Building a Local AI Code Reviewer with ClientAI and Ollama

Ever wanted your own AI-powered code reviewer that runs entirely on your local machine? In this two-part tutorial, we'll build exactly that using ClientAI and Ollama.

Our assistant will analyze Python code structure, identify potential issues, and suggest improvements — all while keeping your code private and secure.

For ClientAI's docs see here and for Github Repo, here.

Series Index

• Part 1: Introduction, Setup, Tool Creation (you are here)
• Part 2: Building the Assistant and Command Line Interface

Project Overview

Our code analysis assistant will be capable of:

• Analyzing code structure and complexity
• Identifying style issues and potential problems
• Generating documentation suggestions
• Providing actionable improvement recommendations

All of this will run locally on your machine, giving you the power of AI-assisted code review while maintaining complete privacy of your code.

Setting Up Our Environment

First, create a new directory for your project:

mkdir local_task_planner
cd local_task_planner

Install ClientAI with Ollama support:

pip install clientai[ollama]

Make sure you have Ollama installed on your system. You can get it from Ollama's website.

Now let's create the file we'll write the code into:

touch code_analyzer.py

And start with our core imports:

import ast
import json
import logging
import re
from dataclasses import dataclass
from typing import List
from clientai import ClientAI
from clientai.agent import (
    Agent,
    ToolConfig,
    act,
    observe,
    run,
    synthesize,
    think,
)
from clientai.ollama import OllamaManager, OllamaServerConfig

Each of these components plays a crucial role:

• ast: Helps us understand Python code by parsing it into a tree structure
• ClientAI: Provides our AI framework
• Various utility modules for data handling and pattern matching

Structuring Our Analysis Results

When analyzing code, we need a clean way to organize our findings. Here's how we'll structure our results:

@dataclass
class CodeAnalysisResult:
    """Results from code analysis."""
    complexity: int
    functions: List[str]
    classes: List[str]
    imports: List[str]
    issues: List[str]

Think of this as our report card for code analysis:

• Complexity score indicates how intricate the code is
• Lists of functions and classes help us understand code structure
• Imports show external dependencies
• Issues track any problems we discover

Building the Core Analysis Engine

Now for the actual core — let's build our code analysis engine:

def analyze_python_code_original(code: str) -> CodeAnalysisResult:
    """Analyze Python code structure and complexity."""
    try:
        tree = ast.parse(code)
        functions = []
        classes = []
        imports = []
        complexity = 0
        for node in ast.walk(tree):
            if isinstance(node, ast.FunctionDef):
                functions.append(node.name)
                complexity += sum(
                    1
                    for _ in ast.walk(node)
                    if isinstance(_, (ast.If, ast.For, ast.While))
                )
            elif isinstance(node, ast.ClassDef):
                classes.append(node.name)
            elif isinstance(node, (ast.Import, ast.ImportFrom)):
                for name in node.names:
                    imports.append(name.name)
        return CodeAnalysisResult(
            complexity=complexity,
            functions=functions,
            classes=classes,
            imports=imports,
            issues=[],
        )
    except Exception as e:
        return CodeAnalysisResult(
            complexity=0, functions=[], classes=[], imports=[], issues=[str(e)]
        )

This function is like our code detective. It:

• Parses code into a tree structure
• Walks through the tree looking for functions, classes, and imports
• Calculates complexity by counting control structures
• Returns a comprehensive analysis result

Implementing Style Checking

Good code isn't just about working correctly — it should be readable and maintainable. Here's our style checker:

mkdir local_task_planner
cd local_task_planner

Our style checker focuses on two key aspects:

• Line length — ensuring code stays readable
• Function naming conventions — enforcing Python's preferred snake_case style

Documentation Helper

Documentation is crucial for maintainable code. Here's our documentation generator:

pip install clientai[ollama]

This helper:

• Identifies functions and classes
• Extracts parameter information
• Generates documentation templates
• Includes placeholders for examples

Making Our Tools AI-Ready

To prepare our tools for integration with the AI system, we need to wrap them in JSON-friendly formats:

touch code_analyzer.py

These wrappers add input validation, JSON serialization and error handling to make our assistant more error proof.

Coming Up in Part 2

In this post we set up our environment, structured our results, and built the functions we will use as tools for our Agent. In the next part, we'll actually create our AI assistant, register these tools, build a command-line interface and see this assistant in action.

Your next step is Part 2: Building the Assistant and Command Line Interface.

To see more about ClientAI, go to the docs.

Connect with Me

If you have any questions, want to discuss tech-related topics, or share your feedback, feel free to reach out to me on social media:

• GitHub: igorbenav
• X/Twitter: @igorbenav
• LinkedIn: Igor

The above is the detailed content of Building a Local AI Code Reviewer with ClientAI and Ollama. For more information, please follow other related articles on the PHP Chinese website!