Skip to content

README_grok.markdown

Latest commit

 

History

History
executable file
·
130 lines (87 loc) · 2.93 KB

README_grok.markdown

File metadata and controls

executable file
·
130 lines (87 loc) · 2.93 KB

Script Analysis with Ollama (analyze_with_ollama_picker_analysis.sh)

Author: Ian Trimble
Version: 1.0.0
Created: April 28, 2025

Overview

The analyze_with_ollama_picker_analysis.sh script is a command-line tool designed to analyze script files using Ollama models and generate detailed Markdown documentation. It supports various script types, including shell and AppleScript, and provides extensive benchmarking to track performance metrics.

Table of Contents

  1. Prerequisites
  2. Installation
  3. Usage
  4. Configuration
  5. Output
  6. Troubleshooting
  7. License

Prerequisites

  • Operating System: macOS
  • Required Tools:
    • jq: For JSON processing.
    • bc: For mathematical calculations.
    • curl: For making API requests.
    • ollama: For running local language models.
  • Ollama Models: At least one model must be installed (e.g., qwen2.5:1.5b).

Install Prerequisites

Use Homebrew to install the required tools:

brew install jq bc curl ollama

Installation

  1. Download the Script
    Save the script as analyze_with_ollama_picker_analysis.sh in your desired directory.

  2. Make the Script Executable
    Run the following command:

    chmod +x analyze_with_ollama_picker_analysis.sh

  3. Start the Ollama Server
    Ensure the Ollama server is running:

    ollama serve &

  4. Pull a Model
    Download at least one Ollama model:

    ollama pull qwen2.5:1.5b

Usage

To analyze a script, run:

./analyze_with_ollama_picker_analysis.sh <input_file> [model]
  • <input_file>: Path to the script file (e.g., .sh, .scpt).
  • [model]: Optional. Specify an Ollama model (e.g., qwen2.5:1.5b).

If no model is specified, the script will prompt for selection or default to the first available model.

Example

./analyze_with_ollama_picker_analysis.sh my_script.sh qwen2.5:1.5b

Configuration

  • Benchmark Directory: ${HOME}/ollama_benchmarks
    Stores benchmark logs and metrics.
  • README File: README.md in the current directory.
    Appends the generated documentation.

Output

The script generates:

  • README.md: Appends a new section with the analysis results.
  • Benchmark Logs:
    • benchmark_log.csv: Detailed performance metrics.
    • metrics_<session_id>.json: Session-specific metrics.
    • response_<session_id>.md: Raw Ollama response.

Troubleshooting

  • Model Not Found:
    • Ensure the model is installed: ollama pull <model>.
  • Dependency Errors:
    • Install missing tools: brew install jq bc curl ollama.
  • API Connection Issues:
    • Start the Ollama server: ollama serve &.

License

This script is provided under the MIT License. See the generated README.md for full license details.