[ENHANCEMENT] Replace magic strings with Enums for severity levels and LLM providers

[advaitpatel]

Overview

Severity levels (`CRITICAL`, `HIGH`, `MEDIUM`, `LOW`, `UNKNOWN`) and LLM provider names (`openai`, `anthropic`, `google`, `ollama`) are repeated as raw strings in at least 8 files. This causes:

• Typo-prone comparisons — a typo like `"CRITCAL"` silently passes validation
• No IDE autocomplete — contributors can't discover valid values
• Scattered validation — the same list of valid providers appears in `utils.py`, `config_manager.py`, and `docksec.py` independently

Proposed Solution

Add an `enums.py` module (or add to `config.py`) with two enums:

# docksec/enums.py
from enum import Enum

class Severity(str, Enum):
    CRITICAL = "CRITICAL"
    HIGH = "HIGH"
    MEDIUM = "MEDIUM"
    LOW = "LOW"
    UNKNOWN = "UNKNOWN"

    @classmethod
    def values(cls) -> list[str]:
        return [e.value for e in cls]

    @classmethod
    def scored_levels(cls) -> list["Severity"]:
        """Severities that affect the security score."""
        return [cls.CRITICAL, cls.HIGH, cls.MEDIUM, cls.LOW]


class LLMProvider(str, Enum):
    OPENAI = "openai"
    ANTHROPIC = "anthropic"
    GOOGLE = "google"
    OLLAMA = "ollama"

    @classmethod
    def values(cls) -> list[str]:
        return [e.value for e in cls]

Files to Update

File	Change
`config.py`	Import `Severity`, replace string literals
`config_manager.py`	Replace provider/severity string lists with Enum
`docker_scanner.py`	Replace all severity string comparisons
`score_calculator.py`	Use `Severity` enum in score weights dict
`utils.py`	Replace provider validation with `LLMProvider`
`docksec.py`	Use `LLMProvider.values()` in argparse choices
`tests/`	Update tests to use enum values

Example Refactor

# Before (docker_scanner.py)
if severity in ["CRITICAL", "HIGH", "MEDIUM", "LOW"]:
    ...

SEVERITY_WEIGHTS = {"CRITICAL": 10, "HIGH": 5, "MEDIUM": 2, "LOW": 1}

# After
from docksec.enums import Severity

if severity in Severity.values():
    ...

SEVERITY_WEIGHTS = {
    Severity.CRITICAL: 10,
    Severity.HIGH: 5,
    Severity.MEDIUM: 2,
    Severity.LOW: 1,
}

Acceptance Criteria

• `Severity` and `LLMProvider` enums created
• All magic string comparisons replaced across the codebase
• All existing tests still pass (no behavior change, refactor only)
• `argparse` provider choices sourced from `LLMProvider.values()`

Skill Level

Beginner Python. Pure refactor — no logic changes required.

Notes

Since `Severity` and `LLMProvider` extend `str`, they are backwards-compatible with existing string comparisons and JSON serialization — no behavior changes needed anywhere.