feat(): lru memoization for js worker

lib/memoize/index.test.ts

@@ -0,0 +1,229 @@
+/* eslint-disable
+  no-unused-vars,
+  @typescript-eslint/explicit-function-return-type,
+  @typescript-eslint/no-unused-vars-experimental,
+  jsdoc/require-param-description
+*/
+/**
+ * Ignore eslint jsdoc rules for mocked class
+ * Ignore eslint unused vars rule for decorator
+ */
+
+import { memoize } from './index';
+import Crypto from '../utils/crypto';
+
+describe('memoize decorator — per-test inline classes', () => {
+  afterEach(() => {
+    jest.useRealTimers();
+    jest.restoreAllMocks();
+    jest.clearAllMocks();
+  });
+
+  it('should memoize return value with concat strategy across several calls', async () => {
+    class Sample {
+      public calls = 0;
+
+      @memoize({ strategy: 'concat', ttl: 60_000, max: 50 })
+      public async run(a: number, b: string) {
+        this.calls += 1;
+        return `${a}-${b}`;
+      }
+    }
+
+    const sample = new Sample();
+
+    /**
+     * First call should memoize the method
+     */
+    expect(await sample.run(1, 'x')).toBe('1-x');
+    /**
+     * In this case 
+     */
+    expect(await sample.run(1, 'x')).toBe('1-x');
+    expect(await sample.run(1, 'x')).toBe('1-x');
+
+    expect(sample.calls).toBe(1);
+  });
+
+  it('should memoize return value with set of arguments with concat strategy across several calls', async () => {
+    class Sample {
+      public calls = 0;
+
+      @memoize({ strategy: 'concat' })
+      public async run(a: unknown, b: unknown) {
+        this.calls += 1;
+        return `${String(a)}|${String(b)}`;
+      }
+    }
+
+    const sample = new Sample();
+
+    /**
+     * Fill the memoization cache with values
+     */
+    await sample.run(1, 'a');
+    await sample.run(2, 'a');
+    await sample.run(1, 'b');
+    await sample.run(true, false);
+    await sample.run(undefined, null);
+
+    expect(sample.calls).toBe(5);
+
+    /**
+     * Those calls should not call the original method, they should return from memoize
+     */
+    await sample.run(1, 'a');
+    await sample.run(2, 'a');
+    await sample.run(1, 'b');
+    await sample.run(true, false);
+    await sample.run(undefined, null);
+
+    expect(sample.calls).toBe(5);
+  });
+
+  it('should memoize return value for stringified objects across several calls', async () => {
+    class Sample {
+      public calls = 0;
+      @memoize({ strategy: 'concat' })
+      public async run(x: unknown, y: unknown) {
+        this.calls += 1;
+        return 'ok';
+      }
+    }
+    const sample = new Sample();
+    const o1 = { a: 1 };
+    const o2 = { b: 2 };
+
+    await sample.run(o1, o2);
+    await sample.run(o1, o2);
+
+    expect(sample.calls).toBe(1);
+  });
+
+  it('should memoize return value for method with non-default arguments (NaN, Infinity, -0, Symbol, Date, RegExp) still cache same-args', async () => {
+    class Sample {
+      public calls = 0;
+      @memoize({ strategy: 'concat' })
+      public async run(...args: unknown[]) {
+        this.calls += 1;
+        return args.map(String).join(',');
+      }
+    }
+    const sample = new Sample();
+
+    const sym = Symbol('t');
+    const d = new Date('2020-01-01T00:00:00Z');
+    const re = /a/i;
+
+    const first = await sample.run(NaN, Infinity, -0, sym, d, re);
+    const second = await sample.run(NaN, Infinity, -0, sym, d, re);
+
+    expect(second).toBe(first);
+    expect(sample.calls).toBe(1);
+  });
+
+  it('should call crypto hash with blake2b512 algo and base64url digest, should memoize return value with hash strategy', async () => {
+    const hashSpy = jest.spyOn(Crypto, 'hash');
+
+    class Sample {
+      public calls = 0;
+      @memoize({ strategy: 'hash' })
+      public async run(...args: unknown[]) {
+        this.calls += 1;
+        return 'ok';
+      }
+    }
+    const sample = new Sample();
+
+    await sample.run({a: 1}, undefined, 0);
+    await sample.run({a: 1}, undefined, 0);
+
+    expect(hashSpy).toHaveBeenCalledWith([{a: 1}, undefined, 0], 'blake2b512', 'base64url');
+    expect(sample.calls).toBe(1);
+  });
+
+  it('should not memoize return value with hash strategy and different arguments', async () => {
+    class Sample {
+      public calls = 0;
+      @memoize({ strategy: 'hash' })
+      public async run(...args: unknown[]) {
+        this.calls += 1;
+        return 'ok';
+      }
+    }
+    const sample = new Sample();
+
+    await sample.run({ v: 1 });
+    await sample.run({ v: 2 });
+    await sample.run({ v: 3 });
+
+    expect(sample.calls).toBe(3);
+  });
+
+  it('should memoize return value with hash strategy across several calls with same args', async () => {
+    class Sample {
+      public calls = 0;
+      @memoize({ strategy: 'hash' })
+      public async run(arg: unknown) {
+        this.calls += 1;
+        return 'ok';
+      }
+    }
+    const sample = new Sample();
+
+    await sample.run({ a: 1 });
+    await sample.run({ a: 1 });
+
+    expect(sample.calls).toBe(1);
+  });
+
+  it('should memoize return value exactly for passed ttl millis', async () => {
+    jest.resetModules();
+    jest.useFakeTimers({ legacyFakeTimers: false });
+    jest.setSystemTime(new Date('2025-01-01T00:00:00Z'));
+
+    const { memoize: memoizeWithMockedTimers } = await import('../memoize/index');
+
+    class Sample {
+      public calls = 0;
+      @memoizeWithMockedTimers({ strategy: 'concat', ttl: 1_000 })
+      public async run(x: string) {
+        this.calls += 1;
+        return x;
+      }
+    }
+    const sample = new Sample();
+
+    await sample.run('k1');
+    expect(sample.calls).toBe(1);
+
+    /**
+     * Skip time beyond the ttl
+     */
+    jest.advanceTimersByTime(1_001);
+
+    await sample.run('k1');
+    expect(sample.calls).toBe(2);
+
+  });
+
+  it('error calls should never be momized', async () => {
+    class Sample {
+      public calls = 0;
+      @memoize()
+      public async run(x: number) {
+        this.calls += 1;
+        if (x === 1) throw new Error('boom');
+        return x * 2;
+      }
+    }
+    const sample = new Sample();
+
+    /**
+     * Compute with throw
+     */
+    await expect(sample.run(1)).rejects.toThrow('boom');
+    await expect(sample.run(1)).rejects.toThrow('boom');
+    expect(sample.calls).toBe(2);
+  });
+});

lib/memoize/index.ts

@@ -0,0 +1,98 @@
+import LRUCache from 'lru-cache';
+import Crypto from '../utils/crypto';
+
+/**
+ * Pick the strategy of cache key form
+ * It could be concatenated list of arguments like 'projectId:eventId'
+ * Or it could be hashed json object — blake2b512 algorithm
+ */
+export type MemoizeKeyStrategy = 'concat' | 'hash';
+
+/**
+ * Options of the memoize decorator
+ */
+export interface MemoizeOptions {
+  /**
+   * Max number of values stored in LRU cache at the same time
+   */
+  max?: number;
+
+  /**
+   * TTL in milliseconds
+   */
+  ttl?: number;
+
+  /**
+   * Strategy for key generation
+   */
+  strategy?: MemoizeKeyStrategy;
+}
+
+/**
+ * Async-only, per-method LRU-backed memoization decorator.
+ * Cache persists for the lifetime of the class instance (e.g. worker).
+ *
+ * @param options
+ */
+export function memoize(options: MemoizeOptions = {}): MethodDecorator {
+  /* eslint-disable @typescript-eslint/no-magic-numbers */
+  const {
+    max = 50,
+    ttl = 1000 * 60 * 30,
+    strategy = 'concat',
+  } = options;
+  /* eslint-enable */
+
+  return function (
+    _target,
+    propertyKey,
+    descriptor: PropertyDescriptor
+  ): PropertyDescriptor {
+    const originalMethod = descriptor.value;
+
+    if (typeof originalMethod !== 'function') {
+      throw new Error('@Memoize can only decorate methods');
+    }
+
+    descriptor.value = async function (...args: unknown[]): Promise<unknown> {
+      /**
+       * Create a cache key for each decorated method
+       */
+      const cacheKey = `memoizeCache:${String(propertyKey)}`;
+
+      /**
+       * Create a new cache if it does not exists yet (for certain function)
+       */
+      const cache: LRUCache<string, any> = this[cacheKey] ??= new LRUCache<string, any>({
+        max,
+        maxAge: ttl,
+      });
+
+      const key = strategy === 'hash'
+        ? Crypto.hash(args, 'blake2b512', 'base64url')
+        : args.map((arg) => JSON.stringify(arg)).join('__ARG_JOIN__');
+
+      /**
+       * Check if we have a cached result
+       */
+      const cachedResult = cache.get(key);
+
+      if (cachedResult !== undefined) {
+        return cachedResult;
+      }
+
+      try {
+        const result = await originalMethod.apply(this, args);
+
+        cache.set(key, result);
+
+        return result;
+      } catch (err) {
+        cache.del(key);
+        throw err;
+      }
+    };
+
+    return descriptor;
+  };
+}

lib/utils/crypto.ts

@@ -1,4 +1,4 @@
-import crypto from 'crypto';
+import crypto, { BinaryToTextEncoding } from 'crypto';
 
 /**
  * Crypto helper
@@ -9,12 +9,13 @@ export default class Crypto {
    *
    * @param value — data to be hashed
    * @param algo — type of algorithm to be used for hashing
+   * @param digest - type of the representation of the hashed value
    */
-  public static hash(value: unknown, algo = 'sha256'): string {
-    const stringifiedValue = JSON.stringify(value);
+  public static hash(value: unknown, algo = 'sha256', digest: BinaryToTextEncoding = 'hex'): string {
+    const stringifiedValue = typeof value === 'string' ? value : JSON.stringify(value);
 
     return crypto.createHash(algo)
       .update(stringifiedValue)
-      .digest('hex');
+      .digest(digest);
   }
 }

tsconfig.json

@@ -58,7 +58,7 @@
         // "inlineSources": true,                 /* Emit the source alongside the sourcemaps within a single file; requires '--inlineSourceMap' or '--sourceMap' to be set. */
 
         /* Experimental Options */
-        // "experimentalDecorators": true,        /* Enables experimental support for ES7 decorators. */
+        "experimentalDecorators": true,        /* Enables experimental support for ES7 decorators. */
         // "emitDecoratorMetadata": true,         /* Enables experimental support for emitting type metadata for decorators. */
 
         /* Advanced Options */

workers/javascript/package.json

@@ -1,6 +1,6 @@
 {
   "name": "hawk-worker-javascript",
-  "version": "0.0.1",
+  "version": "0.1.0",
   "description": "Handles messages from JavaScript Catcher",
   "main": "src/index.ts",
   "license": "UNLICENSED",
@@ -10,7 +10,8 @@
     "@types/useragent": "^2.1.1",
     "source-map-js": "^1.2.0",
     "ts-node": "^8.3.0",
-    "typescript": "^3.5.3"
+    "typescript": "^3.5.3",
+    "lodash.clonedeep": "^4.5.0"
   },
   "dependencies": {
     "useragent": "^2.3.0"