Clones a github repo and puts all the code into a single text file perfect for LLMs

collect_code.sh

#!/bin/bash

# Check if a GitHub URL is provided as an argument
if [ -z "$1" ]; then
    echo "Usage: $0 <github_url>"
    exit 1
fi

# Store the GitHub URL
GIT_URL="$1"

# Extract the repository name from the URL
REPO_NAME=$(basename -s .git "$GIT_URL")

# Clone the GitHub repository
git clone "$GIT_URL"

# Change directory to the cloned repository
cd "$REPO_NAME" || exit 1

# Create or clear the 'code' file
> code

# Define an array of popular code file extensions
EXTENSIONS=("py" "js" "ts" "jsx" "tsx" "rs" "ex" "exs" "go" "java" "c" "cpp" "h" "hpp" \
"cs" "rb" "php" "html" "css" "kt" "swift" "scala" "sh" "pl" "r" "lua" "m" "erl" "hs")

# Build the find command arguments to search for files with the specified extensions
FIND_ARGS=()
for EXT in "${EXTENSIONS[@]}"; do
    FIND_ARGS+=( -iname "*.$EXT" -o )
done
# Remove the last '-o' (logical OR) operator
unset 'FIND_ARGS[${#FIND_ARGS[@]}-1]'

# Find all files matching the extensions and process them
find . -type f \( "${FIND_ARGS[@]}" \) | while read -r FILE; do
    # Append the filename to the 'code' file
    echo "$FILE" >> code
    # Append the file content to the 'code' file
    cat "$FILE" >> code
done

Nice! Thank you.

Awesome

⚠️Only use this for repositories you trust.⚠️

Since the script write its output into a file ("code") inside the repository, the author the repo you clone has full control of what file this might be. It could for example be a symlink to a file you don't want to be written to.