openai_client.py

import os
from dotenv import load_dotenv
from openai import OpenAI
import anthropic

# Load environment variables from .env file
load_dotenv()

client = OpenAI(
    api_key=os.getenv("OPENAI_API_KEY"),
    base_url=os.getenv("https://api.aimlapi.com"),
)

# Initialize the Anthropic client
anthropic_client = anthropic.Anthropic(api_key=os.getenv("ANTHROPIC_API_KEY"))

# Function to get GPT-4o Mini response
def get_code_review_response(prompt, max_tokens=1000):
    try:
        response = anthropic_client.messages.create(
            model="claude-3-5-sonnet-20240620",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": f"You are an AI assistant who helps users in code reviews by deep thinking in points max 5-6 point shortly:\n{prompt}",
                },
            ],
        )
        # Extract feedback text from the response
        review = response.text if hasattr(response, "text") else str(response)

        # Check if feedback is a Message object and extract text if necessary
        if hasattr(response, "content") and isinstance(response.content, list):
            review = "\n\n".join(
                [
                    text_block.text
                    for text_block in response.content
                    if hasattr(text_block, "text")
                ]
            )

        return review
    except Exception as e:
        return "Sorry, an error occurred while generating your idea. Please try again later."


# Function to refactor code
def refactor_code(code_snippet):
    try:
        response = anthropic_client.messages.create(
            model="claude-3-5-sonnet-20240620",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": f"Refactor the following code. Do not provide any explanation or comments, just return the refactored code:\n{code_snippet}",
                },
            ],
        )
        refactor = response.text if hasattr(response, "text") else str(response)

        # Check if feedback is a Message object and extract text if necessary
        if hasattr(response, "content") and isinstance(response.content, list):
            refactor = "\n\n".join(
                [
                    text_block.text
                    for text_block in response.content
                    if hasattr(text_block, "text")
                ]
            )

        return refactor
    except Exception as e:
        return "Sorry, an error occurred while refactoring your code. Please try again later."


# Function to get feedback on code using Anthropic
def code_feedback(code_snippet):
    try:
        response = anthropic_client.messages.create(
            model="claude-3-5-sonnet-20240620",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": f"Please provide feedback on the given code, don't refactor the code:\n{code_snippet}",
                },
            ],
        )

        # Extract feedback text from the response
        feedback = response.text if hasattr(response, "text") else str(response)

        # Check if feedback is a Message object and extract text if necessary
        if hasattr(response, "content") and isinstance(response.content, list):
            feedback = "\n\n".join(
                [
                    text_block.text
                    for text_block in response.content
                    if hasattr(text_block, "text")
                ]
            )

        return feedback

    except Exception as e:
        return "Sorry, an error occurred while getting feedback on your code. Please try again later."


# Function to suggest best coding practices based on given code
def suggest_best_practices(code_snippet):
    try:
        response = anthropic_client.messages.create(
            model="claude-3-5-sonnet-20240620",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": (
                        f"Based on the following code, suggest best practices max 5-6 point shortly"
                        f"for coding patterns that align with industry standards: \n{code_snippet}"
                    ),
                },
            ],
        )

        # Extract suggestions from the response
        best_practices = response.text if hasattr(response, "text") else str(response)

        # Check if the feedback is a Message object and extract text if necessary
        if hasattr(response, "content") and isinstance(response.content, list):
            best_practices = "\n\n".join(
                [
                    text_block.text
                    for text_block in response.content
                    if hasattr(text_block, "text")
                ]
            )

        return best_practices

    except Exception as e:
        return "Sorry, an error occurred while suggesting best practices. Please try again later."


# Function to remove code errors
def remove_code_errors(code_snippet):
    try:
        response = anthropic_client.messages.create(
            model="claude-3-5-sonnet-20240620",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": f"Identify and suggest fixes for errors in the following code:\n{code_snippet}",
                },
            ],
        )
        code_errors = response.text if hasattr(response, "text") else str(response)

        # Check if feedback is a Message object and extract text if necessary
        if hasattr(response, "content") and isinstance(response.content, list):
            code_errors = "\n\n".join(
                [
                    text_block.text
                    for text_block in response.content
                    if hasattr(text_block, "text")
                ]
            )

        return code_errors
    except Exception as e:
        return "Sorry, an error occurred while removing code errors. Please try again later."