prepare-commit-msg

#!/bin/bash

# Define the default commit message file
DEFAULT_COMMIT_MSG_FILE=".git/COMMIT_EDITMSG"
COMMIT_MSG_FILE="${1:-$DEFAULT_COMMIT_MSG_FILE}"

# Logging function with timestamps
log_message() {
    local level="$1"
    local message="$2"
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] [$level] $message" >> /tmp/prepare-commit-msg.log
}

# Function to generate fallback commit message based on file changes
generate_fallback_message() {
    local diff="$1"
    
    # Analyze the diff to determine commit type
    if echo "$diff" | grep -q "^+.*test\|spec"; then
        echo "test: add or update tests"
    elif echo "$diff" | grep -q "^+.*\.md\|README\|doc"; then
        echo "docs: update documentation"
    elif echo "$diff" | grep -qE "^\+.*fix|bug|error"; then
        echo "fix: resolve issue"
    elif echo "$diff" | grep -q "^+"; then
        echo "feat: add new changes"
    else
        echo "chore: update files"
    fi
}

# Read or set default configuration values
DEFAULT_CHATGPT_PROMPT="Suggest commit message based on the following diff:\n\n\`\`\`\n{{diff}}\n\`\`\`\n\ncommit messages must be following these rules:\n - follow conventional commits\n - message format must be in this format: \"<TYPE>: <DESCRIPTION>\"\n - <TYPE> must be the prefix of commit message and must be one of the following: feat, fix, docs, style, refactor, test, chore\n - <DESCRIPTION> must be the description of the commit in lowercase and without any special characters\n\nEXAMPLES COMMIT MESSAGE:\n - fix: add password regex pattern\n - feat: add new test cases\n\nNOTE: Response only commit message, no explanation anymore\n\nACTUAL COMMIT MESSAGE: \n"
DEFAULT_MODEL="gpt-4o-mini"
DEFAULT_MAX_TOKENS=128000
DEFAULT_API_ENDPOINT="https://api.openai.com/v1/chat/completions"

# Get configuration values with validation
CHATGPT_PROMPT=$(git config --get ai.prompt)
if [ -z "$CHATGPT_PROMPT" ]; then
    CHATGPT_PROMPT="$DEFAULT_CHATGPT_PROMPT"
    log_message "INFO" "Using default AI prompt"
fi

MODEL=$(git config --get ai.model)
if [ -z "$MODEL" ]; then
    MODEL="$DEFAULT_MODEL"
    log_message "INFO" "Using default model: $MODEL"
fi

MAX_TOKENS=$(git config --get ai.maxtokens)
if [ -z "$MAX_TOKENS" ]; then
    MAX_TOKENS="$DEFAULT_MAX_TOKENS"
    log_message "INFO" "Using default max tokens: $MAX_TOKENS"
elif ! [[ "$MAX_TOKENS" =~ ^[0-9]+$ ]] || [ "$MAX_TOKENS" -lt 1 ] || [ "$MAX_TOKENS" -gt 200000 ]; then
    log_message "WARN" "Invalid max tokens value: $MAX_TOKENS. Using default: $DEFAULT_MAX_TOKENS"
    MAX_TOKENS="$DEFAULT_MAX_TOKENS"
fi

API_ENDPOINT=$(git config --get ai.endpoint)
if [ -z "$API_ENDPOINT" ]; then
    API_ENDPOINT="$DEFAULT_API_ENDPOINT"
    log_message "INFO" "Using default API endpoint"
elif ! [[ "$API_ENDPOINT" =~ ^https?:// ]]; then
    log_message "ERROR" "Invalid API endpoint format: $API_ENDPOINT. Must start with http:// or https://"
    echo "chore: update files (API endpoint configuration error)"
    exit 0
fi

# Initialize logging
log_message "INFO" "Starting prepare-commit-msg hook with arguments: $@"

# Check if a commit message file is provided or exists
if [ ! -f "$COMMIT_MSG_FILE" ]; then
  log_message "ERROR" "Commit message file does not exist or was not provided: $COMMIT_MSG_FILE"
  echo "chore: update files (commit message file error)"
  exit 0
fi

# Collect the git diff
DIFF=$(git diff --cached)

# Validate that there are staged changes
if [ -z "$DIFF" ]; then
    log_message "WARN" "No staged changes found in git diff --cached"
    echo "chore: update files (no staged changes)"
    exit 0
fi

log_message "INFO" "Found staged changes, diff length: ${#DIFF} characters"

# Replace placeholder with actual diff in the prompt
COMMIT_PROMPT=${CHATGPT_PROMPT/\{\{diff\}\}/"$DIFF"}

log_message "DEBUG" "Generated prompt: $COMMIT_PROMPT"

# Use jq to construct the JSON payload
DATA=$(jq -n --arg prompt "$COMMIT_PROMPT" --arg model "$MODEL" --argjson max_tokens "$MAX_TOKENS" \
'{
  "model": $model,
  "messages": [
    {
      "role": "user",
      "content": $prompt
    }
  ],
  "temperature": 0.5,
  "max_tokens": $max_tokens
}')

# Make sure jq successfully created the payload
if [ -z "$DATA" ]; then
  log_message "ERROR" "Failed to construct JSON payload with jq"
  FALLBACK_MSG=$(generate_fallback_message "$DIFF")
  echo "$FALLBACK_MSG"
  exit 0
fi

log_message "INFO" "JSON payload constructed successfully"

# Get the OpenAI API key from the git configuration
API_KEY=$(git config --get openai.apikey)

if [ -z "$API_KEY" ]; then
    log_message "ERROR" "OpenAI API key is not set in git configuration"
    FALLBACK_MSG=$(generate_fallback_message "$DIFF")
    echo "$FALLBACK_MSG"
    exit 0
elif [ ${#API_KEY} -lt 20 ]; then
    log_message "ERROR" "OpenAI API key appears to be invalid (too short)"
    FALLBACK_MSG=$(generate_fallback_message "$DIFF")
    echo "$FALLBACK_MSG"
    exit 0
fi

log_message "INFO" "API key validated successfully"

# Make the API request
API_URL="$API_ENDPOINT"
log_message "INFO" "Making API request to: $API_URL"

RESPONSE=$(curl -s -X POST "$API_URL" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $API_KEY" \
--data "$DATA" \
--max-time 30 \
--write-out "HTTPSTATUS:%{http_code}")

# Extract HTTP status code and response body
HTTP_STATUS=$(echo "$RESPONSE" | grep -o "HTTPSTATUS:[0-9]*" | cut -d: -f2)
RESPONSE_BODY=$(echo "$RESPONSE" | sed 's/HTTPSTATUS:[0-9]*$//')

log_message "INFO" "API response HTTP status: $HTTP_STATUS"

# Check if the response was successful
if [ -z "$RESPONSE_BODY" ]; then
  log_message "ERROR" "No response from the API"
  FALLBACK_MSG=$(generate_fallback_message "$DIFF")
  echo "$FALLBACK_MSG"
  exit 0
fi

# Check HTTP status
if [ "$HTTP_STATUS" != "200" ]; then
  log_message "ERROR" "API request failed with HTTP status $HTTP_STATUS"
  # Try to extract error message from response
  API_ERROR=$(echo "$RESPONSE_BODY" | jq -r '.error.message // .error // "Unknown API error"' 2>/dev/null || echo "Unknown API error")
  log_message "ERROR" "API error: $API_ERROR"
  FALLBACK_MSG=$(generate_fallback_message "$DIFF")
  echo "$FALLBACK_MSG"
  exit 0
fi

# Validate JSON response structure
if ! echo "$RESPONSE_BODY" | jq -e '.choices[0].message.content' > /dev/null 2>&1; then
  log_message "ERROR" "Invalid API response structure"
  log_message "DEBUG" "Response body: $RESPONSE_BODY"
  FALLBACK_MSG=$(generate_fallback_message "$DIFF")
  echo "$FALLBACK_MSG"
  exit 0
fi

# Parse the response to get the commit message
COMMIT_MSG=$(echo "$RESPONSE_BODY" | jq -r '.choices[0].message.content')

# Check if the commit message is valid
if [ -z "$COMMIT_MSG" ] || [ "$COMMIT_MSG" = "null" ]; then
  log_message "ERROR" "AI generated null or empty commit message"
  FALLBACK_MSG=$(generate_fallback_message "$DIFF")
  echo "$FALLBACK_MSG"
  exit 0
fi

# Validate commit message format (basic check)
if ! echo "$COMMIT_MSG" | grep -qE "^(feat|fix|docs|style|refactor|test|chore):\s*.+"; then
  log_message "WARN" "AI generated message doesn't follow conventional commits format: $COMMIT_MSG"
  # Still use it but log the warning
fi

log_message "INFO" "Successfully generated commit message: $COMMIT_MSG"

# Output the commit message
echo "$COMMIT_MSG"