().endswith(".txt"):
|
| 133 |
-
raise HTTPException(
|
| 134 |
-
status_code=415,
|
| 135 |
-
detail="Solo se aceptan archivos .txt",
|
| 136 |
-
)
|
| 137 |
|
| 138 |
-
# Lectura del contenido
|
| 139 |
try:
|
| 140 |
-
|
| 141 |
-
text = content_bytes.decode("utf-8")
|
| 142 |
except UnicodeDecodeError:
|
| 143 |
-
raise HTTPException(
|
| 144 |
-
status_code=422,
|
| 145 |
-
detail="No se pudo decodificar el archivo como UTF-8.",
|
| 146 |
-
)
|
| 147 |
|
| 148 |
if not text.strip():
|
| 149 |
-
raise HTTPException(
|
| 150 |
-
status_code=422,
|
| 151 |
-
detail="El archivo está vacío.",
|
| 152 |
-
)
|
| 153 |
|
| 154 |
-
# NER
|
| 155 |
try:
|
| 156 |
result = run_ner(text)
|
| 157 |
except RuntimeError as e:
|
| 158 |
raise HTTPException(status_code=503, detail=str(e))
|
| 159 |
|
| 160 |
-
return JSONResponse(
|
| 161 |
-
content={"filename": file.filename, **result}
|
| 162 |
-
)
|
| 163 |
|
| 164 |
|
| 165 |
-
@app.post(
|
| 166 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 167 |
"""
|
| 168 |
-
Alternativa: envía el texto directamente
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 169 |
|
| 170 |
**Body:**
|
| 171 |
```json
|
|
@@ -182,3 +218,157 @@ async def annotate_text(payload: dict):
|
|
| 182 |
raise HTTPException(status_code=503, detail=str(e))
|
| 183 |
|
| 184 |
return JSONResponse(content=result)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
FastAPI endpoint para el proyecto lanc-ner
|
| 5 |
Recibe un archivo .txt y devuelve las anotaciones NER en formato JSON.
|
| 6 |
|
| 7 |
+
Autenticación: Bearer token via variable de entorno API_TOKEN.
|
|
|
|
| 8 |
|
| 9 |
+
Uso:
|
| 10 |
+
export API_TOKEN="mi-token-secreto"
|
| 11 |
+
uvicorn api_ner:app --host 0.0.0.0 --port 7860 --reload
|
|
|
|
| 12 |
"""
|
| 13 |
|
| 14 |
+
from fastapi import FastAPI, File, UploadFile, HTTPException, Depends, Security
|
| 15 |
+
from fastapi.responses import JSONResponse, HTMLResponse
|
| 16 |
+
from fastapi.security import HTTPBearer, HTTPAuthorizationCredentials
|
| 17 |
+
from fastapi.openapi.utils import get_openapi
|
| 18 |
import spacy
|
| 19 |
from spacy import displacy
|
| 20 |
import os
|
| 21 |
import logging
|
|
|
|
| 22 |
|
| 23 |
+
# ── Logging ───────────────────────────────────────────────────────────────────
|
| 24 |
logging.basicConfig(level=logging.INFO)
|
| 25 |
logger = logging.getLogger(__name__)
|
| 26 |
|
| 27 |
+
# ── Configuración ─────────────────────────────────────────────────────────────
|
| 28 |
+
NLP_MODEL_PATH = os.getenv("NER_MODEL", "en_core_web_sm")
|
| 29 |
+
API_TOKEN = os.getenv("API_TOKEN", "") # obligatorio en producción
|
| 30 |
+
|
| 31 |
+
if not API_TOKEN:
|
| 32 |
+
logger.warning("⚠️ API_TOKEN no definido — la API está desprotegida.")
|
| 33 |
+
|
| 34 |
# ── App ───────────────────────────────────────────────────────────────────────
|
| 35 |
app = FastAPI(
|
| 36 |
title="Lancaster NER API",
|
| 37 |
+
description=(
|
| 38 |
+
"API REST para detección de entidades nombradas (NER) basada en spaCy.\n\n"
|
| 39 |
+
"**Autenticación:** todos los endpoints protegidos requieren un "
|
| 40 |
+
"`Authorization: Bearer <token>` en la cabecera.\n\n"
|
| 41 |
+
"Consulta `/docs` para la documentación interactiva (Swagger UI) "
|
| 42 |
+
"o `/redoc` para la versión ReDoc."
|
| 43 |
+
),
|
| 44 |
+
version="1.1.0",
|
| 45 |
+
docs_url="/docs",
|
| 46 |
+
redoc_url="/redoc",
|
| 47 |
)
|
| 48 |
|
| 49 |
+
# ── Modelo spaCy ──────────────────────────────────────────────────────────────
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
try:
|
| 51 |
nlp = spacy.load(NLP_MODEL_PATH)
|
| 52 |
logger.info(f"Modelo cargado: {NLP_MODEL_PATH}")
|
| 53 |
except OSError:
|
| 54 |
+
logger.error(f"Modelo '{NLP_MODEL_PATH}' no encontrado.")
|
|
|
|
|
|
|
|
|
|
| 55 |
nlp = None
|
| 56 |
|
| 57 |
+
# ── Seguridad: Bearer token ───────────────────────────────────────────────────
|
| 58 |
+
bearer_scheme = HTTPBearer(auto_error=True)
|
| 59 |
+
|
| 60 |
+
def verify_token(credentials: HTTPAuthorizationCredentials = Security(bearer_scheme)):
|
| 61 |
+
"""Valida el Bearer token contra la variable de entorno API_TOKEN."""
|
| 62 |
+
if API_TOKEN and credentials.credentials != API_TOKEN:
|
| 63 |
+
raise HTTPException(
|
| 64 |
+
status_code=401,
|
| 65 |
+
detail="Token inválido o expirado.",
|
| 66 |
+
headers={"WWW-Authenticate": "Bearer"},
|
| 67 |
+
)
|
| 68 |
+
return credentials.credentials
|
| 69 |
|
| 70 |
+
|
| 71 |
+
# ── Función NER ───────────────────────────────────────────────────────────────
|
| 72 |
def run_ner(text: str) -> dict:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
if nlp is None:
|
| 74 |
raise RuntimeError("El modelo NLP no está disponible.")
|
| 75 |
|
| 76 |
doc = nlp(text)
|
| 77 |
|
|
|
|
| 78 |
entities = [
|
| 79 |
{
|
| 80 |
"text": ent.text,
|
|
|
|
| 86 |
for ent in doc.ents
|
| 87 |
]
|
| 88 |
|
|
|
|
| 89 |
tokens = [
|
| 90 |
{
|
| 91 |
"token": token.text,
|
|
|
|
| 98 |
for token in doc
|
| 99 |
]
|
| 100 |
|
|
|
|
| 101 |
html = displacy.render(doc, style="ent", page=False, minify=True)
|
| 102 |
|
| 103 |
return {
|
|
|
|
| 110 |
|
| 111 |
# ── Endpoints ─────────────────────────────────────────────────────────────────
|
| 112 |
|
| 113 |
+
@app.get(
|
| 114 |
+
"/",
|
| 115 |
+
tags=["Health"],
|
| 116 |
+
summary="Health check",
|
| 117 |
+
response_description="Estado de la API y del modelo cargado.",
|
| 118 |
+
)
|
| 119 |
def health_check():
|
| 120 |
+
"""
|
| 121 |
+
Comprueba que la API está en marcha.
|
| 122 |
+
**No requiere autenticación.**
|
| 123 |
+
"""
|
| 124 |
+
return {
|
| 125 |
+
"status": "ok",
|
| 126 |
+
"model": NLP_MODEL_PATH,
|
| 127 |
+
"model_status": "loaded" if nlp is not None else "not loaded",
|
| 128 |
+
"auth_enabled": bool(API_TOKEN),
|
| 129 |
+
}
|
| 130 |
|
| 131 |
|
| 132 |
+
@app.post(
|
| 133 |
+
"/annotate",
|
| 134 |
+
tags=["NER"],
|
| 135 |
+
summary="Anotar documento .txt",
|
| 136 |
+
response_description="Entidades NER detectadas, tokens y visualización HTML.",
|
| 137 |
+
)
|
| 138 |
+
async def annotate(
|
| 139 |
+
file: UploadFile = File(..., description="Archivo de texto plano (.txt) a anotar."),
|
| 140 |
+
token: str = Depends(verify_token),
|
| 141 |
+
):
|
| 142 |
"""
|
| 143 |
+
Recibe un archivo `.txt` y devuelve sus anotaciones NER.
|
|
|
|
| 144 |
|
| 145 |
+
**Cabecera requerida:**
|
| 146 |
+
```
|
| 147 |
+
Authorization: Bearer <tu-token>
|
| 148 |
+
```
|
| 149 |
+
|
| 150 |
+
**Respuesta:**
|
| 151 |
```json
|
| 152 |
{
|
| 153 |
"filename": "documento.txt",
|
| 154 |
+
"num_entities": 2,
|
| 155 |
"entities": [
|
| 156 |
+
{
|
| 157 |
+
"text": "London",
|
| 158 |
+
"label": "GPE",
|
| 159 |
+
"label_description": "Countries, cities, states",
|
| 160 |
+
"start": 5,
|
| 161 |
+
"end": 11
|
| 162 |
+
}
|
| 163 |
],
|
| 164 |
"tokens": [...],
|
| 165 |
"html_visualization": "<div>...</div>"
|
| 166 |
}
|
| 167 |
```
|
| 168 |
"""
|
|
|
|
| 169 |
if not file.filename.lower().endswith(".txt"):
|
| 170 |
+
raise HTTPException(status_code=415, detail="Solo se aceptan archivos .txt.")
|
|
|
|
|
|
|
|
|
|
| 171 |
|
|
|
|
| 172 |
try:
|
| 173 |
+
text = (await file.read()).decode("utf-8")
|
|
|
|
| 174 |
except UnicodeDecodeError:
|
| 175 |
+
raise HTTPException(status_code=422, detail="El archivo no es UTF-8 válido.")
|
|
|
|
|
|
|
|
|
|
| 176 |
|
| 177 |
if not text.strip():
|
| 178 |
+
raise HTTPException(status_code=422, detail="El archivo está vacío.")
|
|
|
|
|
|
|
|
|
|
| 179 |
|
|
|
|
| 180 |
try:
|
| 181 |
result = run_ner(text)
|
| 182 |
except RuntimeError as e:
|
| 183 |
raise HTTPException(status_code=503, detail=str(e))
|
| 184 |
|
| 185 |
+
return JSONResponse(content={"filename": file.filename, **result})
|
|
|
|
|
|
|
| 186 |
|
| 187 |
|
| 188 |
+
@app.post(
|
| 189 |
+
"/annotate/text",
|
| 190 |
+
tags=["NER"],
|
| 191 |
+
summary="Anotar texto plano (JSON)",
|
| 192 |
+
response_description="Entidades NER detectadas, tokens y visualización HTML.",
|
| 193 |
+
)
|
| 194 |
+
async def annotate_text(
|
| 195 |
+
payload: dict,
|
| 196 |
+
token: str = Depends(verify_token),
|
| 197 |
+
):
|
| 198 |
"""
|
| 199 |
+
Alternativa a `/annotate`: envía el texto directamente en el body JSON.
|
| 200 |
+
|
| 201 |
+
**Cabecera requerida:**
|
| 202 |
+
```
|
| 203 |
+
Authorization: Bearer <tu-token>
|
| 204 |
+
```
|
| 205 |
|
| 206 |
**Body:**
|
| 207 |
```json
|
|
|
|
| 218 |
raise HTTPException(status_code=503, detail=str(e))
|
| 219 |
|
| 220 |
return JSONResponse(content=result)
|
| 221 |
+
|
| 222 |
+
|
| 223 |
+
@app.get(
|
| 224 |
+
"/docu",
|
| 225 |
+
tags=["Documentación"],
|
| 226 |
+
summary="Documentación HTML del servicio",
|
| 227 |
+
response_class=HTMLResponse,
|
| 228 |
+
include_in_schema=False, # no aparece en /docs para no duplicar
|
| 229 |
+
)
|
| 230 |
+
def docu():
|
| 231 |
+
"""
|
| 232 |
+
Página HTML autodocumentada con la descripción completa del servicio,
|
| 233 |
+
sus endpoints, parámetros y ejemplos de uso.
|
| 234 |
+
**No requiere autenticación.**
|
| 235 |
+
"""
|
| 236 |
+
schema = get_openapi(
|
| 237 |
+
title=app.title,
|
| 238 |
+
version=app.version,
|
| 239 |
+
description=app.description,
|
| 240 |
+
routes=app.routes,
|
| 241 |
+
)
|
| 242 |
+
|
| 243 |
+
endpoints_html = ""
|
| 244 |
+
for path, methods in schema.get("paths", {}).items():
|
| 245 |
+
for method, info in methods.items():
|
| 246 |
+
badge_color = {
|
| 247 |
+
"get": "#61affe", "post": "#49cc90",
|
| 248 |
+
"put": "#fca130", "delete": "#f93e3e",
|
| 249 |
+
}.get(method.lower(), "#888")
|
| 250 |
+
|
| 251 |
+
params_html = ""
|
| 252 |
+
for p in info.get("parameters", []):
|
| 253 |
+
params_html += f"""
|
| 254 |
+
<tr>
|
| 255 |
+
<td><code>{p.get('name')}</code></td>
|
| 256 |
+
<td>{p.get('in')}</td>
|
| 257 |
+
<td>{'✅' if p.get('required') else '—'}</td>
|
| 258 |
+
<td>{p.get('description','')}</td>
|
| 259 |
+
</tr>"""
|
| 260 |
+
|
| 261 |
+
params_section = (
|
| 262 |
+
f"""<table class="param-table">
|
| 263 |
+
<thead><tr><th>Nombre</th><th>En</th><th>Req.</th><th>Descripción</th></tr></thead>
|
| 264 |
+
<tbody>{params_html}</tbody>
|
| 265 |
+
</table>"""
|
| 266 |
+
if params_html else ""
|
| 267 |
+
)
|
| 268 |
+
|
| 269 |
+
endpoints_html += f"""
|
| 270 |
+
<div class="endpoint">
|
| 271 |
+
<div class="ep-header">
|
| 272 |
+
<span class="badge" style="background:{badge_color}">{method.upper()}</span>
|
| 273 |
+
<span class="ep-path">{path}</span>
|
| 274 |
+
<span class="ep-summary">{info.get('summary','')}</span>
|
| 275 |
+
</div>
|
| 276 |
+
<p class="ep-desc">{info.get('description','').strip().split(chr(10))[0]}</p>
|
| 277 |
+
{params_section}
|
| 278 |
+
</div>"""
|
| 279 |
+
|
| 280 |
+
html = f"""<!DOCTYPE html>
|
| 281 |
+
<html lang="es">
|
| 282 |
+
<head>
|
| 283 |
+
<meta charset="UTF-8"/>
|
| 284 |
+
<meta name="viewport" content="width=device-width, initial-scale=1.0"/>
|
| 285 |
+
<title>{app.title} — Documentación</title>
|
| 286 |
+
<style>
|
| 287 |
+
*, *::before, *::after {{ box-sizing: border-box; margin: 0; padding: 0; }}
|
| 288 |
+
body {{ font-family: 'Segoe UI', system-ui, sans-serif; background: #f5f7fa;
|
| 289 |
+
color: #1a1a2e; line-height: 1.6; }}
|
| 290 |
+
header {{ background: linear-gradient(135deg,#1a1a2e,#16213e);
|
| 291 |
+
color:#fff; padding: 2.5rem 2rem; }}
|
| 292 |
+
header h1 {{ font-size: 2rem; margin-bottom:.4rem; }}
|
| 293 |
+
header p {{ opacity:.8; font-size:1rem; }}
|
| 294 |
+
.version {{ display:inline-block; background:#e94560; color:#fff;
|
| 295 |
+
font-size:.75rem; padding:.2rem .6rem; border-radius:12px;
|
| 296 |
+
margin-left:.6rem; vertical-align:middle; }}
|
| 297 |
+
main {{ max-width: 860px; margin: 2rem auto; padding: 0 1rem 4rem; }}
|
| 298 |
+
section {{ margin-bottom: 2.5rem; }}
|
| 299 |
+
h2 {{ font-size: 1.25rem; border-bottom: 2px solid #e94560;
|
| 300 |
+
padding-bottom:.4rem; margin-bottom:1rem; color:#16213e; }}
|
| 301 |
+
.auth-box {{ background:#fff3cd; border-left:4px solid #ffc107;
|
| 302 |
+
padding:1rem 1.2rem; border-radius:6px; margin-bottom:1.5rem; }}
|
| 303 |
+
.auth-box code {{ background:#ffe69c; padding:.1rem .4rem; border-radius:3px; }}
|
| 304 |
+
.endpoint {{ background:#fff; border:1px solid #e0e4ef; border-radius:8px;
|
| 305 |
+
padding:1.2rem 1.4rem; margin-bottom:1rem;
|
| 306 |
+
box-shadow:0 1px 4px rgba(0,0,0,.06); }}
|
| 307 |
+
.ep-header {{ display:flex; align-items:center; gap:.8rem; flex-wrap:wrap;
|
| 308 |
+
margin-bottom:.5rem; }}
|
| 309 |
+
.badge {{ color:#fff; font-size:.7rem; font-weight:700; padding:.25rem .6rem;
|
| 310 |
+
border-radius:4px; letter-spacing:.05em; }}
|
| 311 |
+
.ep-path {{ font-family:monospace; font-size:1rem; font-weight:600; }}
|
| 312 |
+
.ep-summary {{ color:#555; font-size:.9rem; }}
|
| 313 |
+
.ep-desc {{ color:#444; font-size:.9rem; margin-bottom:.6rem; }}
|
| 314 |
+
.param-table {{ width:100%; border-collapse:collapse; font-size:.85rem; margin-top:.5rem; }}
|
| 315 |
+
.param-table th {{ background:#f0f2f8; text-align:left;
|
| 316 |
+
padding:.4rem .6rem; border:1px solid #dde; }}
|
| 317 |
+
.param-table td {{ padding:.4rem .6rem; border:1px solid #eef; }}
|
| 318 |
+
pre {{ background:#1e2235; color:#cdd6f4; padding:1rem; border-radius:6px;
|
| 319 |
+
overflow-x:auto; font-size:.85rem; margin-top:.5rem; }}
|
| 320 |
+
.links {{ display:flex; gap:1rem; flex-wrap:wrap; }}
|
| 321 |
+
.links a {{ display:inline-block; background:#16213e; color:#fff; text-decoration:none;
|
| 322 |
+
padding:.5rem 1.1rem; border-radius:6px; font-size:.9rem; }}
|
| 323 |
+
.links a:hover {{ background:#e94560; }}
|
| 324 |
+
</style>
|
| 325 |
+
</head>
|
| 326 |
+
<body>
|
| 327 |
+
<header>
|
| 328 |
+
<h1>{app.title} <span class="version">v{app.version}</span></h1>
|
| 329 |
+
<p>Detección de entidades nombradas (NER) sobre documentos de texto.</p>
|
| 330 |
+
</header>
|
| 331 |
+
<main>
|
| 332 |
+
|
| 333 |
+
<section>
|
| 334 |
+
<h2>Documentación interactiva</h2>
|
| 335 |
+
<div class="links">
|
| 336 |
+
<a href="/docs">Swagger UI</a>
|
| 337 |
+
<a href="/redoc">ReDoc</a>
|
| 338 |
+
</div>
|
| 339 |
+
</section>
|
| 340 |
+
|
| 341 |
+
<section>
|
| 342 |
+
<h2>Autenticación</h2>
|
| 343 |
+
<div class="auth-box">
|
| 344 |
+
Todos los endpoints de anotación requieren un <strong>Bearer token</strong>
|
| 345 |
+
en la cabecera HTTP:<br/><br/>
|
| 346 |
+
<code>Authorization: Bearer <tu-token></code><br/><br/>
|
| 347 |
+
El token se configura mediante la variable de entorno <code>API_TOKEN</code>
|
| 348 |
+
en el Space de Hugging Face.
|
| 349 |
+
</div>
|
| 350 |
+
</section>
|
| 351 |
+
|
| 352 |
+
<section>
|
| 353 |
+
<h2>Endpoints</h2>
|
| 354 |
+
{endpoints_html}
|
| 355 |
+
</section>
|
| 356 |
+
|
| 357 |
+
<section>
|
| 358 |
+
<h2>Ejemplos de uso</h2>
|
| 359 |
+
<p><strong>Subir archivo .txt:</strong></p>
|
| 360 |
+
<pre>curl -X POST "https://<space>.hf.space/annotate" \\
|
| 361 |
+
-H "Authorization: Bearer <token>" \\
|
| 362 |
+
-F "file=@documento.txt"</pre>
|
| 363 |
+
|
| 364 |
+
<p style="margin-top:1rem"><strong>Enviar texto como JSON:</strong></p>
|
| 365 |
+
<pre>curl -X POST "https://<space>.hf.space/annotate/text" \\
|
| 366 |
+
-H "Authorization: Bearer <token>" \\
|
| 367 |
+
-H "Content-Type: application/json" \\
|
| 368 |
+
-d '{{"text": "John lives in New York."}}'</pre>
|
| 369 |
+
</section>
|
| 370 |
+
|
| 371 |
+
</main>
|
| 372 |
+
</body>
|
| 373 |
+
</html>"""
|
| 374 |
+
return HTMLResponse(content=html)
|