fix: missing export in deno.json

add: .idea to VCS, the refactor guide

.gitignore

@@ -51,7 +51,6 @@ internal/
 !tests/cases/projects/projectOption/**/node_modules
 !tests/cases/projects/NodeModulesSearch/**/*
 !tests/baselines/reference/project/nodeModules*/**/*
-.idea
 yarn.lock
 yarn-error.log
 .parallelperf.*
@@ -78,10 +77,11 @@ node_modules/
 # project specific
 logs/
 __pycache__
-filter/runs
-pred/runs
-pred/checkpoints
-data/
-filter/checkpoints
+ml/filter/runs
+ml/pred/runs
+ml/pred/checkpoints
+ml/pred/observed
+ml/data/
+ml/filter/checkpoints
 scripts
 model/

.idea/.gitignore

@@ -0,0 +1,9 @@
+# Default ignored files
+/shelf/
+/workspace.xml
+# Editor-based HTTP Client requests
+/httpRequests/
+# Datasource local storage ignored files
+/dataSources/
+/dataSources.local.xml
+dataSources.xml

.idea/cvsa.iml

@@ -0,0 +1,21 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="WEB_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$">
+      <excludeFolder url="file://$MODULE_DIR$/.tmp" />
+      <excludeFolder url="file://$MODULE_DIR$/temp" />
+      <excludeFolder url="file://$MODULE_DIR$/tmp" />
+      <excludeFolder url="file://$MODULE_DIR$/ml/data" />
+      <excludeFolder url="file://$MODULE_DIR$/doc" />
+      <excludeFolder url="file://$MODULE_DIR$/ml/filter/checkpoints" />
+      <excludeFolder url="file://$MODULE_DIR$/ml/filter/runs" />
+      <excludeFolder url="file://$MODULE_DIR$/ml/lab/data" />
+      <excludeFolder url="file://$MODULE_DIR$/ml/lab/temp" />
+      <excludeFolder url="file://$MODULE_DIR$/logs" />
+      <excludeFolder url="file://$MODULE_DIR$/model" />
+      <excludeFolder url="file://$MODULE_DIR$/src/db" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>

.idea/inspectionProfiles/Project_Default.xml

@@ -0,0 +1,12 @@
+<component name="InspectionProjectProfileManager">
+  <profile version="1.0">
+    <option name="myName" value="Project Default" />
+    <inspection_tool class="GrazieInspection" enabled="false" level="GRAMMAR_ERROR" enabled_by_default="false" />
+    <inspection_tool class="LanguageDetectionInspection" enabled="false" level="WARNING" enabled_by_default="false" />
+    <inspection_tool class="SpellCheckingInspection" enabled="false" level="TYPO" enabled_by_default="false">
+      <option name="processCode" value="true" />
+      <option name="processLiterals" value="true" />
+      <option name="processComments" value="true" />
+    </inspection_tool>
+  </profile>
+</component>

.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/cvsa.iml" filepath="$PROJECT_DIR$/.idea/cvsa.iml" />
+    </modules>
+  </component>
+</project>

.idea/sqldialects.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="SqlDialectMappings">
+    <file url="PROJECT" dialect="PostgreSQL" />
+  </component>
+</project>

.idea/vcs.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="VcsDirectoryMappings">
+    <mapping directory="" vcs="Git" />
+  </component>
+</project>

README-refactor.md

@@ -0,0 +1,65 @@
+# 项目重构方案
+
+## 目标架构
+采用monorepo结构管理三个独立部分：
+1. `packages/crawler` - 现有爬虫功能
+2. `packages/frontend` - 基于Astro的前端
+3. `packages/backend` - 基于Hono的API后端
+
+## 目录结构调整方案
+
+### 新结构
+```
+.
+├── packages/
+│   ├── crawler/       # 爬虫组件
+│   ├── frontend/      # Astro前端 
+│   ├── backend/       # Hono后端API
+│   └── core/          # 共享代码(未来提取)
+├── docs/              # 文档
+├── scripts/           # 项目脚本
+└── package.json       # 根项目配置
+```
+
+### 具体迁移方案
+
+#### 1. 爬虫部分(crawler)
+保留以下目录/文件：
+- `lib/` (除前端相关)
+- `src/db/raw/` 
+- `src/filterWorker.ts`
+- `src/worker.ts`
+- `test/`
+- `deno.json`
+- `.gitignore`
+
+需要移除：
+- Fresh框架相关文件
+- 前端组件(`components/`)
+- 静态资源(`static/`)
+
+#### 2. 前端部分(frontend)
+全新创建Astro项目，不保留任何现有前端代码
+
+#### 3. 后端部分(backend)
+全新创建Hono项目
+
+#### 4. 共享代码(core)
+未来可从爬虫中提取以下内容到core package：
+- 数据库相关：`lib/db/`
+- 消息队列：`lib/mq/`
+- 网络请求：`lib/net/`
+- 工具函数：`lib/utils/`
+
+## 重构步骤建议
+
+1. 初始化monorepo结构
+2. 迁移爬虫代码到`packages/crawler`
+3. 创建新的Astro项目在`packages/frontend`
+4. 创建新的Hono项目在`packages/backend`
+5. 逐步提取共享代码到`packages/core`
+
+## 注意事项
+- 机器学习相关代码(`pred/`, `filter/`, `lab/`)保持现状
+- 文档(`doc/`)可以迁移到`docs/`目录
+- 需要更新CI/CD流程支持monorepo

components/Button.tsx

@@ -1,12 +0,0 @@
-import { JSX } from "preact";
-import { IS_BROWSER } from "$fresh/runtime.ts";
-
-export function Button(props: JSX.HTMLAttributes<HTMLButtonElement>) {
-	return (
-		<button
-			{...props}
-			disabled={!IS_BROWSER || props.disabled}
-			class="px-2 py-1 border-gray-500 border-2 rounded bg-white hover:bg-gray-200 transition-colors"
-		/>
-	);
-}

data/filter/1.py

@@ -1,55 +0,0 @@
-import json
-import random
-
-def process_data(input_file, output_file):
-    """
-    从输入文件中读取数据，找出model和human不一致的行，
-    删除"model"键，将"human"键重命名为"label"，
-    然后将处理后的数据添加到输出文件中。
-    在写入之前，它会加载output_file中的所有样本，
-    并使用aid键进行去重过滤。
-
-    Args:
-        input_file (str): 输入文件的路径。
-        output_file (str): 输出文件的路径。
-    """
-
-    # 加载output_file中已有的数据，用于去重
-    existing_data = set()
-    try:
-        with open(output_file, 'r', encoding='utf-8') as f_out:
-            for line in f_out:
-                try:
-                    data = json.loads(line)
-                    existing_data.add(data['aid'])
-                except json.JSONDecodeError:
-                    pass  # 忽略JSON解码错误，继续读取下一行
-    except FileNotFoundError:
-        pass  # 如果文件不存在，则忽略
-
-    with open(input_file, 'r', encoding='utf-8') as f_in, open(output_file, 'a', encoding='utf-8') as f_out:
-        for line in f_in:
-            try:
-                data = json.loads(line)
-
-                if data['model'] != data['human'] or random.random() < 0.2:
-                    if data['aid'] not in existing_data:  # 检查aid是否已存在
-                        del data['model']
-                        data['label'] = data['human']
-                        del data['human']
-                        f_out.write(json.dumps(data, ensure_ascii=False) + '\n')
-                        existing_data.add(data['aid'])  # 将新的aid添加到集合中
-
-            except json.JSONDecodeError as e:
-                print(f"JSON解码错误: {e}")
-                print(f"错误行内容: {line.strip()}")
-            except KeyError as e:
-                print(f"KeyError: 键 '{e}' 不存在")
-                print(f"错误行内容: {line.strip()}")
-
-# 调用函数处理数据
-input_file = 'real_test.jsonl'
-output_file = 'labeled_data.jsonl'
-process_data(input_file, output_file)
-print(f"处理完成，结果已写入 {output_file}")
-

deno.json

@@ -1,60 +1,8 @@
 {
-	"lock": false,
+    "lock": false,
+    "workspace": ["./packages/crawler", "./packages/frontend", "./packages/backend", "./packages/core"],
+    "nodeModulesDir": "auto",
 	"tasks": {
-		"crawl-raw-bili": "deno --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run src/db/raw/insertAidsToDB.ts",
-		"crawl-bili-aids": "deno --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run src/db/raw/fetchAids.ts",
-		"check": "deno fmt --check && deno lint && deno check **/*.ts && deno check **/*.tsx",
-		"cli": "echo \"import '\\$fresh/src/dev/cli.ts'\" | deno run --unstable -A -",
-		"manifest": "deno task cli manifest $(pwd)",
-		"start": "deno run -A --watch=static/,routes/ dev.ts",
-		"build": "deno run -A dev.ts build",
-		"preview": "deno run -A main.ts",
-		"update": "deno run -A -r https://fresh.deno.dev/update .",
-		"worker:main": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net --allow-write --allow-run ./src/worker.ts",
-		"worker:filter": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net --allow-write ./src/filterWorker.ts",
-		"adder": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net ./src/jobAdder.ts",
-		"bullui": "deno run --allow-read --allow-env --allow-ffi --allow-net ./src/bullui.ts",
-		"all": "concurrently 'deno task worker:main' 'deno task adder' 'deno task bullui' 'deno task worker:filter'",
-		"test": "deno test ./test/ --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run"
-	},
-	"lint": {
-		"rules": {
-			"tags": ["fresh", "recommended"]
-		}
-	},
-	"exclude": ["**/_fresh/*"],
-	"imports": {
-		"@std/assert": "jsr:@std/assert@1",
-		"$fresh/": "https://deno.land/x/fresh@1.7.3/",
-		"preact": "https://esm.sh/preact@10.22.0",
-		"preact/": "https://esm.sh/preact@10.22.0/",
-		"@preact/signals": "https://esm.sh/*@preact/signals@1.2.2",
-		"@preact/signals-core": "https://esm.sh/*@preact/signals-core@1.5.1",
-		"tailwindcss": "npm:tailwindcss@3.4.1",
-		"tailwindcss/": "npm:/tailwindcss@3.4.1/",
-		"tailwindcss/plugin": "npm:/tailwindcss@3.4.1/plugin.js",
-		"$std/": "https://deno.land/std@0.216.0/",
-		"@huggingface/transformers": "npm:@huggingface/transformers@3.0.0",
-		"bullmq": "npm:bullmq",
-		"lib/": "./lib/",
-		"ioredis": "npm:ioredis",
-		"@bull-board/api": "npm:@bull-board/api",
-		"@bull-board/express": "npm:@bull-board/express",
-		"express": "npm:express",
-		"src/": "./src/",
-		"onnxruntime": "npm:onnxruntime-node@1.19.2",
-		"chalk": "npm:chalk"
-	},
-	"compilerOptions": {
-		"jsx": "react-jsx",
-		"jsxImportSource": "preact"
-	},
-	"nodeModulesDir": "auto",
-	"fmt": {
-		"useTabs": true,
-		"lineWidth": 120,
-		"indentWidth": 4,
-		"semiColons": true,
-		"proseWrap": "always"
+		"crawler": "deno task --filter 'crawler' all"
 	}
 }

dev.ts

@@ -1,7 +0,0 @@
-#!/usr/bin/env -S deno run -A --watch=static/,routes/
-
-import dev from "$fresh/dev.ts";
-import config from "./fresh.config.ts";
-
-import "$std/dotenv/load.ts";
-await dev(import.meta.url, "./main.ts", config);

fresh.config.ts

@@ -1,6 +0,0 @@
-import { defineConfig } from "$fresh/server.ts";
-import tailwind from "$fresh/plugins/tailwind.ts";
-
-export default defineConfig({
-	plugins: [tailwind()],
-});

fresh.gen.ts

@@ -1,27 +0,0 @@
-// DO NOT EDIT. This file is generated by Fresh.
-// This file SHOULD be checked into source version control.
-// This file is automatically updated during development when running `dev.ts`.
-
-import * as $_404 from "./routes/_404.tsx";
-import * as $_app from "./routes/_app.tsx";
-import * as $api_joke from "./routes/api/joke.ts";
-import * as $greet_name_ from "./routes/greet/[name].tsx";
-import * as $index from "./routes/index.tsx";
-import * as $Counter from "./islands/Counter.tsx";
-import type { Manifest } from "$fresh/server.ts";
-
-const manifest = {
-	routes: {
-		"./routes/_404.tsx": $_404,
-		"./routes/_app.tsx": $_app,
-		"./routes/api/joke.ts": $api_joke,
-		"./routes/greet/[name].tsx": $greet_name_,
-		"./routes/index.tsx": $index,
-	},
-	islands: {
-		"./islands/Counter.tsx": $Counter,
-	},
-	baseUrl: import.meta.url,
-} satisfies Manifest;
-
-export default manifest;

islands/Counter.tsx

@@ -1,16 +0,0 @@
-import type { Signal } from "@preact/signals";
-import { Button } from "../components/Button.tsx";
-
-interface CounterProps {
-	count: Signal<number>;
-}
-
-export default function Counter(props: CounterProps) {
-	return (
-		<div class="flex gap-8 py-6">
-			<Button onClick={() => props.count.value -= 1}>-1</Button>
-			<p class="text-3xl tabular-nums">{props.count}</p>
-			<Button onClick={() => props.count.value += 1}>+1</Button>
-		</div>
-	);
-}

lib/ml/mantis.ts

@@ -1,22 +0,0 @@
-import { AIManager } from "lib/ml/manager.ts";
-import * as ort from "onnxruntime";
-import logger from "lib/log/logger.ts";
-import { WorkerError } from "lib/mq/schema.ts";
-
-const modelPath = "./model/model.onnx";
-
-class MantisProto extends AIManager {
-	constructor() {
-		super();
-		this.models = {
-			"predictor": modelPath,
-		};
-	}
-
-	public override async init(): Promise<void> {
-		await super.init();
-	}
-}
-
-const Mantis = new MantisProto();
-export default Mantis;

lib/mq/executors.ts

@@ -1 +0,0 @@
-export * from "lib/mq/exec/getLatestVideos.ts";

main.ts

@@ -1,13 +0,0 @@
-/// <reference no-default-lib="true" />
-/// <reference lib="dom" />
-/// <reference lib="dom.iterable" />
-/// <reference lib="dom.asynciterable" />
-/// <reference lib="deno.ns" />
-
-import "$std/dotenv/load.ts";
-
-import { start } from "$fresh/server.ts";
-import manifest from "./fresh.gen.ts";
-import config from "./fresh.config.ts";
-
-await start(manifest, config);

packages/crawler/db/allData.ts

@@ -1,6 +1,6 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { AllDataType, BiliUserType } from "lib/db/schema.d.ts";
-import Akari from "lib/ml/akari.ts";
+import { AllDataType, BiliUserType } from "db/schema.d.ts";
+import Akari from "ml/akari.ts";
 
 export async function videoExistsInAllData(client: Client, aid: number) {
 	return await client.queryObject<{ exists: boolean }>(

packages/crawler/db/init.ts

@@ -1,5 +1,5 @@
 import { Pool } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { postgresConfig } from "lib/db/pgConfig.ts";
+import { postgresConfig } from "db/pgConfig.ts";
 
 const pool = new Pool(postgresConfig, 12);
 

packages/crawler/db/snapshot.ts

@@ -1,5 +1,5 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { LatestSnapshotType } from "lib/db/schema.d.ts";
+import { LatestSnapshotType } from "db/schema.d.ts";
 
 export async function getVideosNearMilestone(client: Client) {
 	const queryResult = await client.queryObject<LatestSnapshotType>(`

packages/crawler/db/snapshotSchedule.ts

@@ -1,9 +1,9 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { formatTimestampToPsql } from "lib/utils/formatTimestampToPostgre.ts";
+import { formatTimestampToPsql } from "utils/formatTimestampToPostgre.ts";
 import { SnapshotScheduleType } from "./schema.d.ts";
-import logger from "lib/log/logger.ts";
+import logger from "log/logger.ts";
 import { MINUTE } from "$std/datetime/constants.ts";
-import { redis } from "lib/db/redis.ts";
+import { redis } from "db/redis.ts";
 import { Redis } from "ioredis";
 
 const REDIS_KEY = "cvsa:snapshot_window_counts";

packages/crawler/db/songs.ts

@@ -1,5 +1,5 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { parseTimestampFromPsql } from "lib/utils/formatTimestampToPostgre.ts";
+import { parseTimestampFromPsql } from "utils/formatTimestampToPostgre.ts";
 
 export async function getNotCollectedSongs(client: Client) {
 	const queryResult = await client.queryObject<{ aid: number }>(`

packages/crawler/deno.json

@@ -0,0 +1,50 @@
+{
+    "name": "@cvsa/crawler",
+	"tasks": {
+		"crawl-raw-bili": "deno --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run src/db/raw/insertAidsToDB.ts",
+		"crawl-bili-aids": "deno --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run src/db/raw/fetchAids.ts",
+		"check": "deno fmt --check && deno lint && deno check **/*.ts && deno check **/*.tsx",
+		"manifest": "deno task cli manifest $(pwd)",
+		"start": "deno run -A --watch=static/,routes/ dev.ts",
+		"build": "deno run -A dev.ts build",
+		"preview": "deno run -A main.ts",
+		"worker:main": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net --allow-write --allow-run ./src/worker.ts",
+		"worker:filter": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net --allow-write ./src/filterWorker.ts",
+		"adder": "deno run --env-file=.env --allow-env --allow-read --allow-ffi --allow-net ./src/jobAdder.ts",
+		"bullui": "deno run --allow-read --allow-env --allow-ffi --allow-net ./src/bullui.ts",
+		"all": "concurrently 'deno task worker:main' 'deno task adder' 'deno task bullui' 'deno task worker:filter'",
+		"test": "deno test ./test/ --allow-env --allow-ffi --allow-read --allow-net --allow-write --allow-run"
+	},
+	"lint": {
+		"rules": {
+			"tags": ["recommended"]
+		}
+	},
+	"imports": {
+		"@std/assert": "jsr:@std/assert@1",
+		"$std/": "https://deno.land/std@0.216.0/",
+		"@huggingface/transformers": "npm:@huggingface/transformers@3.0.0",
+		"bullmq": "npm:bullmq",
+		"mq/": "./mq/",
+        "db/": "./db/",
+        "log/": "./log/",
+        "net/": "./net/",
+        "ml/": "./ml/",
+        "utils/": "./utils/",
+		"ioredis": "npm:ioredis",
+		"@bull-board/api": "npm:@bull-board/api",
+		"@bull-board/express": "npm:@bull-board/express",
+		"express": "npm:express",
+		"src/": "./src/",
+		"onnxruntime": "npm:onnxruntime-node@1.19.2",
+		"chalk": "npm:chalk"
+	},
+	"fmt": {
+		"useTabs": true,
+		"lineWidth": 120,
+		"indentWidth": 4,
+		"semiColons": true,
+		"proseWrap": "always"
+	},
+    "exports": "./main.ts"
+}

packages/crawler/log/test.ts

@@ -1,4 +1,4 @@
-import logger from "lib/log/logger.ts";
+import logger from "log/logger.ts";
 
 logger.error(Error("test error"), "test service");
 logger.debug(`some string`);

packages/crawler/main.ts

@@ -0,0 +1,7 @@
+// DENO ASK ME TO EXPORT SOMETHING WHEN 'name' IS SPECIFIED
+// AND IF I DON'T SPECIFY 'name', THE --filter FLAG IN `deno task` WON'T WORK.
+// I DONT'T KNOW WHY
+// SO HERE'S A PLACHOLDER EXPORT FOR DENO:
+export const DENO = "FUCK YOU DENO";
+// Oh, maybe export the version is a good idea
+export const VERSION = "1.0.12";

packages/crawler/ml/akari.ts

@@ -1,7 +1,7 @@
-import { AIManager } from "lib/ml/manager.ts";
+import { AIManager } from "ml/manager.ts";
 import * as ort from "onnxruntime";
-import logger from "lib/log/logger.ts";
-import { WorkerError } from "lib/mq/schema.ts";
+import logger from "log/logger.ts";
+import { WorkerError } from "mq/schema.ts";
 import { AutoTokenizer, PreTrainedTokenizer } from "@huggingface/transformers";
 
 const tokenizerModel = "alikia2x/jina-embedding-v3-m2v-1024";

packages/crawler/ml/manager.ts

@@ -1,6 +1,6 @@
 import * as ort from "onnxruntime";
-import logger from "lib/log/logger.ts";
-import { WorkerError } from "lib/mq/schema.ts";
+import logger from "log/logger.ts";
+import { WorkerError } from "mq/schema.ts";
 
 export class AIManager {
 	public sessions: { [key: string]: ort.InferenceSession } = {};

packages/crawler/mq/exec/classifyVideo.ts

@@ -1,13 +1,13 @@
 import { Job } from "bullmq";
-import { db } from "lib/db/init.ts";
-import { getUnlabelledVideos, getVideoInfoFromAllData, insertVideoLabel } from "lib/db/allData.ts";
-import Akari from "lib/ml/akari.ts";
-import { ClassifyVideoQueue } from "lib/mq/index.ts";
-import logger from "lib/log/logger.ts";
-import { lockManager } from "lib/mq/lockManager.ts";
-import { aidExistsInSongs } from "lib/db/songs.ts";
-import { insertIntoSongs } from "lib/mq/task/collectSongs.ts";
-import { scheduleSnapshot } from "lib/db/snapshotSchedule.ts";
+import { db } from "db/init.ts";
+import { getUnlabelledVideos, getVideoInfoFromAllData, insertVideoLabel } from "db/allData.ts";
+import Akari from "ml/akari.ts";
+import { ClassifyVideoQueue } from "mq/index.ts";
+import logger from "log/logger.ts";
+import { lockManager } from "mq/lockManager.ts";
+import { aidExistsInSongs } from "db/songs.ts";
+import { insertIntoSongs } from "mq/task/collectSongs.ts";
+import { scheduleSnapshot } from "db/snapshotSchedule.ts";
 import { MINUTE } from "$std/datetime/constants.ts";
 
 export const classifyVideoWorker = async (job: Job) => {

packages/crawler/mq/exec/getLatestVideos.ts

@@ -1,8 +1,8 @@
 import { Job } from "bullmq";
-import { queueLatestVideos } from "lib/mq/task/queueLatestVideo.ts";
-import { db } from "lib/db/init.ts";
-import { insertVideoInfo } from "lib/mq/task/getVideoDetails.ts";
-import { collectSongs } from "lib/mq/task/collectSongs.ts";
+import { queueLatestVideos } from "mq/task/queueLatestVideo.ts";
+import { db } from "db/init.ts";
+import { insertVideoInfo } from "mq/task/getVideoDetails.ts";
+import { collectSongs } from "mq/task/collectSongs.ts";
 
 export const getLatestVideosWorker = async (_job: Job): Promise<void> => {
 	const client = await db.connect();

packages/crawler/mq/exec/snapshotTick.ts

@@ -1,6 +1,6 @@
 import { Job } from "bullmq";
-import { db } from "lib/db/init.ts";
-import { getLatestVideoSnapshot, getVideosNearMilestone } from "lib/db/snapshot.ts";
+import { db } from "db/init.ts";
+import { getLatestVideoSnapshot, getVideosNearMilestone } from "db/snapshot.ts";
 import {
 	bulkGetVideosWithoutProcessingSchedules,
 	bulkScheduleSnapshot,
@@ -16,18 +16,18 @@ import {
 	snapshotScheduleExists,
 	videoHasProcessingSchedule,
 	getBulkSnapshotsInNextSecond
-} from "lib/db/snapshotSchedule.ts";
+} from "db/snapshotSchedule.ts";
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
 import { HOUR, MINUTE, SECOND, WEEK } from "$std/datetime/constants.ts";
-import logger from "lib/log/logger.ts";
-import { SnapshotQueue } from "lib/mq/index.ts";
-import { insertVideoSnapshot } from "lib/mq/task/getVideoStats.ts";
-import { NetSchedulerError } from "lib/mq/scheduler.ts";
-import { getBiliVideoStatus, setBiliVideoStatus } from "lib/db/allData.ts";
-import { truncate } from "lib/utils/truncate.ts";
-import { lockManager } from "lib/mq/lockManager.ts";
-import { getSongsPublihsedAt } from "lib/db/songs.ts";
-import { bulkGetVideoStats } from "lib/net/bulkGetVideoStats.ts";
+import logger from "log/logger.ts";
+import { SnapshotQueue } from "mq/index.ts";
+import { insertVideoSnapshot } from "mq/task/getVideoStats.ts";
+import { NetSchedulerError } from "mq/scheduler.ts";
+import { getBiliVideoStatus, setBiliVideoStatus } from "db/allData.ts";
+import { truncate } from "utils/truncate.ts";
+import { lockManager } from "mq/lockManager.ts";
+import { getSongsPublihsedAt } from "db/songs.ts";
+import { bulkGetVideoStats } from "net/bulkGetVideoStats.ts";
 
 const priorityMap: { [key: string]: number } = {
 	"milestone": 1,

packages/crawler/mq/executors.ts

@@ -0,0 +1 @@
+export * from "mq/exec/getLatestVideos.ts";

packages/crawler/mq/init.ts

@@ -1,9 +1,9 @@
 import { MINUTE, SECOND } from "$std/datetime/constants.ts";
-import { ClassifyVideoQueue, LatestVideosQueue, SnapshotQueue } from "lib/mq/index.ts";
-import logger from "lib/log/logger.ts";
-import { initSnapshotWindowCounts } from "lib/db/snapshotSchedule.ts";
-import { db } from "lib/db/init.ts";
-import { redis } from "lib/db/redis.ts";
+import { ClassifyVideoQueue, LatestVideosQueue, SnapshotQueue } from "mq/index.ts";
+import logger from "log/logger.ts";
+import { initSnapshotWindowCounts } from "db/snapshotSchedule.ts";
+import { db } from "db/init.ts";
+import { redis } from "db/redis.ts";
 
 export async function initMQ() {
 	const client = await db.connect();

packages/crawler/mq/lockManager.ts

@@ -1,5 +1,5 @@
 import { Redis } from "ioredis";
-import { redis } from "lib/db/redis.ts";
+import { redis } from "db/redis.ts";
 
 class LockManager {
 	private redis: Redis;

packages/crawler/mq/rateLimiter.ts

@@ -1,4 +1,4 @@
-import { SlidingWindow } from "lib/mq/slidingWindow.ts";
+import { SlidingWindow } from "mq/slidingWindow.ts";
 
 export interface RateLimiterConfig {
 	window: SlidingWindow;

packages/crawler/mq/scheduler.ts

@@ -1,7 +1,7 @@
-import logger from "lib/log/logger.ts";
-import { RateLimiter, RateLimiterConfig } from "lib/mq/rateLimiter.ts";
-import { SlidingWindow } from "lib/mq/slidingWindow.ts";
-import { redis } from "lib/db/redis.ts";
+import logger from "log/logger.ts";
+import { RateLimiter, RateLimiterConfig } from "mq/rateLimiter.ts";
+import { SlidingWindow } from "mq/slidingWindow.ts";
+import { redis } from "db/redis.ts";
 import Redis from "ioredis";
 import { SECOND } from "$std/datetime/constants.ts";
 

packages/crawler/mq/task/collectSongs.ts

@@ -1,7 +1,7 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { aidExistsInSongs, getNotCollectedSongs } from "lib/db/songs.ts";
-import logger from "lib/log/logger.ts";
-import { scheduleSnapshot } from "lib/db/snapshotSchedule.ts";
+import { aidExistsInSongs, getNotCollectedSongs } from "db/songs.ts";
+import logger from "log/logger.ts";
+import { scheduleSnapshot } from "db/snapshotSchedule.ts";
 import { MINUTE } from "$std/datetime/constants.ts";
 
 export async function collectSongs(client: Client) {

packages/crawler/mq/task/getVideoDetails.ts

@@ -1,9 +1,9 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { getVideoDetails } from "lib/net/getVideoDetails.ts";
-import { formatTimestampToPsql } from "lib/utils/formatTimestampToPostgre.ts";
-import logger from "lib/log/logger.ts";
-import { ClassifyVideoQueue } from "lib/mq/index.ts";
-import { userExistsInBiliUsers, videoExistsInAllData } from "lib/db/allData.ts";
+import { getVideoDetails } from "net/getVideoDetails.ts";
+import { formatTimestampToPsql } from "utils/formatTimestampToPostgre.ts";
+import logger from "log/logger.ts";
+import { ClassifyVideoQueue } from "mq/index.ts";
+import { userExistsInBiliUsers, videoExistsInAllData } from "db/allData.ts";
 import { HOUR, SECOND } from "$std/datetime/constants.ts";
 
 export async function insertVideoInfo(client: Client, aid: number) {

packages/crawler/mq/task/getVideoStats.ts

@@ -1,7 +1,7 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { getVideoInfo } from "lib/net/getVideoInfo.ts";
-import { LatestSnapshotType } from "lib/db/schema.d.ts";
-import logger from "lib/log/logger.ts";
+import { getVideoInfo } from "net/getVideoInfo.ts";
+import { LatestSnapshotType } from "db/schema.d.ts";
+import logger from "log/logger.ts";
 
 /*
  * Fetch video stats from bilibili API and insert into database

packages/crawler/mq/task/queueLatestVideo.ts

@@ -1,10 +1,10 @@
 import { Client } from "https://deno.land/x/postgres@v0.19.3/mod.ts";
-import { getLatestVideoAids } from "lib/net/getLatestVideoAids.ts";
-import { videoExistsInAllData } from "lib/db/allData.ts";
-import { sleep } from "lib/utils/sleep.ts";
+import { getLatestVideoAids } from "net/getLatestVideoAids.ts";
+import { videoExistsInAllData } from "db/allData.ts";
+import { sleep } from "utils/sleep.ts";
 import { SECOND } from "$std/datetime/constants.ts";
-import logger from "lib/log/logger.ts";
-import { LatestVideosQueue } from "lib/mq/index.ts";
+import logger from "log/logger.ts";
+import { LatestVideosQueue } from "mq/index.ts";
 
 export async function queueLatestVideos(
 	client: Client,

packages/crawler/net/bulkGetVideoStats.ts

@@ -1,6 +1,6 @@
-import netScheduler from "lib/mq/scheduler.ts";
-import { MediaListInfoData, MediaListInfoResponse } from "lib/net/bilibili.d.ts";
-import logger from "lib/log/logger.ts";
+import netScheduler from "mq/scheduler.ts";
+import { MediaListInfoData, MediaListInfoResponse } from "net/bilibili.d.ts";
+import logger from "log/logger.ts";
 
 /*
  * Bulk fetch video metadata from bilibili API

packages/crawler/net/getLatestVideoAids.ts

@@ -1,6 +1,6 @@
-import { VideoListResponse } from "lib/net/bilibili.d.ts";
-import logger from "lib/log/logger.ts";
-import netScheduler from "lib/mq/scheduler.ts";
+import { VideoListResponse } from "net/bilibili.d.ts";
+import logger from "log/logger.ts";
+import netScheduler from "mq/scheduler.ts";
 
 export async function getLatestVideoAids(page: number = 1, pageSize: number = 10): Promise<number[]> {
 	const startFrom = 1 + pageSize * (page - 1);

packages/crawler/net/getVideoDetails.ts

@@ -1,6 +1,6 @@
-import netScheduler from "lib/mq/scheduler.ts";
-import { VideoDetailsData, VideoDetailsResponse } from "lib/net/bilibili.d.ts";
-import logger from "lib/log/logger.ts";
+import netScheduler from "mq/scheduler.ts";
+import { VideoDetailsData, VideoDetailsResponse } from "net/bilibili.d.ts";
+import logger from "log/logger.ts";
 
 export async function getVideoDetails(aid: number): Promise<VideoDetailsData | null> {
 	const url = `https://api.bilibili.com/x/web-interface/view/detail?aid=${aid}`;

packages/crawler/net/getVideoInfo.ts

@@ -1,6 +1,6 @@
-import netScheduler from "lib/mq/scheduler.ts";
-import { VideoInfoData, VideoInfoResponse } from "lib/net/bilibili.d.ts";
-import logger from "lib/log/logger.ts";
+import netScheduler from "mq/scheduler.ts";
+import { VideoInfoData, VideoInfoResponse } from "net/bilibili.d.ts";
+import logger from "log/logger.ts";
 
 /*
  * Fetch video metadata from bilibili API

packages/crawler/src/bullui.ts

@@ -2,7 +2,7 @@ import express from "express";
 import { createBullBoard } from "@bull-board/api";
 import { BullMQAdapter } from "@bull-board/api/bullMQAdapter.js";
 import { ExpressAdapter } from "@bull-board/express";
-import { ClassifyVideoQueue, LatestVideosQueue, SnapshotQueue } from "lib/mq/index.ts";
+import { ClassifyVideoQueue, LatestVideosQueue, SnapshotQueue } from "mq/index.ts";
 
 const serverAdapter = new ExpressAdapter();
 serverAdapter.setBasePath("/");

packages/crawler/src/filterWorker.ts

@@ -1,10 +1,10 @@
 import { ConnectionOptions, Job, Worker } from "bullmq";
-import { redis } from "lib/db/redis.ts";
-import logger from "lib/log/logger.ts";
-import { classifyVideosWorker, classifyVideoWorker } from "lib/mq/exec/classifyVideo.ts";
-import { WorkerError } from "lib/mq/schema.ts";
-import { lockManager } from "lib/mq/lockManager.ts";
-import Akari from "lib/ml/akari.ts";
+import { redis } from "db/redis.ts";
+import logger from "log/logger.ts";
+import { classifyVideosWorker, classifyVideoWorker } from "mq/exec/classifyVideo.ts";
+import { WorkerError } from "mq/schema.ts";
+import { lockManager } from "mq/lockManager.ts";
+import Akari from "ml/akari.ts";
 
 Deno.addSignalListener("SIGINT", async () => {
 	logger.log("SIGINT Received: Shutting down workers...", "mq");

packages/crawler/src/jobAdder.ts

@@ -0,0 +1,3 @@
+import { initMQ } from "mq/init.ts";
+
+await initMQ();

packages/crawler/src/worker.ts

@@ -1,10 +1,10 @@
 import { ConnectionOptions, Job, Worker } from "bullmq";
-import { collectSongsWorker, getLatestVideosWorker } from "lib/mq/executors.ts";
-import { redis } from "lib/db/redis.ts";
-import logger from "lib/log/logger.ts";
-import { lockManager } from "lib/mq/lockManager.ts";
-import { WorkerError } from "lib/mq/schema.ts";
-import { getVideoInfoWorker } from "lib/mq/exec/getLatestVideos.ts";
+import { collectSongsWorker, getLatestVideosWorker } from "mq/executors.ts";
+import { redis } from "db/redis.ts";
+import logger from "log/logger.ts";
+import { lockManager } from "mq/lockManager.ts";
+import { WorkerError } from "mq/schema.ts";
+import { getVideoInfoWorker } from "mq/exec/getLatestVideos.ts";
 import {
 	collectMilestoneSnapshotsWorker,
 	regularSnapshotsWorker,
@@ -13,7 +13,7 @@ import {
 	scheduleCleanupWorker,
 	takeBulkSnapshotForVideosWorker,
 	bulkSnapshotTickWorker
-} from "lib/mq/exec/snapshotTick.ts";
+} from "mq/exec/snapshotTick.ts";
 
 Deno.addSignalListener("SIGINT", async () => {
 	logger.log("SIGINT Received: Shutting down workers...", "mq");