@dan-uni/dan-any-plugin-detaolu 0.0.2

package/dist/index.umd.min.js.LICENSE.txt

@@ -0,0 +1,28 @@
+/*!
+ * The buffer module from node.js, for the browser.
+ *
+ * @author   Feross Aboukhadijeh <feross@feross.org> <http://feross.org>
+ * @license  MIT
+ */
+
+/*!
+ * The buffer module from node.js, for the browser.
+ *
+ * @author   Feross Aboukhadijeh <https://feross.org>
+ * @license  MIT
+ */
+
+/*! For license information please see index.umd.min.js.LICENSE.txt */
+
+/*! ieee754. BSD-3-Clause License. Feross Aboukhadijeh <https://feross.org/opensource> */
+
+/*! safe-buffer. MIT License. Feross Aboukhadijeh <https://feross.org/opensource> */
+
+/*! zero-fill. MIT License. Feross Aboukhadijeh <https://feross.org/opensource> */
+
+/**
+ * @author: xmcp(代码主要逻辑来源)
+ * @see: https://github.com/xmcp/pakku.js
+ * @license: GPL-3.0
+ * 本文件内代码来源见上，经部分修改，并整合config注释
+ */

package/dist/pakku.js/index.d.ts

@@ -0,0 +1,47 @@
+/**
+ * @author: xmcp(代码主要逻辑来源)
+ * @see: https://github.com/xmcp/pakku.js
+ * @license: GPL-3.0
+ * 本文件内代码来源见上，经部分修改，并整合config注释
+ */
+import type { DanmuChunk, DanmuClusterOutput, DanmuObject } from './types';
+export declare const DEFAULT_CONFIG: {
+    /**
+     * 时间阈值(合并n秒内的弹幕)：
+     * 超长（大概 60 秒以上？）的阈值可能会导致程序运行缓慢
+     */
+    THRESHOLD: number;
+    /**
+     * 编辑距离合并阈值：
+     * 根据编辑距离判断不完全一致但内容相近（例如有错别字）的弹幕,
+     * 能有效击杀 "<code>你指尖跃动的电光</code>" 和 "<code>你<b>之间</b>跃动的电光</code>" 等
+     * @example 禁用(0),轻微(≤3),中等(≤5),强力(≤8)
+     */
+    MAX_DIST: number;
+    /**
+     * 词频向量合并阈值：
+     * 根据 2-Gram 频率向量的夹角判断不完全一致但内容类似的弹幕,
+     * 能有效击杀 "<code>yeah!~</code>" 和 "<code>yeah!~yeah!~yeah!~yeah!~</code>" 等
+     * @example 禁用(1000),轻微(60%),中等(45%),强力(30%)
+     */
+    MAX_COSINE: number;
+    /**
+     * 识别谐音弹幕：
+     * 将常用汉字转换为拼音再进行比较,
+     * 能有效击杀 "<code>布拉迪巴特福来</code>" 和 "<code>布拉迪·八德福莱</code>" 等
+     */
+    TRIM_PINYIN: boolean;
+    TRIM_ENDING: boolean;
+    TRIM_SPACE: boolean;
+    TRIM_WIDTH: boolean;
+    FORCELIST: string[][];
+    WHITELIST: [string, string][];
+    BLACKLIST: [string, string][];
+    CROSS_MODE: boolean;
+    PROC_TYPE7: boolean;
+    PROC_TYPE4: boolean;
+    PROC_POOL1: boolean;
+};
+export type Config = Partial<typeof DEFAULT_CONFIG>;
+declare function merge(chunk: DanmuChunk<DanmuObject>, config?: Config): Promise<DanmuClusterOutput>;
+export default merge;

package/dist/pakku.js/similarity_stub.d.ts

@@ -0,0 +1,9 @@
+import type { Config } from '.';
+import type { int, Stats } from './types';
+export declare function init(wasm_module: ArrayBuffer): Promise<void>;
+export declare function begin_chunk(config: Config): void;
+export declare function begin_index_lock(): void;
+export declare function detect_similarity(str: string, mode: number, index_l: int, S: Stats): null | {
+    reason: string;
+    idx_diff: int;
+};

package/dist/pakku.js/types.d.ts

@@ -0,0 +1,70 @@
+import type { UniDM } from '@dan-uni/dan-any';
+export type int = number;
+export type float = number;
+export type AnyObject = {
+    [k: string]: any;
+};
+export interface DanmuObject {
+    time_ms: int;
+    mode: int;
+    content: string;
+    pool: int;
+    danuni_dan: UniDM;
+}
+export interface DanmuObjectPeer extends DanmuObject {
+    pakku: {
+        sim_reason: string;
+    };
+}
+export interface DanmuObjectRepresentative extends DanmuObject {
+    pakku: {
+        peers: DanmuObjectPeer[];
+        desc: string[];
+        disp_str: string;
+    };
+}
+export interface DanmuChunk<ObjectType extends DanmuObject> {
+    objs: ObjectType[];
+}
+export interface DanmuCluster {
+    peers: DanmuObjectPeer[];
+    desc: string[];
+    chosen_str: string;
+}
+export interface DanmuClusterPtr {
+    peers_ptr: [int, string][];
+    desc: string[];
+    chosen_str: string;
+    danuni_count: int;
+    danuni_dans: DanmuObject[];
+}
+export interface DanmuClusterOutput {
+    clusters: DanmuClusterPtr[];
+    stats: Stats;
+}
+export declare class Stats {
+    combined_identical: number;
+    combined_edit_distance: number;
+    combined_pinyin_distance: number;
+    combined_cosine_distance: number;
+    deleted_blacklist: number;
+    deleted_blacklist_each: {
+        [k: string]: int;
+    };
+    ignored_whitelist: number;
+    ignored_type: number;
+    num_taolu_matched: number;
+}
+export declare class Queue<T> {
+    storage: {
+        [k: int]: T;
+    };
+    index_l: int;
+    index_r: int;
+    constructor(init?: T[]);
+    push(item: T): void;
+    pop(): void;
+    peek(): T | null;
+    size(): number;
+    [Symbol.iterator](): IterableIterator<T>;
+}

package/package.json

@@ -0,0 +1,39 @@
+{
+  "name": "@dan-uni/dan-any-plugin-detaolu",
+  "version": "0.0.2",
+  "description": "A filter, dedupe and anti-spam plugin of dan-any, a danmaku transformer lib, based on pakku.js.",
+  "keywords": [
+    "bangumi",
+    "danmaku"
+  ],
+  "license": "GPL-3.0-or-later",
+  "homepage": "https://github.com/ani-uni/danuni/tree/master/packages/dan-any-plugin-detaolu#readme",
+  "bugs": {
+    "url": "https://github.com/ani-uni/danuni/issues"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/ani-uni/danuni.git"
+  },
+  "author": "rinne",
+  "main": "dist/index.js",
+  "module": "dist/index.js",
+  "types": "dist/src/index.d.ts",
+  "browser": "dist/index.umd.min.js",
+  "scripts": {
+    "test": "vitest",
+    "test:ui": "vitest --ui",
+    "coverage": "vitest run --coverage",
+    "bench": "vitest bench",
+    "lint": "eslint",
+    "build": "rslib build",
+    "build:types": "tsc"
+  },
+  "dependencies": {
+    "@dan-uni/dan-any": "workspace:^",
+    "fs-extra": "^11.3.0"
+  },
+  "devDependencies": {
+    "@types/fs-extra": "^11.0.4"
+  }
+}

package/rslib.config.ts

@@ -0,0 +1,27 @@
+import { pluginNodePolyfill } from '@rsbuild/plugin-node-polyfill'
+import { defineConfig } from '@rslib/core'
+
+import pkg from './package.json'
+
+export default defineConfig({
+  lib: [
+    {
+      format: 'esm',
+      output: {
+        filename: { js: 'index.js' },
+        target: 'node',
+      },
+      dts: true,
+    },
+    {
+      format: 'umd',
+      output: {
+        filename: { js: 'index.umd.min.js' },
+        target: 'web',
+      },
+      dts: true,
+      umdName: pkg.name,
+      plugins: [pluginNodePolyfill()],
+    },
+  ],
+})

package/src/index.test.ts

@@ -0,0 +1,39 @@
+//基于以下注释，根据vitest生成测试用例
+import { describe, it } from 'vitest'
+
+import { UniPool } from '@dan-uni/dan-any'
+
+import detaolu from '.'
+
+const xml = `<i>
+<chatserver>chat.bilibili.com</chatserver>
+<chatid>1156756312</chatid>
+<mission>0</mission>
+<maxlimit>2947</maxlimit>
+<state>0</state>
+<real_name>0</real_name>
+<source>k-v</source>
+<d p="13.213,1,25,16777215,1686314041,3,ff41173d,1335658005672492032">喜欢</d>
+<d p="13.213,1,25,16777215,1686590010,0,296b35b5,1337972999512832512">来了 哈哈~~</d>
+<d p="13.246,1,25,16777215,1686276875,0,5664cfc4,1335346233459549696">就是</d>
+<d p="13.266,1,25,16777215,1686283375,0,c7e6646f,1335400761013670912">什么鬼？</d>
+<d p="13.284,1,25,16777215,1686291338,0,38662881,1335467554877267456">哇哦</d>
+<d p="13.306,1,25,16777215,1686268410,0,4c01de10,1335275224983600896">试试</d>
+<d p="13.331,1,25,16777215,1686948453,3,56a3c5d5,1340979831550069760">不喜欢</d>
+<d p="13.374,1,25,16777215,1686300770,3,647fe355,1335546672880933888">不喜欢</d>
+<d p="13.376,1,25,16777215,1686297921,0,469d94b8,1335522778300134400">哦豁</d>
+<d p="13.419,1,25,8700107,1686268005,0,be402447,1335271828100244224">太酷啦</d>
+<d p="13.419,1,25,16777215,1686316828,3,7ffb6619,1335681385016736768">喜欢</d>
+<d p="13.459,1,25,16777215,1686299729,0,45834405,1335537942797634048">一般，不好看</d>
+<d p="13.462,1,25,16777215,1686302133,0,3cab672c,1335558106620590080">哈哈哈</d>
+<d p="13.481,1,25,16777215,1686297342,0,ce67fafd,1335517923728804864">？</d>
+<d p="13.499,1,25,16777215,1686301548,3,2848bf1c,1335553202649003264">不喜欢</d>
+</i>`
+
+describe('其它', () => {
+  const pool = UniPool.fromBiliXML(xml)
+  it('反套路(基于pakku.js实现的弹幕去重)', async () => {
+    const d = (await pool.pipe(detaolu({ MAX_COSINE: 1000 }))).minify()
+    console.info(d)
+  })
+})

package/src/index.ts

@@ -0,0 +1,70 @@
+import type { Config as DeTaoLuConfig } from './pakku.js'
+
+import { UniDM, UniDMTools, UniPool } from '@dan-uni/dan-any'
+
+import pakkujs from './pakku.js'
+
+async function detaolu(that: UniPool, config?: DeTaoLuConfig) {
+  const p = await pakkujs(
+    {
+      objs: that.dans.map((d) => ({
+        time_ms: d.progress * 1000,
+        mode: d.mode, //TODO
+        content: d.content,
+        pool: d.pool,
+        // danuni_sender: d.senderID,
+        danuni_dan: d,
+      })),
+    },
+    config,
+  )
+  const selected = p.clusters.map((p) => {
+    if (p.danuni_dans.length === 1) {
+      return p.danuni_dans[0].danuni_dan
+    } else {
+      const dans = p.danuni_dans,
+        pool = new UniPool(dans.map((d) => d.danuni_dan))
+      function isAllBottomMode(p: UniPool) {
+        return p.dans.every((d) => d.mode === UniDMTools.Modes.Bottom)
+      }
+      const progess = pool.dans.map((d) => d.progress)
+      return UniDM.create({
+        SOID: pool.shared.SOID ?? pool.dans[0].SOID,
+        progress: dans[0].danuni_dan.progress,
+        mode:
+          pool.shared.mode ??
+          (isAllBottomMode(pool)
+            ? UniDMTools.Modes.Bottom
+            : UniDMTools.Modes.Top),
+        fontsize: dans.length > 0 ? 36 : 25,
+        color: pool.shared.color ?? pool.most.color,
+        senderID: 'detaolu[bot]@dan-any',
+        content: p.chosen_str,
+        weight: 10,
+        pool: pool.shared.pool ?? pool.most.pool,
+        attr: ['Protect'],
+        platform: pool.shared.platform ?? pool.most.platform,
+        extra: {
+          danuni: {
+            merge: {
+              count: p.danuni_count,
+              duration: Math.max(...progess) - Math.min(...progess),
+              senders: pool.dans
+                .filter((d) => d.content === p.chosen_str)
+                .map((d) => d.senderID),
+              taolu_count: pool.dans.length,
+              taolu_senders: pool.dans.map((d) => d.senderID),
+            },
+          },
+        },
+      })
+    }
+  })
+  return new UniPool(selected)
+}
+
+function detaolu_constructor(config?: DeTaoLuConfig) {
+  return (that: UniPool) => detaolu(that, config)
+}
+
+export default detaolu_constructor