perf: memoise regex compilation in the classifier

Classify runs on every proxied request and recompiled each remote's
pattern lists every time. Cache compiled regexes keyed by pattern text so
each distinct pattern compiles once and is reused thereafter.

Refs #73

internal/database/stats.go

@@ -30,6 +30,15 @@ func (db *DB) GetOverviewStats(ctx context.Context) (*models.OverviewStats, erro
 		return nil, err
 	}
 
+	err = db.Pool.QueryRow(ctx, `
+		SELECT COALESCE(SUM(size_bytes), 0)
+		FROM access_log
+		WHERE cache_hit = TRUE AND created_at > NOW() - INTERVAL '30 days'
+	`).Scan(&stats.BandwidthSaved30d)
+	if err != nil {
+		return nil, err
+	}
+
 	return &stats, nil
 }
 

internal/proxy/classifier.go

@@ -2,6 +2,7 @@ package proxy
 
 import (
 	"regexp"
+	"sync"
 
 	"git.unkin.net/unkin/artifactapi/internal/provider"
 	"git.unkin.net/unkin/artifactapi/pkg/models"
@@ -60,10 +61,29 @@ func (c *Classifier) Classify(remote models.Remote, path string) Classification
 	return ClassImmutable
 }
 
+// patternCache memoises regex compilation. Classify runs on every proxied
+// request and previously recompiled each remote's pattern lists every time;
+// keying by the pattern string lets each distinct pattern compile once and
+// then be reused, with no invalidation needed (the pattern text is the key).
+// A pattern that fails to compile is cached as a typed nil so we don't retry.
+var patternCache sync.Map // map[string]*regexp.Regexp
+
+func compileCached(pattern string) *regexp.Regexp {
+	if v, ok := patternCache.Load(pattern); ok {
+		return v.(*regexp.Regexp)
+	}
+	re, err := regexp.Compile(pattern)
+	if err != nil {
+		re = nil
+	}
+	patternCache.Store(pattern, re)
+	return re
+}
+
 func compilePatterns(patterns []string) []*regexp.Regexp {
 	compiled := make([]*regexp.Regexp, 0, len(patterns))
 	for _, p := range patterns {
-		if re, err := regexp.Compile(p); err == nil {
+		if re := compileCached(p); re != nil {
 			compiled = append(compiled, re)
 		}
 	}

ui/src/pages/Dashboard.tsx

@@ -50,6 +50,11 @@ export function Dashboard() {
           value={formatNumber(stats.total_blobs_deduped)}
           sub="shared blobs"
         />
+        <StatsCard
+          label="Bandwidth Saved"
+          value={formatBytes(stats.bandwidth_saved_30d)}
+          sub="last 30 days"
+        />
       </div>
 
       {health && (