gollm

Core API Reference

GollmCore

Main orchestrator class for goLLM functionality.

Initialization

from gollm import GollmCore

# Initialize with default configuration
gollm = GollmCore()

# Initialize with custom config path
gollm = GollmCore(config_path="path/to/config.json")

Methods

validate_file(file_path: str) -> Dict

Validate a single file against configured rules.

Parameters:

• file_path: Path to the file to validate

Returns:

{
    "valid": bool,
    "violations": List[Violation],
    "score": float,
    "summary": Dict[str, int]
}

validate_project(directory: str = ".") -> Dict

Validate all Python files in a directory.

Parameters:

• directory: Directory path to scan (default: current directory)

Returns:

{
    "valid": bool,
    "files": Dict[str, FileResult],
    "summary": Dict[str, int],
    "score": float
}

async handle_code_generation(prompt: str, context: Dict = None) -> Dict

Generate code using the configured LLM.

Parameters:

• prompt: User’s code generation request
• context: Additional context (optional)

Returns:

{
    "code": str,
    "explanation": str,
    "valid": bool,
    "score": float
}

CodeValidator

Validates Python code against quality rules.

Initialization

from gollm.validation.validators import CodeValidator
from gollm.config.config import GollmConfig

config = GollmConfig.load()
validator = CodeValidator(config)

Methods

validate_file(file_path: str) -> Dict

Validate a single file.

validate_project(directory: str = ".") -> Dict

Validate all Python files in a directory.

validate_code(code: str, filename: str = "<string>") -> Dict

Validate a string containing Python code.

LLMOrchestrator

Manages interactions with LLM providers.

Initialization

from gollm.llm.orchestrator import LLMOrchestrator
from gollm.config.config import GollmConfig

config = GollmConfig.load()
orchestrator = LLMOrchestrator(config)

Methods

async generate_code(prompt: str, context: Dict = None) -> Dict

Generate code based on the given prompt.

Parameters:

• prompt: Code generation prompt
• context: Additional context (optional)

Returns:

{
    "code": str,
    "explanation": str,
    "metadata": Dict
}

async chat(messages: List[Dict]) -> Dict

Have a conversation with the LLM.

Parameters:

• messages: List of message dictionaries with ‘role’ and ‘content’

Returns:

{
    "response": str,
    "metadata": Dict
}

FileUtils

Utility functions for file operations.

Methods

backup_file(file_path: str) -> str

Create a backup of a file.

safe_write(file_path: str, content: str) -> None

Safely write content to a file with backup.

get_file_hash(file_path: str) -> str

Calculate MD5 hash of a file.

Violation

Represents a code quality violation.

Attributes

• rule_id: str - Identifier of the violated rule
• message: str - Description of the violation
• filename: str - Path to the file
• line: int - Line number
• column: int - Column number
• severity: str - One of: ‘error’, ‘warning’, ‘info’
• context: str - Code context around the violation

Configuration

See Configuration Reference for detailed configuration options.

Error Handling

All API methods raise specific exceptions:

• GollmError: Base exception class
• ValidationError: Raised for validation failures
• LLMError: Raised for LLM-related errors
• ConfigError: Raised for configuration issues

Example error handling:

try:
    result = gollm.validate_file("myfile.py")
except gollm.ValidationError as e:
    print(f"Validation failed: {e}")
except gollm.GollmError as e:
    print(f"Error: {e}")

Logging

Configure logging using Python’s standard logging module:

import logging

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger("gollm")

This site is open source. Improve this page.