Script to output a git repo's content to a single text file. Useful for prompting with LLMs where large volume of input is supported

repo2text.sh

#!/bin/bash
set -euo pipefail  # Exit on error, unset variables, and prevent errors in pipelines

# Usage information
usage() {
    echo "Usage: $0 <path-to-git-repo> <output-file>"
    exit 1
}

# Check for mandatory arguments
if [ "$#" -ne 2 ]; then
    usage
fi

GIT_REPO_PATH="$1"
OUTPUT_FILE="$2"

# Define ignore patterns for file types
declare -a ignore_patterns=(
    "*.bin"
    "*.png"
    "*.jpg"
    "*.gif"
    "*.mp4"
    "*.jpeg"
)

# # Verify the provided path is a git repository
# if [ ! -d "$GIT_REPO_PATH/.git" ]; then
#     echo "The provided path does not seem to be a git repository."
#     exit 1
# fi

# Create or clear the output file
: > "$OUTPUT_FILE"

# Output the directory tree to the top of the file
{
    echo "Directory tree of the repository:"
    tree "$GIT_REPO_PATH"
    echo -e "\n\n"
} >> "$OUTPUT_FILE"

# Function to output file contents with separators
output_file_contents() {
    local repo_path="$1"
    local output_file="$2"

    # Start the find command
    local find_command="find $repo_path -type f ! -path '*.git/*'"

    # Exclude defined ignore patterns
    for pattern in "${ignore_patterns[@]}"; do
        find_command+=" ! -name '$pattern'"
    done

    # Execute find command and process files
    eval "$find_command" | while read -r file; do
        {
            echo "---------- $file ----------"
            cat "$file"
            echo -e "\n"
        } >> "$output_file"
    done
}

# Call function to output file contents
output_file_contents "$GIT_REPO_PATH" "$OUTPUT_FILE"

echo "Script completed. Output is in $OUTPUT_FILE"