app/models/patterns.py

import re
from dataclasses import dataclass, field
from typing import List, Dict, Any


@dataclass
class VulnerabilityPattern:
    name: str
    description: str
    patterns: List[str]
    severity: str          # HIGH | MEDIUM | LOW
    category: str
    recommendation: str = ""


VULNERABILITY_PATTERNS: List[VulnerabilityPattern] = [
    VulnerabilityPattern(
        name="reentrancy",
        description="Reentrancy vulnerability – external call before state update",
        patterns=[r"\.call\s*\(", r"\.send\s*\(", r"msg\.sender\.call"],
        severity="HIGH",
        category="reentrancy",
        recommendation=(
            "Use the checks-effects-interactions pattern: update state before "
            "external calls. Consider OpenZeppelin's ReentrancyGuard."
        ),
    ),
    VulnerabilityPattern(
        name="integer_overflow",
        description="Integer overflow / underflow (Solidity < 0.8)",
        patterns=[r"pragma solidity\s+\^?0\.[0-7]", r"\+\+", r"--", r"\+=", r"-=", r"\*="],
        severity="HIGH",
        category="arithmetic",
        recommendation="Use Solidity ^0.8.0 (built-in checks) or SafeMath for older versions.",
    ),
    VulnerabilityPattern(
        name="unchecked_call",
        description="External call return value not checked",
        patterns=[r"\.call\s*\([^)]*\)\s*;", r"\.send\s*\([^)]*\)\s*;"],
        severity="MEDIUM",
        category="unchecked_call",
        recommendation="Always capture and verify the boolean return of low-level calls.",
    ),
    VulnerabilityPattern(
        name="tx_origin",
        description="Authorization via tx.origin (phishing risk)",
        patterns=[r"tx\.origin"],
        severity="MEDIUM",
        category="authorization",
        recommendation="Replace tx.origin with msg.sender for access control checks.",
    ),
    VulnerabilityPattern(
        name="access_control",
        description="Public function missing access control",
        patterns=[r"function\s+\w+\s*\([^)]*\)\s*(public|external)(?!\s+view|\s+pure)"],
        severity="MEDIUM",
        category="access_control",
        recommendation="Add onlyOwner or role-based modifiers to sensitive functions.",
    ),
    VulnerabilityPattern(
        name="timestamp_dependency",
        description="Logic depends on block.timestamp (miner manipulation)",
        patterns=[r"block\.timestamp", r"\bnow\b", r"block\.number"],
        severity="LOW",
        category="timestamp",
        recommendation="Avoid using block.timestamp for critical decisions; use Chainlink VRF where randomness is needed.",
    ),
]

SEVERITY_SCORE = {"HIGH": 3, "MEDIUM": 2, "LOW": 1}


def analyze_with_patterns(solidity_code: str) -> Dict[str, Any]:
    """Run regex-based vulnerability detection and return structured report."""
    lines = solidity_code.split("\n")
    vulnerabilities: List[Dict[str, Any]] = []
    severity_score = 0

    for vp in VULNERABILITY_PATTERNS:
        finding: Dict[str, Any] = {
            "name": vp.name,
            "description": vp.description,
            "severity": vp.severity,
            "category": vp.category,
            "recommendation": vp.recommendation,
            "detected": False,
            "line_numbers": [],
            "matches": [],
        }

        for i, line in enumerate(lines, start=1):
            for pat in vp.patterns:
                for m in re.finditer(pat, line, re.IGNORECASE):
                    finding["detected"] = True
                    finding["line_numbers"].append(i)
                    finding["matches"].append(
                        {"line": i, "text": line.strip(), "match": m.group()}
                    )

        if finding["detected"]:
            finding["line_numbers"] = sorted(set(finding["line_numbers"]))
            vulnerabilities.append(finding)
            severity_score += SEVERITY_SCORE[vp.severity]

    high = [v for v in vulnerabilities if v["severity"] == "HIGH"]
    medium = [v for v in vulnerabilities if v["severity"] == "MEDIUM"]
    low = [v for v in vulnerabilities if v["severity"] == "LOW"]

    if high:
        risk_level = "CRITICAL"
    elif medium:
        risk_level = "MEDIUM"
    elif low:
        risk_level = "LOW"
    else:
        risk_level = "SAFE"

    version_match = re.search(r"pragma\s+solidity\s+([^;]+);", solidity_code)
    solidity_version = version_match.group(1).strip() if version_match else "unknown"

    return {
        "analysis_type": "pattern",
        "solidity_version": solidity_version,
        "total_lines": len(lines),
        "risk_level": risk_level,
        "severity_score": severity_score,
        "vulnerabilities_count": len(vulnerabilities),
        "vulnerabilities": vulnerabilities,
        "high_count": len(high),
        "medium_count": len(medium),
        "low_count": len(low),
    }