fix: scale extractor to 3 replicas in paper stage

The extraction queue had 3000+ SEC filings backed up with a single
extractor pod processing them at 10-115s each. Ollama handles
concurrent requests so multiple extractor pods can share the GPU.

infra/helm/stonks-oracle/values-paper.yaml

@@ -16,6 +16,7 @@ config:
   REDIS_DB: "2"
   DEPLOY_STAGE: "paper"
   POSTGRES_USER: "stonks_paper"
+  OLLAMA_BASE_URL: "http://10.1.1.12:2701"
 
 ## Secrets override: Alpaca paper trading API endpoint
 secrets:
@@ -31,3 +32,9 @@ ingress:
     superset: stonks-paper-dash.celestium.life
     trino: stonks-paper-trino.celestium.life
     tradingEngine: stonks-paper-trading.celestium.life
+
+## Scale extractor to 3 replicas — Ollama handles concurrent requests
+## and the backlog of SEC filings needs higher throughput
+services:
+  extractor:
+    replicas: 3