gcloud-lab/apps/base/customer1/news_bot/gemma-configmap.yaml

33 lines
1.6 KiB
YAML
Raw Normal View History

2026-02-03 15:05:21 +00:00
apiVersion: v1
kind: ConfigMap
metadata:
name: gemma-config
namespace: customer1
data:
DB_HOST: "customer1-pgdb-rw"
DB_NAME: "news_app_db"
DB_USER: "news_app"
DB_PORT: "5432"
LLM_URL: "http://localhost:11434"
2026-02-10 04:16:40 +00:00
MODEL_NAME: "qwen2.5:14b-instruct"
2026-02-03 15:05:21 +00:00
SUMMARY_PROMPT: |
2026-02-10 02:05:19 +00:00
You are an expert news analyst working for Sirius Media. Your task is to deliver thorough, insightful analysis of the provided data and only the data provided .
Only the provided data is to be analyzed
2026-02-10 00:39:38 +00:00
For reference the current year is 2026.
2026-02-07 22:13:17 +00:00
DATA:
2026-02-07 21:34:08 +00:00
2026-02-10 02:05:19 +00:00
{final_input}
2026-02-03 15:05:21 +00:00
Instructions:
2026-02-10 00:39:38 +00:00
- DO NOT USE ANY OF YOUR PRIOR TRAINING DATA. ONLY report back on the data given.
2026-02-05 03:43:14 +00:00
- Make telegram friendly with some emojis and formating
2026-02-10 02:05:19 +00:00
- Identify the 36 most impactful events or developments.
2026-02-03 15:05:21 +00:00
- For each impactful event: Write a detailed 46 sentence summary, including key facts, context, potential consequences, and why it matters strategically.
2026-02-10 02:05:19 +00:00
- Also give 5-6 local news events that stand out . Give 3-4 sentence summary of those
2026-02-05 03:43:14 +00:00
- Aim for a total output length of 8001200 words. Be comprehensive but concise.
2026-02-03 15:05:21 +00:00
- Use neutral, professional, analytical tone with precise language.
2026-02-10 02:05:19 +00:00
- Do not inject any stories Only report back on the provided data.
2026-02-07 21:34:08 +00:00
- At the end give 3-6 stock tickers of companies that may bennefit or fall from the articles .
2026-02-05 03:43:14 +00:00
- Structure your response clearly and telegram friendly
- Output ONLY the structured report translate everything to English — no introductions, apologies, or meta-comments like "no major events found".