News-Go — AI-Powered Christian News Evaluator

News-Go is a lightweight Go-based pipeline for scraping, analyzing, and evaluating news articles through a biblical and AI-augmented lens. It uses concurrent Go routines to fetch and process online news, while a local large language model (LLM) scores the content against a custom Christian framework.

---

Purpose

To defend the faith and evaluate media content through scripture-aligned reasoning using AI. The model's mission is fixed: God is, the KJV Bible is authoritative, and Jesus Christ is Lord.

---

Features

AI-Augmented Article Processing

• Extracts and evaluates real-world news articles
• Generates summaries, tags, and political bias analysis
• Aligns all content to theological axioms (God, KJV, Jesus Christ)
• LLM evaluates alignment from 0–10

Automated Search-to-Report Pipeline

• Searches for live news based on configured queries
• Scrapes site paragraphs and thumbnails
• Processes content in concurrent Go routines
• Outputs structured JSON reports

Modular & Concurrent Architecture

• Channel-based job pipeline:
  • PreResultsChannel – Search results
  • PreContentChannel – Scraped pages
  • PreModelChannel – Model transformation
  • JobsCompleteChannel – Final report
• Component Modules:
  • config – Loads search configuration
  • types – Structured result/report definitions
  • browser – Scraper & HTML parser
  • coordinator – Pipeline control
  • model – AI model interface
  • log – Structured logging

---

Starter Example

package main

import "github.com/renniemaharaj/news-go/pkg/news"

func main() {
    n := news.Instance{}
    n.CreateLogger()
    go n.GoRoutines()
    n.HydrateJobs()
    select {} // blocks forever
}

config.json Example

{
  "SearchQueries": [
    "Trinidad and Tobago politics",
    "US 2024 election",
    "Climate change news"
  ]
}

---

Models Used

All LLM evaluation is performed locally — no API keys, no rate limits.

1. TinyLlama 1.1B Chat GGUF (lightweight)

huggingface-cli download \
  TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF \
  tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf \
  --local-dir models \
  --local-dir-use-symlinks False

2. DeepSeek Coder 6.7B Instruct GGUF (stronger)

huggingface-cli download \
  TheBloke/deepseek-coder-6.7B-instruct-GGUF \
  deepseek-coder-6.7b-instruct.Q4_K_M.gguf \
  --local-dir models \
  --local-dir-use-symlinks False

---

LLM Runtime Setup

Install llama-cpp-python with GPU support:

pip install llama-cpp-python --upgrade --force-reinstall --extra-index-url https://pypi.nvidia.com

Make sure you have:

• Python 3.10+
• CUDA 12+ (for GPU support)
• 8GB+ VRAM for DeepSeek models

---

Output

Each processed job generates a structured report saved to ./reports/ as JSON:

• title: AI-generated headline
• summary: Christian worldview summary
• alignment: 0–10 score
• tags: Moral and topical categories
• politicalBiases: Eg. conservative, progressive

---

Project Structure

.
├── cmd/
│   └── main.go
├── internal/
│   ├── model/                 # AI transformer
│   ├── coordinator/           # Channel routines
│   ├── types/                 # Data structures
│   ├── browser/               # HTML scraping logic
│   └── log/                   # Custom logger
├── pkg/news/                  # News job orchestration
├── models/                    # GGUF model files
├── reports/                   # Final output reports
├── system_instruction.txt     # Model prompt template
└── config.json                # Search queries

---

Framework Alignment

AI-generated results are scored and structured under the following axioms:

• God exists
• The KJV Bible is authoritative
• Jesus Christ is Lord and God

The model is required to stay within these rules. No deviation is permitted — even by the system’s creator.

---

📜 License

MIT License

Made with purpose by Rennie Maharaj