genai-git-commit/git_commit_ai.py

import subprocess
import os
import google.generativeai as genai
import re


def get_staged_diff():
    """
    Retrieves the diff of staged files using git.

    Returns:
        str: The diff of the staged files, or None on error.
    """
    try:
        # Use subprocess.run for better control and error handling
        process = subprocess.run(
            ["git", "diff", "--staged"],  # Corrected: --staged is the correct option
            capture_output=True,
            text=True,  # Ensure output is returned as text
            check=True,  # Raise an exception for non-zero exit codes
        )
        return process.stdout
    except subprocess.CalledProcessError as e:
        print(f"Error getting staged diff: {e}")
        print(f"  stderr: {e.stderr}")  # Print stderr for more details
        return None
    except FileNotFoundError:
        print(
            "Error: git command not found.  Please ensure Git is installed and in your PATH."
        )
        return None
    except Exception as e:
        print(f"An unexpected error occurred: {e}")
        return None


def get_project_files():
    """Gets a list of all files tracked in the latest commit (HEAD)."""
    try:
        process = subprocess.run(
            # Changed command to list files in the last commit
            ["git", "ls-tree", "-r", "--name-only", "HEAD"],
            capture_output=True,
            text=True,
            check=True,
            cwd=os.getcwd(),  # Ensure it runs in the correct directory
        )
        return process.stdout.splitlines()
    except subprocess.CalledProcessError as e:
        print(f"Error getting project file list: {e}")
        print(f"  stderr: {e.stderr}")
        return []  # Return empty list on error
    except FileNotFoundError:
        print("Error: git command not found. Is Git installed and in your PATH?")
        return []
    except Exception as e:
        print(f"An unexpected error occurred while listing files: {e}")
        return []


def get_file_content(filepath):
    """Reads the content of a file relative to the script's CWD."""
    # Consider adding checks to prevent reading files outside the repo
    try:
        # Assuming the script runs from the repo root
        with open(filepath, "r", encoding="utf-8") as f:
            return f.read()
    except FileNotFoundError:
        print(f"Warning: File not found: {filepath}")
        return None
    except IsADirectoryError:
        print(f"Warning: Path is a directory, not a file: {filepath}")
        return None
    except Exception as e:
        print(f"Warning: Error reading file {filepath}: {e}")
        return None


def generate_commit_message(diff, gemini_api_key):
    """
    Generates a commit message using the Gemini API, given the diff.

    Args:
        diff (str): The diff of the staged files.
        gemini_api_key (str): Your Gemini API key.

    Returns:
        str: The generated commit message, or None on error.
    """
    if not diff:
        print("Error: No diff provided to generate commit message.")
        return None

    genai.configure(api_key=gemini_api_key)
    # Use the intended model name
    model = genai.GenerativeModel("gemini-1.5-flash")

    project_files_list = get_project_files()

    # Define prompt as a regular string, not f-string, placeholders will be filled by .format()
    prompt = f"""
    You are an expert assistant that generates Git commit messages following conventional commit standards.
    Analyze the following diff of staged files and generate ONLY the commit message (subject and body) adhering to standard Git conventions.

    1.  **Subject Line:** Write a concise, imperative subject line summarizing the change (max 50 characters). Start with a capital letter. Do not end with a period. Use standard commit types like 'feat:', 'fix:', 'refactor:', 'docs:', 'test:', 'chore:', etc.
    2.  **Blank Line:** Leave a single blank line between the subject and the body.
    3.  **Body:** Write a detailed but precise body explaining the 'what' and 'why' of the changes. Wrap lines at 72 characters. Focus on the motivation for the change and contrast its implementation with the previous behavior. If the change is trivial, the body can be omitted.

    **Project Files:**
    Here is a list of files in the project:
    ```
    {project_files_list}
    ```

    **Contextual Understanding:**
    *   The diff shows changes in the context of the project files listed above.
    *   If understanding the relationship between the changed files and other parts of the project is necessary to write an accurate commit message, you may request the content of specific files from the list above.
    *   To request file content, respond *only* with the exact phrase: `Request content for file: <path/to/file>` where `<path/to/file>` is the relative path from the repository root. Do not add any other text to your response if you are requesting a file.

    Diff:
    ```diff
    {diff}
    ```

    Generate ONLY the commit message text, without any introductory phrases like "Here is the commit message:", unless you need to request file content.
    """

    try:
        # Get project files to include in the prompt
        project_files = get_project_files()
        project_files_list = (
            "\n".join(project_files)
            if project_files
            else "(Could not list project files)"
        )

        # Format the prompt with the diff and file list
        formatted_prompt = prompt.format(
            diff=diff, project_files_list=project_files_list
        )

        # Use a conversation history for potential back-and-forth
        conversation = [formatted_prompt]
        max_requests = 3  # Limit the number of file requests
        requests_made = 0

        while requests_made < max_requests:
            response = model.generate_content("\n".join(conversation))
            message = response.text.strip()

            # Check if the AI is requesting a file
            request_match = re.match(r"^Request content for file: (.*)$", message)

            if request_match:
                filepath = request_match.group(1).strip()
                print(f"AI requests content for: {filepath}")

                user_input = input(f"Allow access to '{filepath}'? (y/n): ").lower()

                if user_input == "y":
                    file_content = get_file_content(filepath)
                    if file_content:
                        # Provide content to AI
                        conversation.append(
                            f"Response for file '{filepath}':\n```\n{file_content}\n```\nNow, generate the commit message based on the diff and this context."
                        )
                    else:
                        # Inform AI file couldn't be read
                        conversation.append(
                            f"File '{filepath}' could not be read or was not found. Continue generating the commit message based on the original diff."
                        )
                else:
                    # Inform AI permission denied
                    conversation.append(
                        f"User denied access to file '{filepath}'. Continue generating the commit message based on the original diff."
                    )

                requests_made += 1
            else:
                # AI did not request a file, assume it's the commit message
                break  # Exit the loop
        else:
            # Max requests reached
            print(
                "Warning: Maximum number of file requests reached. Generating commit message without further context."
            )
            # Make one last attempt to generate the message without the last request fulfilled
            response = model.generate_content(
                "\n".join(conversation[:-1])
                + "\nGenerate the commit message now based on the available information."
            )  # Use conversation up to the last request
            message = response.text.strip()

        # Extract the final message, remove potential markdown code blocks, and strip whitespace
        # Ensure message is not None before processing
        if message:
            message = re.sub(
                r"^\s*```[a-zA-Z]*\s*\n?", "", message, flags=re.MULTILINE
            )  # Remove leading code block start
            message = re.sub(
                r"\n?```\s*$", "", message, flags=re.MULTILINE
            )  # Remove trailing code block end
            message = message.strip()  # Strip leading/trailing whitespace
        else:
            # Handle case where response.text might be None or empty after failed requests
            print(
                "Error: Failed to get a valid response from the AI after handling requests."
            )
            return None

        # Basic validation: Check if the message seems plausible (not empty, etc.)
        if not message or len(message) < 5:  # Arbitrary short length check
            print(
                f"Warning: Generated commit message seems too short or empty: '{message}'"
            )
            # Optionally, you could add retry logic here or return None

        return message
    except Exception as e:
        # Provide more context in the error message
        print(f"Error generating commit message with Gemini: {e}")
        # Consider logging response details if available, e.g., response.prompt_feedback
        if hasattr(response, "prompt_feedback"):
            print(f"Prompt Feedback: {response.prompt_feedback}")
        return None


def create_commit(message):
    """
    Creates a git commit with the given message.

    Args:
        message (str): The commit message.

    Returns:
        bool: True if the commit was successful, False otherwise.
    """
    if not message:
        print("Error: No commit message provided to create commit.")
        return False

    try:
        process = subprocess.run(
            ["git", "commit", "-m", message],
            check=True,  # Important: Raise exception on non-zero exit
            capture_output=True,  # capture the output
            text=True,
        )
        print(process.stdout)  # print the output
        return True
    except subprocess.CalledProcessError as e:
        print(f"Error creating git commit: {e}")
        print(e.stderr)
        return False
    except FileNotFoundError:
        print("Error: git command not found.  Is Git installed and in your PATH?")
        return False
    except Exception as e:
        print(f"An unexpected error occurred: {e}")
        return False


def main():
    """
    Main function to orchestrate the process of:
    1. Getting the staged diff.
    2. Generating a commit message using Gemini.
    3. Creating a git commit with the generated message.
    """
    gemini_api_key = os.environ.get("GEMINI_API_KEY")
    if not gemini_api_key:
        print(
            "Error: GEMINI_API_KEY environment variable not set.\n"
            "  Please obtain an API key from Google Cloud and set the environment variable.\n"
            "  For example: export GEMINI_API_KEY='YOUR_API_KEY'"
        )
        return

    diff = get_staged_diff()
    if diff is None:
        print("Aborting commit due to error getting diff.")
        return  # Exit the script

    if not diff.strip():  # check if the diff is empty
        print("Aborting: No changes staged to commit.")
        return

    message = generate_commit_message(diff, gemini_api_key)
    if message is None:
        print("Aborting commit due to error generating message.")
        return  # Exit if message generation failed

    print(f"Generated commit message:\n{message}")  # Print the message for review

    # Prompt the user for confirmation before committing
    user_input = input(
        "Do you want to create the commit with this message? (y/n): "
    ).lower()
    if user_input == "y":
        if create_commit(message):
            print("Commit created successfully.")
        else:
            print("Commit failed.")
    else:
        print("Commit aborted by user.")


if __name__ == "__main__":
    main()