feat: transition bestseller analysis storage to SQLite and add category blacklist

- Replaces Excel output with structured database tables for tracking category analysis runs and product results.
- Implements a blacklist to exclude specific category IDs from the bestseller pipeline.
- Adds unit tests for category processing and enhances logging with levels and timestamps.
- Introduces foreign key enforcement and updated schema definitions in the database module.

src/types.ts

@@ -73,3 +73,45 @@ export interface AnalysisResult {
   product: EnrichedProduct;
   verdict: LlmVerdict;
 }
+
+export interface CategoryRunSummaryDb {
+  categoryId: number;
+  categoryLabel: string;
+  runTimestamp: string;
+  topAsinsChecked: number;
+  availableAsins: number;
+  fbaCount: number;
+  fbmCount: number;
+  skipCount: number;
+  status: "ok" | "empty" | "failed";
+  errorMessage?: string;
+}
+
+export interface ProductAnalysisResultDb {
+  asin: string;
+  runId: number;
+  name: string;
+  brand?: string;
+  category?: string;
+  unitCost?: number;
+  currentPrice?: number;
+  avgPrice90d?: number;
+  avgPrice90dSheet?: number;
+  sellingPriceSheet?: number;
+  salesRank?: number;
+  salesRankAvg90d?: number;
+  sellerCount?: number;
+  monthlySold?: number;
+  rankDrops30d?: number;
+  rankDrops90d?: number;
+  fbaFee?: number;
+  fbmFee?: number;
+  referralPercent?: number;
+  canSell?: string;
+  sellabilityStatus?: string;
+  sellabilityReason?: string;
+  verdict: string;
+  confidence: number;
+  reasoning?: string;
+  fetchedAt: string;
+}