WuYanBo
/
CodeReview
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Improve report generation and statistics by normalizing severity comparisons and leveraging backend-calculated task metrics.

This commit is contained in:
lintsinghua 2025-12-13 20:42:10 +08:00
parent d449e2ba78
commit 2a332d6eda
3 changed files with 57 additions and 75 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
backend/app/api/v1/endpoints/agent_tasks.py
View File

@ -185,7 +185,8 @@ class AgentFindingResponse(BaseModel):
code_snippet: Optional[str]
is_verified: bool
confidence: float
# 🔥 FIX: Map from ai_confidence in ORM, make Optional with default
confidence: Optional[float] = Field(default=0.5, validation_alias="ai_confidence")
status: str
suggestion: Optional[str] = None
@ -193,8 +194,10 @@ class AgentFindingResponse(BaseModel):
created_at: datetime
class Config:
from_attributes = True
model_config = {
"from_attributes": True,
"populate_by_name": True, # Allow both 'confidence' and 'ai_confidence'
}
class TaskSummaryResponse(BaseModel):
@ -2201,6 +2204,16 @@ async def generate_audit_report(
)
findings = findings.scalars().all()
# 🔥 Helper function to normalize severity for comparison (case-insensitive)
def normalize_severity(sev: str) -> str:
return str(sev).lower().strip() if sev else ""
# Log findings for debugging
logger.info(f"[Report] Task {task_id}: Found {len(findings)} findings from database")
if findings:
for i, f in enumerate(findings[:3]): # Log first 3
logger.debug(f"[Report] Finding {i+1}: severity='{f.severity}', title='{f.title[:50] if f.title else 'N/A'}'")
if format == "json":
# Enhanced JSON report with full metadata
return {
@ -2218,10 +2231,10 @@ async def generate_audit_report(
"total_findings": len(findings),
"verified_findings": sum(1 for f in findings if f.is_verified),
"severity_distribution": {
"critical": sum(1 for f in findings if f.severity == 'critical'),
"high": sum(1 for f in findings if f.severity == 'high'),
"medium": sum(1 for f in findings if f.severity == 'medium'),
"low": sum(1 for f in findings if f.severity == 'low'),
"critical": sum(1 for f in findings if normalize_severity(f.severity) == 'critical'),
"high": sum(1 for f in findings if normalize_severity(f.severity) == 'high'),
"medium": sum(1 for f in findings if normalize_severity(f.severity) == 'medium'),
"low": sum(1 for f in findings if normalize_severity(f.severity) == 'low'),
},
"agent_metrics": {
"total_iterations": task.total_iterations,
@ -2258,10 +2271,10 @@ async def generate_audit_report(
# Calculate statistics
total = len(findings)
critical = sum(1 for f in findings if f.severity == 'critical')
high = sum(1 for f in findings if f.severity == 'high')
medium = sum(1 for f in findings if f.severity == 'medium')
low = sum(1 for f in findings if f.severity == 'low')
critical = sum(1 for f in findings if normalize_severity(f.severity) == 'critical')
high = sum(1 for f in findings if normalize_severity(f.severity) == 'high')
medium = sum(1 for f in findings if normalize_severity(f.severity) == 'medium')
low = sum(1 for f in findings if normalize_severity(f.severity) == 'low')
verified = sum(1 for f in findings if f.is_verified)
with_poc = sum(1 for f in findings if f.has_poc)
@ -2323,13 +2336,13 @@ async def generate_audit_report(
md_lines.append(f"| Severity | Count | Verified |")
md_lines.append(f"|----------|-------|----------|")
if critical > 0:
md_lines.append(f"| **CRITICAL** | {critical} | {sum(1 for f in findings if f.severity == 'critical' and f.is_verified)} |")
md_lines.append(f"| **CRITICAL** | {critical} | {sum(1 for f in findings if normalize_severity(f.severity) == 'critical' and f.is_verified)} |")
if high > 0:
md_lines.append(f"| **HIGH** | {high} | {sum(1 for f in findings if f.severity == 'high' and f.is_verified)} |")
md_lines.append(f"| **HIGH** | {high} | {sum(1 for f in findings if normalize_severity(f.severity) == 'high' and f.is_verified)} |")
if medium > 0:
md_lines.append(f"| **MEDIUM** | {medium} | {sum(1 for f in findings if f.severity == 'medium' and f.is_verified)} |")
md_lines.append(f"| **MEDIUM** | {medium} | {sum(1 for f in findings if normalize_severity(f.severity) == 'medium' and f.is_verified)} |")
if low > 0:
md_lines.append(f"| **LOW** | {low} | {sum(1 for f in findings if f.severity == 'low' and f.is_verified)} |")
md_lines.append(f"| **LOW** | {low} | {sum(1 for f in findings if normalize_severity(f.severity) == 'low' and f.is_verified)} |")
md_lines.append(f"| **Total** | {total} | {verified} |")
md_lines.append("")
@ -2353,7 +2366,7 @@ async def generate_audit_report(
else:
# Group findings by severity
for severity_level, severity_name in [('critical', 'Critical'), ('high', 'High'), ('medium', 'Medium'), ('low', 'Low')]:
severity_findings = [f for f in findings if f.severity == severity_level]
severity_findings = [f for f in findings if normalize_severity(f.severity) == severity_level]
if not severity_findings:
continue

68
frontend/src/pages/AgentAudit/components/ReportExportDialog.tsx
View File

@ -104,20 +104,18 @@ function formatBytes(bytes: number): string {
// ============ Sub Components ============
// Report stats summary
// Report stats summary - uses task statistics for reliability
const ReportStats = memo(function ReportStats({
task,
findings,
}: {
task: AgentTask;
findings: AgentFinding[];
findings: AgentFinding[]; // Keep for API compatibility, but use task stats
}) {
const severityCounts = {
critical: findings.filter(f => f.severity.toLowerCase() === "critical").length,
high: findings.filter(f => f.severity.toLowerCase() === "high").length,
medium: findings.filter(f => f.severity.toLowerCase() === "medium").length,
low: findings.filter(f => f.severity.toLowerCase() === "low").length,
};
// Use task's reliable statistics instead of computing from findings array
// This ensures consistency even when findings array is empty or not loaded
const totalFindings = task.findings_count || 0;
const criticalAndHigh = (task.critical_count || 0) + (task.high_count || 0);
const verified = task.verified_count || 0;
return (
<div className="grid grid-cols-4 gap-2 mb-4">
@ -137,7 +135,7 @@ const ReportStats = memo(function ReportStats({
<span className="text-[9px] text-slate-500 uppercase tracking-wider">Findings</span>
</div>
<div className="text-lg font-bold font-mono text-white">
{findings.length}
{totalFindings}
</div>
</div>
@ -147,7 +145,7 @@ const ReportStats = memo(function ReportStats({
<span className="text-[9px] text-slate-500 uppercase tracking-wider">Critical</span>
</div>
<div className="text-lg font-bold font-mono text-rose-400">
{severityCounts.critical + severityCounts.high}
{criticalAndHigh}
</div>
</div>
@ -157,7 +155,7 @@ const ReportStats = memo(function ReportStats({
<span className="text-[9px] text-slate-500 uppercase tracking-wider">Verified</span>
</div>
<div className="text-lg font-bold font-mono text-teal-400">
{findings.filter(f => f.is_verified).length}
{verified}
</div>
</div>
</div>
@ -351,49 +349,15 @@ export const ReportExportDialog = memo(function ReportExportDialog({
setPreview(prev => ({ ...prev, loading: true, error: null }));
try {
// For JSON, we can generate it client-side
// For JSON, fetch from backend API to ensure data consistency
// The backend properly queries findings from the database
if (format === "json") {
const reportData = {
report_metadata: {
task_id: task.id,
project_id: task.project_id,
generated_at: new Date().toISOString(),
task_status: task.status,
duration_seconds: task.completed_at && task.started_at
? Math.round((new Date(task.completed_at).getTime() - new Date(task.started_at).getTime()) / 1000)
: null,
},
summary: {
security_score: task.security_score,
total_files_analyzed: task.analyzed_files,
total_findings: findings.length,
verified_findings: findings.filter(f => f.is_verified).length,
severity_distribution: {
critical: task.critical_count,
high: task.high_count,
medium: task.medium_count,
low: task.low_count,
},
},
findings: findings.map(f => ({
id: f.id,
vulnerability_type: f.vulnerability_type,
severity: f.severity,
title: f.title,
description: f.description,
file_path: f.file_path,
line_start: f.line_start,
line_end: f.line_end,
code_snippet: f.code_snippet,
is_verified: f.is_verified,
has_poc: f.has_poc,
suggestion: f.suggestion,
ai_confidence: f.ai_confidence,
})),
};
const response = await apiClient.get(`/agent-tasks/${task.id}/report`, {
params: { format: "json" },
});
setPreview({
content: JSON.stringify(reportData, null, 2),
content: JSON.stringify(response.data, null, 2),
format: "json",
loading: false,
error: null,

15
frontend/src/pages/AgentAudit/components/StatsPanel.tsx
View File

@ -8,7 +8,6 @@
import { memo } from "react";
import { Activity, FileCode, Repeat, Zap, Bug, Shield, AlertTriangle } from "lucide-react";
import { Badge } from "@/components/ui/badge";
import { calculateSeverityCounts } from "../utils";
import type { StatsPanelProps } from "../types";
// Circular progress component
@ -86,8 +85,15 @@ function MetricCard({ icon, label, value, suffix = "", colorClass = "text-slate-
export const StatsPanel = memo(function StatsPanel({ task, findings }: StatsPanelProps) {
if (!task) return null;
const severityCounts = calculateSeverityCounts(findings);
const totalFindings = Object.values(severityCounts).reduce((a, b) => a + b, 0);
// 🔥 Use task's reliable statistics instead of computing from findings array
// This ensures consistency even when findings array is empty or not loaded
const severityCounts = {
critical: task.critical_count || 0,
high: task.high_count || 0,
medium: task.medium_count || 0,
low: task.low_count || 0,
};
const totalFindings = task.findings_count || 0;
const progressPercent = task.progress_percentage || 0;
// Determine score color
@ -212,8 +218,7 @@ export const StatsPanel = memo(function StatsPanel({ task, findings }: StatsPane
color={getScoreColor(task.security_score)}
/>
<div className="absolute inset-0 flex items-center justify-center">
<span className={`text-sm font-bold font-mono ${
task.security_score >= 80 ? 'text-emerald-400' :
<span className={`text-sm font-bold font-mono ${task.security_score >= 80 ? 'text-emerald-400' :
task.security_score >= 60 ? 'text-amber-400' :
'text-rose-400'
}`}>