benchmarks

.gitignore

@@ -2,4 +2,5 @@ node_modules
 dist
 bun.lock
 .claude
-CLAUDE.md
+CLAUDE.md
+.idea

benchmarks/.env.example

@@ -0,0 +1,7 @@
+# Prisma Benchmarks Configuration
+
+# Note: Prisma 7 uses prisma.config.ts for database URL configuration
+# This .env file is kept for any other environment-specific settings
+
+# Example: Custom configuration (if needed)
+# NODE_ENV=production

benchmarks/.gitignore

@@ -0,0 +1,13 @@
+# Benchmark artifacts
+*.db
+*.db-shm
+*.db-wal
+results/
+node_modules/
+
+# Environment
+.env
+
+# Prisma
+migrations/
+generated/

benchmarks/README.md

@@ -0,0 +1,312 @@
+# Prisma Adapter Benchmarks
+
+This benchmark suite compares the performance of the Bun SQLite adapter (`@synapsenwerkstatt/prisma-bun-sqlite-adapter`) against Prisma's official LibSQL adapter (`@prisma/adapter-libsql`).
+
+**Built for Prisma 7** with `prisma.config.ts` configuration.
+
+## Methodology
+
+The benchmarking approach follows [Prisma's official methodology](https://www.prisma.io/blog/performance-benchmarks-comparing-query-latency-across-typescript-orms-and-databases):
+
+- **Query execution timing**: Uses `performance.now()` before and after each query
+- **Multiple iterations**: Each query runs 500 times (configurable) to reduce variance
+- **Outlier removal**: Removes measurements above the 99th percentile
+- **Deterministic data**: Uses faker.js with a fixed seed for reproducible datasets
+- **Comprehensive metrics**: Tracks median, mean, min, max, percentiles (P50-P99), and standard deviation
+
+## Installation
+
+Install the required dependencies:
+
+```bash
+# Install benchmark dependencies
+bun add -d @faker-js/faker
+
+# Install LibSQL adapter for comparison (optional)
+bun add -d @prisma/adapter-libsql @libsql/client
+```
+
+## Database Schema
+
+The benchmark uses a realistic e-commerce schema with 8 models:
+- **User** - Basic user information
+- **Profile** - User profile (1:1 with User)
+- **Post** - Blog posts with many-to-many tags
+- **Category** - Post categories
+- **Tag** - Post tags (many-to-many)
+- **Order** - User orders
+- **OrderItem** - Order line items
+- **Product** - Product catalog
+
+## Setup
+
+### 1. Configuration
+
+Prisma 7 uses `prisma.config.ts` for configuration. The benchmark database URL is configured there:
+
+```typescript
+// benchmarks/prisma.config.ts
+import { defineConfig } from 'prisma/config'
+
+export default defineConfig({
+  schema: 'schema.prisma',
+  migrations: {
+    path: 'migrations',
+  },
+  datasource: {
+    url: 'file:./benchmark.db',
+  },
+})
+```
+
+### 2. Generate Prisma Client
+
+```bash
+cd benchmarks
+bunx prisma generate
+```
+
+### 3. Create and Seed Database
+
+```bash
+# Create database and run migrations
+bunx prisma migrate dev --name init
+
+# Seed with test data (1000 users, 3000 posts, 5000 orders, etc.)
+bun seed.ts
+```
+
+## Running Benchmarks
+
+### Benchmark Single Adapter (Bun SQLite)
+
+```bash
+# Run full benchmark suite (500 iterations per query)
+bun runner.ts --seed
+
+# Quick benchmark (6 selected queries)
+bun runner.ts --quick
+
+# Custom iteration count
+bun runner.ts --iterations=1000 --warmup=20
+```
+
+### Compare Both Adapters
+
+```bash
+# Compare Bun SQLite vs LibSQL
+bun compare-adapters.ts --seed
+
+# Quick comparison
+bun compare-adapters.ts --quick
+
+# Custom iterations
+bun compare-adapters.ts --iterations=1000 --warmup=20
+```
+
+### Command Line Options
+
+- `--seed` - Drop and reseed database before benchmarking
+- `--quick` - Run only 6 selected queries instead of full suite
+- `--iterations=N` - Number of iterations per query (default: 500)
+- `--warmup=N` - Number of warmup iterations (default: 10)
+
+## Query Suite
+
+The benchmark includes 27 queries covering various patterns:
+
+### Simple Queries
+- `findMany-users-simple` - Select all users
+- `findMany-users-limit` - Select with limit
+- `findUnique-user` - Find by ID
+
+### Filtered Queries
+- `findMany-posts-published` - Boolean filter
+- `findMany-orders-status` - String filter
+- `findMany-products-price-range` - Numeric range filter
+
+### Queries with Relations
+- `findMany-users-with-profile` - Single relation
+- `findMany-posts-with-author` - Inverse relation
+- `findMany-posts-with-relations` - Multiple relations
+- `findMany-users-with-all-relations` - Complex nested relations
+- `findUnique-user-nested` - Deeply nested includes
+
+### Aggregations
+- `aggregate-users-count` - Simple count
+- `aggregate-orders-sum` - Sum aggregation
+- `aggregate-products-stats` - Multiple aggregations
+- `groupBy-posts-by-category` - Group by with count
+- `groupBy-orders-by-status` - Group by with sum
+
+### Complex Queries
+- `findMany-posts-complex-filter` - Multiple conditions with OR
+- `findMany-orders-with-items-and-products` - Multiple nested relations
+
+### Write Operations
+- `create-user` - Simple create
+- `update-user` - Simple update
+- `create-post-with-relations` - Create with relations
+
+### Raw Queries
+- `raw-query-simple` - Basic SQL
+- `raw-query-join` - SQL with joins
+
+### Transactions
+- `transaction-simple` - Transaction with multiple queries
+
+## Results
+
+Results are saved to `benchmarks/results/` with timestamps:
+
+- **JSON format**: `results-YYYY-MM-DDTHH-MM-SS.json` - Raw data for analysis
+- **Markdown report**: Auto-generated with comparison tables and summary
+
+### Visualize Results
+
+```bash
+# Generate visualizations from saved results
+bun visualize.ts results/results-2024-01-15T10-30-00.json
+```
+
+This generates:
+- Comparison tables showing performance differences
+- ASCII bar charts for selected queries
+- Markdown report with detailed analysis
+
+## Understanding Results
+
+### Metrics Explained
+
+- **Median**: Middle value, less affected by outliers (primary metric)
+- **Mean**: Average of all measurements
+- **Min/Max**: Fastest and slowest measurements
+- **P50-P99**: Percentiles showing distribution
+- **Std Dev**: Variation in measurements
+
+### Interpreting Comparisons
+
+The comparison output shows:
+- **Absolute timings**: Raw latency for each adapter
+- **Percentage difference**: How much faster/slower one adapter is
+- **Visual indicators**: 🟢 (faster), 🔴 (slower), ⚪ (same)
+
+Example output:
+```
+Bun SQLite - findMany-users-limit
+  Median: 2.45ms
+
+LibSQL - findMany-users-limit
+  Median: 3.12ms
+
+Comparison: +27.35% slower
+```
+
+## Dataset Configuration
+
+Edit the `CONFIG` object in `seed.ts` to adjust data volume:
+
+```typescript
+const CONFIG = {
+  users: 1000,           // Number of users
+  postsPerUser: 3,       // Posts per user
+  categoriesCount: 20,   // Total categories
+  tagsCount: 50,         // Total tags
+  tagsPerPost: 3,        // Tags per post
+  ordersPerUser: 5,      // Orders per user
+  productsCount: 200,    // Total products
+  itemsPerOrder: 4,      // Items per order
+}
+```
+
+Default dataset:
+- 1,000 users
+- 1,000 profiles
+- 3,000 posts
+- 5,000 orders
+- 20,000 order items
+- 200 products
+- 50 tags
+
+## Troubleshooting
+
+### "Module not found: @prisma/client"
+
+Run Prisma generate first:
+```bash
+cd benchmarks
+bunx prisma generate
+```
+
+### "Table does not exist"
+
+Run migrations:
+```bash
+bunx prisma migrate dev --name init
+```
+
+### "Cannot find module 'prisma/config'"
+
+Make sure you're using Prisma 7:
+```bash
+bunx prisma --version  # Should show 7.x.x
+bun install            # Update dependencies
+```
+
+### "Cannot find module @prisma/adapter-libsql"
+
+The LibSQL adapter is optional. To use it:
+```bash
+bun add -d @prisma/adapter-libsql @libsql/client
+```
+
+Or run only the Bun adapter benchmark:
+```bash
+bun runner.ts
+```
+
+### Out of Memory
+
+Reduce iterations or dataset size:
+```bash
+bun runner.ts --iterations=100
+```
+
+Or edit `seed.ts` to reduce data volume.
+
+## Requirements
+
+- **Bun**: >= 1.0.0
+- **Node.js**: >= 20.19.0 (required by Prisma 7)
+- **TypeScript**: >= 5.4.0
+- **Prisma**: >= 7.0.0
+- **@prisma/client**: >= 7.0.0
+- **@prisma/driver-adapter-utils**: >= 7.0.0
+
+### Compatibility
+
+| Prisma Version | Adapter Version | Status |
+|----------------|-----------------|--------|
+| 7.0.x+         | 2.0.x          | ✅ Stable |
+| 6.13.x+        | 1.x.x          | ✅ Legacy |
+
+## Contributing
+
+To add new queries:
+
+1. Add to `queries.ts`:
+```typescript
+{
+  name: 'my-new-query',
+  description: 'Description of what it tests',
+  query: async (prisma) => {
+    return prisma.model.findMany({ ... })
+  },
+}
+```
+
+2. Run benchmarks to test
+
+## License
+
+MIT