feat(web-search): tool DuckDuckGo abierta en /web-search para watsonx Orchestrate

Sub-app FastAPI montada en /web-search con búsqueda vía ddgs.
OpenAPI 3.0.3 + servers absoluto + operationId webSearch listo para
importar como external tool en wxO. Sin auth (uso de taller/demo).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.env.example

@@ -6,3 +6,11 @@ BASE_URL=https://taller-wox.fitlabs.dev
 MATERIAL_DIR=./material
 DB_PATH=./leads.db
 REPORTS_OUTPUT_DIR=./app/data/reports_output
+
+# --- Web Search Tool para watsonx Orchestrate (montada en /web-search) ---
+# Endpoint abierto (sin auth) — apto para taller/demo, no para producción.
+# Por defecto se calcula como ${BASE_URL}/web-search. Override sólo si hace falta.
+# WEB_SEARCH_PUBLIC_BASE_URL=https://taller-wox.fitlabs.dev/web-search
+WEB_SEARCH_DEFAULT_MAX_RESULTS=5
+WEB_SEARCH_DEFAULT_REGION=wt-wt
+WEB_SEARCH_REQUEST_TIMEOUT=12

app/main.py

@@ -7,6 +7,7 @@ from fastapi.staticfiles import StaticFiles
 from app import admin, benefits_api, frontend, reports_api
 from app.config import get_settings
 from app.db import init_db
+from app.web_search import web_search_app
 
 settings = get_settings()
 
@@ -36,6 +37,8 @@ app.mount(
     name="reports_output",
 )
 
+app.mount("/web-search", web_search_app)
+
 
 @app.on_event("startup")
 def on_startup() -> None:

app/web_search/__init__.py

@@ -0,0 +1,3 @@
+from app.web_search.api import web_search_app
+
+__all__ = ["web_search_app"]

app/web_search/api.py

@@ -0,0 +1,154 @@
+"""Sub-app FastAPI con la web-search tool para watsonx Orchestrate.
+
+Se monta en la app principal bajo `/web-search`, por lo que:
+- `GET /web-search/search`      → endpoint de búsqueda (abierto, sin auth)
+- `GET /web-search/health`      → health check (fuera del schema)
+- `GET /web-search/openapi.json`→ spec OpenAPI 3.0.3 listo para importar en wxO
+- `GET /web-search/docs`        → Swagger UI para probar
+"""
+from __future__ import annotations
+
+from typing import Annotated
+
+from fastapi import Depends, FastAPI, HTTPException, Query, status
+from fastapi.openapi.utils import get_openapi
+from pydantic import BaseModel, Field
+
+from app.web_search.config import WebSearchSettings, get_web_search_settings
+from app.web_search.search import SearchProviderError, search_web
+
+
+OPERATION_DESCRIPTION = (
+    "Busca en la web y devuelve los resultados más relevantes (título, URL y "
+    "resumen). Úsala cuando necesites información actual o reciente, noticias, "
+    "datos que puedan haber cambiado, o cualquier información que no esté en tu "
+    "conocimiento base."
+)
+
+
+class SearchResultModel(BaseModel):
+    title: str = Field(..., description="Título del resultado de búsqueda.")
+    url: str = Field(..., description="URL del resultado.")
+    snippet: str = Field(
+        ...,
+        description="Fragmento de texto / resumen del resultado (texto plano, sin HTML, máx. 300 caracteres).",
+    )
+
+
+class SearchResponse(BaseModel):
+    query: str = Field(..., description="Consulta original recibida.")
+    count: int = Field(..., description="Cantidad de resultados devueltos.")
+    results: list[SearchResultModel] = Field(
+        ..., description="Lista de resultados, ordenados de mayor a menor relevancia."
+    )
+
+
+web_search_app = FastAPI(
+    title="FIT Web Search Tool",
+    version="1.0.0",
+    description=(
+        "Herramienta de búsqueda web para agentes de watsonx Orchestrate. "
+        "Devuelve resultados normalizados desde DuckDuckGo."
+    ),
+    docs_url="/docs",
+    redoc_url=None,
+    openapi_url="/openapi.json",
+)
+
+
+@web_search_app.get(
+    "/health",
+    include_in_schema=False,
+)
+def health() -> dict[str, str]:
+    return {"status": "ok"}
+
+
+@web_search_app.get(
+    "/search",
+    operation_id="webSearch",
+    summary="Buscar en la web",
+    description=OPERATION_DESCRIPTION,
+    response_model=SearchResponse,
+    responses={
+        422: {"description": "Parámetros de entrada inválidos."},
+        502: {"description": "Error en el proveedor de búsqueda."},
+    },
+)
+def search(
+    settings: Annotated[WebSearchSettings, Depends(get_web_search_settings)],
+    query: Annotated[
+        str,
+        Query(
+            min_length=1,
+            max_length=400,
+            description="Consulta de búsqueda en lenguaje natural.",
+            examples=["últimas noticias IBM watsonx orchestrate"],
+        ),
+    ],
+    max_results: Annotated[
+        int,
+        Query(
+            ge=1,
+            le=10,
+            description="Cantidad de resultados a devolver (1–10). Por defecto 5.",
+        ),
+    ] = 5,
+    region: Annotated[
+        str,
+        Query(
+            min_length=2,
+            max_length=10,
+            description=(
+                "Región/idioma del motor de búsqueda. Ejemplos: 'wt-wt' (global), "
+                "'cl-es' (Chile/español), 'mx-es' (México/español), 'us-en' (EEUU/inglés)."
+            ),
+            examples=["wt-wt", "cl-es", "mx-es", "us-en"],
+        ),
+    ] = "wt-wt",
+) -> SearchResponse:
+    try:
+        results = search_web(
+            query=query,
+            max_results=max_results,
+            region=region,
+            timeout=settings.request_timeout,
+        )
+    except SearchProviderError:
+        raise HTTPException(
+            status_code=status.HTTP_502_BAD_GATEWAY,
+            detail="Error en el proveedor de búsqueda",
+        )
+
+    return SearchResponse(
+        query=query,
+        count=len(results),
+        results=[
+            SearchResultModel(title=r.title, url=r.url, snippet=r.snippet) for r in results
+        ],
+    )
+
+
+def _custom_openapi() -> dict:
+    if web_search_app.openapi_schema:
+        return web_search_app.openapi_schema
+
+    settings = get_web_search_settings()
+    schema = get_openapi(
+        title=web_search_app.title,
+        version=web_search_app.version,
+        description=web_search_app.description,
+        routes=web_search_app.routes,
+    )
+
+    # wxO prefiere 3.0.x (FastAPI emite 3.1.0 por defecto)
+    schema["openapi"] = "3.0.3"
+
+    # Bloque servers absoluto — obligatorio para wxO
+    schema["servers"] = [{"url": settings.public_base_url}]
+
+    web_search_app.openapi_schema = schema
+    return schema
+
+
+web_search_app.openapi = _custom_openapi  # type: ignore[assignment]

app/web_search/config.py

@@ -0,0 +1,36 @@
+import os
+from functools import lru_cache
+from dataclasses import dataclass
+
+
+@dataclass(frozen=True)
+class WebSearchSettings:
+    public_base_url: str
+    default_max_results: int
+    default_region: str
+    request_timeout: int
+
+
+def _int_env(name: str, default: int) -> int:
+    raw = os.environ.get(name)
+    if raw is None or raw == "":
+        return default
+    try:
+        return int(raw)
+    except ValueError:
+        return default
+
+
+@lru_cache
+def get_web_search_settings() -> WebSearchSettings:
+    base_url = os.environ.get("BASE_URL", "https://taller-wox.fitlabs.dev").rstrip("/")
+    public_base_url = os.environ.get(
+        "WEB_SEARCH_PUBLIC_BASE_URL",
+        f"{base_url}/web-search",
+    ).rstrip("/")
+    return WebSearchSettings(
+        public_base_url=public_base_url,
+        default_max_results=_int_env("WEB_SEARCH_DEFAULT_MAX_RESULTS", 5),
+        default_region=os.environ.get("WEB_SEARCH_DEFAULT_REGION", "wt-wt"),
+        request_timeout=_int_env("WEB_SEARCH_REQUEST_TIMEOUT", 12),
+    )

app/web_search/search.py

@@ -0,0 +1,65 @@
+"""Único punto del servicio que toca ddgs.
+
+Si se cambia el backend (Brave, Google CSE, etc.) sólo se reescribe este módulo;
+el contrato `search_web(...)` -> list[SearchResult] no cambia.
+"""
+from __future__ import annotations
+
+import html
+import re
+from dataclasses import dataclass
+
+from ddgs import DDGS
+
+
+SNIPPET_MAX_LEN = 300
+_TAG_RE = re.compile(r"<[^>]+>")
+_WS_RE = re.compile(r"\s+")
+
+
+class SearchProviderError(RuntimeError):
+    """Error opaco del proveedor de búsqueda. main.py lo traduce a HTTP 502."""
+
+
+@dataclass(frozen=True)
+class SearchResult:
+    title: str
+    url: str
+    snippet: str
+
+
+def _clean_snippet(raw: str | None) -> str:
+    if not raw:
+        return ""
+    text = _TAG_RE.sub("", raw)
+    text = html.unescape(text)
+    text = _WS_RE.sub(" ", text).strip()
+    if len(text) > SNIPPET_MAX_LEN:
+        text = text[: SNIPPET_MAX_LEN - 1].rstrip() + "…"
+    return text
+
+
+def search_web(query: str, max_results: int, region: str, timeout: int) -> list[SearchResult]:
+    try:
+        raw = DDGS(timeout=timeout).text(
+            query,
+            region=region,
+            max_results=max_results,
+        )
+    except Exception as exc:  # ddgs lanza varias excepciones distintas; las unificamos
+        raise SearchProviderError(str(exc)) from exc
+
+    results: list[SearchResult] = []
+    for item in raw or []:
+        title = (item.get("title") or "").strip()
+        url = (item.get("href") or item.get("url") or "").strip()
+        if not url:
+            continue
+        results.append(
+            SearchResult(
+                title=title or url,
+                url=url,
+                snippet=_clean_snippet(item.get("body") or item.get("snippet")),
+            )
+        )
+    return results

docker-compose.yml

@@ -23,6 +23,10 @@ services:
       MATERIAL_DIR: /app/material
       DB_PATH: /app/db/leads.db
       REPORTS_OUTPUT_DIR: /app/app/data/reports_output
+      WEB_SEARCH_PUBLIC_BASE_URL: ${WEB_SEARCH_PUBLIC_BASE_URL:-https://taller-wox.fitlabs.dev/web-search}
+      WEB_SEARCH_DEFAULT_MAX_RESULTS: ${WEB_SEARCH_DEFAULT_MAX_RESULTS:-5}
+      WEB_SEARCH_DEFAULT_REGION: ${WEB_SEARCH_DEFAULT_REGION:-wt-wt}
+      WEB_SEARCH_REQUEST_TIMEOUT: ${WEB_SEARCH_REQUEST_TIMEOUT:-12}
     volumes:
       - taller-wox-db:/app/db
       - taller-wox-reports:/app/app/data/reports_output

requirements.txt

@@ -10,3 +10,4 @@ pandas==2.2.3
 plotly==5.24.1
 pytest==8.3.3
 httpx==0.27.2
+ddgs>=9,<10