fuentes automaticas: el CODIGO agrega la fuente, no el LLM (ronda 13-14)
ci / lint-test (pull_request) Successful in 2s
isolation-test / probe (pull_request) Successful in 0s

- sources.mjs (nuevo): selectSource() + attribute() — seleccion de material
  ≤72h con score≥6, atribucion en formato fijo.
- core.mjs: las 4 piezas (digest/opinion/reflection/events) generan sobre
  la fuente seleccionada y el codigo agrega la atribucion DESPUES del LLM.
- guard.mjs:
  * R1/R2/R4/R5/R8 evaluan el TEXTO SIN la linea de atribucion
  * R3/R7 ven el texto completo (con atribucion) — la fuente garantiza
    que pasen
  * R9 (nueva): la pieza debe compartir terminos del dominito con su fuente
  * R8: superlativos estrechos (sin never/only que dejaban pasar aforismos)
  * deletion anadido a METADATA_TYPES (no cuenta hacia el tope diario)
- guard.test.mjs: 24/24 verdes (incluye tests R14a/c/d)
This commit is contained in:
2026-11-02 04:08:26 +00:00
parent 4e4fffaa16
commit dd8e933ba5
5 changed files with 152 additions and 12 deletions
+45
View File
@@ -0,0 +1,45 @@
// sources.mjs — selección de material de origen y atribución automática
// Ronda 13: el CÓDIGO agrega la fuente, no el LLM.
import { readFileSync } from 'node:fs'
import { createRequire } from 'node:module'
const require_ = createRequire(import.meta.url)
let _pool = null
function pool() {
if (!_pool) {
const pg = require_('/srv/cayce-core/node_modules/pg')
const env = Object.fromEntries(
readFileSync('/srv/hardcap-news/backend/.env', 'utf8').split('\n').filter(l => l.includes('=')).map(l => [l.slice(0, l.indexOf('=')), l.slice(l.indexOf('=') + 1)])
)
_pool = new pg.Pool({ connectionString: env.DATABASE_URL })
}
return _pool
}
export async function selectSource(topic = null) {
const p = pool()
const q = topic
? `SELECT id, title, "originalUrl", "publishedAt"::text FROM articles WHERE status='PUBLISHED' AND "relevanceScore">=6 AND "publishedAt">now()-interval '72 hours' AND $1=ANY(topics) ORDER BY "relevanceScore" DESC LIMIT 1`
: `SELECT id, title, "originalUrl", "publishedAt"::text FROM articles WHERE status='PUBLISHED' AND "relevanceScore">=6 AND "publishedAt">now()-interval '72 hours' ORDER BY "relevanceScore" DESC LIMIT 1`
const { rows } = await p.query(q, topic ? [topic] : [])
return rows[0] || null
}
export async function selectBatch(n = 5, topic = null) {
const p = pool()
const q = topic
? `SELECT id, title, summary, "originalUrl", "publishedAt"::text, "relevanceScore" FROM articles WHERE status='PUBLISHED' AND "relevanceScore">=6 AND "publishedAt">now()-interval '72 hours' AND $1=ANY(topics) ORDER BY "relevanceScore" DESC LIMIT $2`
: `SELECT id, title, summary, "originalUrl", "publishedAt"::text, "relevanceScore" FROM articles WHERE status='PUBLISHED' AND "relevanceScore">=6 AND "publishedAt">now()-interval '72 hours' ORDER BY "relevanceScore" DESC LIMIT $1`
const { rows } = await p.query(q, topic ? [topic, n] : [n])
return rows
}
export function attribute(source) {
if (!source) return ''
const date = (source.publishedAt || '').slice(0, 10) || new Date().toISOString().slice(0, 10)
return `\n\n— fuente: ${source.originalUrl} · ${date}`
}
export function attributeURL(url, dateStr) {
return `\n\n— fuente: ${url} · ${dateStr}`
}