@yorha2b-lab/autodev 5.4.0 → 5.5.1

package/.github/scripts/analyze/deep_intelligence.py

@@ -0,0 +1,159 @@
+import os
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+from sklearn.cluster import DBSCAN
+from sklearn.preprocessing import StandardScaler
+
+
+def sigmoid(x):
+    """
+    概率平滑函数：
+    避免指标达到某个值后直接封顶，
+    让异常程度呈渐进变化。
+    """
+    return 1 / (1 + np.exp(-x))
+
+
+def map_rank(cluster_id):
+    """
+    DBSCAN 的 cluster 数量不是固定的，
+    不能写死 0/1/2，否则新增聚类会出现 NaN。
+    """
+    if cluster_id == -1:
+        return 'Legendary'
+    if cluster_id == 0:
+        return 'Veteran'
+    return f'Tactical Unit {cluster_id}'
+
+
+def execute_deep_intelligence():
+
+    print("🤖 Pod 042: 启动‘全量情报融合协议’...")
+
+    clone_path = 'ghrs-data/clones_ledger.csv'
+    view_path = 'ghrs-data/views_ledger.csv'
+
+    if not os.path.exists(clone_path) or not os.path.exists(view_path):
+        print("⚠️ 数据账本缺失，取消分析。")
+        return
+
+    # 1. 物理对账
+    df_c = pd.read_csv(clone_path)
+    df_v = pd.read_csv(view_path)
+
+    df = pd.merge(df_c, df_v, on='date', how='inner').fillna(0)
+
+    # ==================================================
+    # 2. 特征工程
+    # ==================================================
+
+    # 访问转化强度
+    df['ratio'] = (df['clones'] / df['views'].replace(0, 1))
+
+    # Github 行为天然长尾，因此进入 Log 空间
+    df['log_volume'] = np.log1p(df['clones'])
+    df['log_intensity'] = np.log1p(df['ratio'])
+
+    # ==================================================
+    # 3. 自动化概率审计
+    # ==================================================
+
+    # 原始 ratio 越高，越像自动化行为
+    ratio_score = sigmoid(df['ratio'] - 3)
+
+    # 单日爆发越高，异常概率越高
+    burst_score = sigmoid((df['clones'] - 100) / 100)
+
+    # 综合异常系数
+    df['bot_prob'] = (ratio_score * 0.6 + burst_score * 0.4)
+
+    # ==================================================
+    # 4. DBSCAN 行为聚类
+    # ==================================================
+    X = StandardScaler().fit_transform(df[['log_volume', 'log_intensity']])
+    db = DBSCAN(eps=0.4, min_samples=3).fit(X)
+    df['cluster'] = db.labels_
+
+    # ==================================================
+    # 5. 战力结算
+    # ==================================================
+    bot_clones = (df['clones'] * df['bot_prob']).sum()
+    total_clones = df['clones'].sum()
+    human_clones = max(0, total_clones - bot_clones)
+    df['rank'] = df['cluster'].apply(map_rank)
+    rank_counts = (df['rank'].value_counts())
+
+    # ==================================================
+    # 6. 全频道视觉矩阵
+    # ==================================================
+
+    plt.style.use('dark_background')
+    fig = plt.figure(figsize=(24, 8))
+
+    gs = fig.add_gridspec(1, 3, width_ratios=[1, 1.2, 1.8])
+
+    # ----------------------------
+    # A. 自动化组成
+    # ----------------------------
+
+    ax1 = fig.add_subplot(gs[0])
+
+    ax1.pie(
+        [human_clones, bot_clones],
+        labels=['Human', 'Machine'],
+        autopct='%1.1f%%',
+        startangle=90,
+        pctdistance=0.85,
+        explode=(0.05, 0)
+    )
+
+    ax1.add_artist(plt.Circle((0, 0), 0.70, fc='#0d1117'))
+    ax1.set_title("Operational Composition", color='#33cc33', pad=20)
+
+    # ----------------------------
+    # B. 行为层级
+    # ----------------------------
+
+    ax2 = fig.add_subplot(gs[1])
+    sns.barplot(x=rank_counts.index, y=rank_counts.values,
+                palette='viridis', ax=ax2)
+
+    ax2.set_title("Commander Hierarchy", color='#0066ff', pad=20)
+
+    ax2.set_ylabel("Days Count")
+
+    ax2.grid(axis='y', alpha=0.1)
+
+    # ----------------------------
+    # C. 行为拓扑
+    # ----------------------------
+
+    ax3 = fig.add_subplot(gs[2])
+
+    # sqrt 缩放避免极端 clone 撑爆气泡
+    bubble_size = (np.sqrt(df['clones'] + 1) * 20)
+    ax3.scatter(
+        df['log_volume'],
+        df['log_intensity'],
+        c=df['cluster'],
+        cmap='coolwarm',
+        s=bubble_size,
+        alpha=0.6,
+        edgecolors='w'
+    )
+    ax3.set_title("Sovereign Logistic Matrix (Log Space)",
+                  color='#ff00ff', pad=20)
+    ax3.set_xlabel("Log(Total Clones)")
+    ax3.set_ylabel("Log(Intensity)")
+    ax3.grid(alpha=0.1)
+    plt.tight_layout()
+    os.makedirs('plots', exist_ok=True)
+    plt.savefig('plots/bunker_intelligence_fusion.png', dpi=120)
+
+    print("✅ 全量情报视觉矩阵封存成功！")
+
+
+if __name__ == "__main__":
+    execute_deep_intelligence()

package/.github/scripts/deploy/publish.py

@@ -33,9 +33,7 @@ def assemble_readme():
 ![Peak Recon](./plots/bunker_peak_recon.png)
 
 ### 🧠 4. 深度情报审计 (Intelligence Intelligence)
-| 成分分析 (Meat vs Machine) | 战力梯度 (Hierarchy) |
-| :---: | :---: |
-| ![Bot](./plots/bot_composition.png) | ![Cluster](./plots/commander_profiles.png) |
+![Fusion](./plots/bunker_intelligence_fusion.png)
 
 #### 🛰️ 流量来源实时追踪
 {ref_table}

package/.github/scripts/render/dashboard.py

@@ -181,43 +181,87 @@ def intelligence_grid():
     plt.savefig('plots/bunker_intelligence_grid.png', dpi=100)
 
 
-def load_clean(path, col):
+def load_clean(path, col_name):
     df = pd.read_csv(path)
     df['date'] = pd.to_datetime(df['date'], errors='coerce')
     df = df.dropna(subset=['date']).copy()
-    df[col] = pd.to_numeric(df[col.lower()], errors='coerce').fillna(0)
-    df['uniques'] = pd.to_numeric(df['uniques'], errors='coerce').fillna(0)
-    return df[['date', col, 'uniques']]
+
+    # 物理强转数字
+    df[col_name] = pd.to_numeric(
+        df[col_name.lower()], errors='coerce').fillna(0)
+
+    # 💡 核心修正：直接在这里把 uniques 重命名为 [类别]_Uniques
+    # 这样合并后，列名就是固定的 'Clones_Uniques' 和 'Views_Uniques'
+    new_unique_name = f"{col_name}_Uniques"
+    df[new_unique_name] = pd.to_numeric(
+        df['uniques'], errors='coerce').fillna(0)
+
+    return df[['date', col_name, new_unique_name]]
 
 
 def main_dashboard(df_c, df_v):
+    # 💡 1. 数据融合与脱水
     df = pd.merge(df_c, df_v, on='date', how='outer').fillna(
         0).sort_values('date')
     df['cumulative'] = df['Clones'].cumsum()
-    df['ratio'] = df['Clones'] / df['uniques'].replace(0, 1)
-    df['clones_ma7'] = df['Clones'].rolling(window=7, min_periods=1).mean()
+    df['ratio'] = df['Clones'] / \
+        df['Views_Uniques'].replace(0, 1)
+
+    # 💡 2. 核心算法注入：计算“信号纯度” (Purity)
+    df['cumulative'] = df['Clones'].cumsum()
+
+    # 计算“信号纯度”：克隆数 / 访客数
+    df['ratio'] = df['Clones'] / df['Views_Uniques'].replace(0, 1)
+
+    # 算法：(比例/10 * 0.6) + (总量/500 * 0.4)
+    df['bot_prob'] = (df['ratio'] / 10).clip(upper=1) * 0.6 + \
+        (df['Clones'] / 500).clip(upper=1) * 0.4
+    df['purity'] = (1.0 - df['bot_prob']) * 100
 
-    fig1 = plt.figure(figsize=(12, 10))
-    gs1 = fig1.add_gridspec(3, 1)
+    # 💡 3. 画布物理扩容：由 3 行 1 列 升级为 4 行 1 列
+    fig1 = plt.figure(figsize=(12, 14))  # 增加高度以容纳四个雷达
+    gs1 = fig1.add_gridspec(4, 1)
+
+    # --- 子图 1: 实时侦察 (Views vs Clones) ---
     ax1 = fig1.add_subplot(gs1[0])
     ax1.plot(df['date'].tail(30), df['Views'].tail(
         30), alpha=0.5, ls='--', label='Views (Web)')
     ax1.plot(df['date'].tail(30), df['Clones'].tail(30),
              color='#33cc33', lw=2, label='Clones (Terminal)')
     ax1.set_title(
-        'Strategic Recon: Views vs Clones (Last 30D)', color='#33cc33')
+        'Strategic Recon: Web Visitors vs Terminal Commandos (30D)', color='#33cc33')
     ax1.legend(loc='upper left', frameon=True, fontsize='small')
+
+    # --- 子图 2: 累计荣光 (Growth) ---
     ax2 = fig1.add_subplot(gs1[1])
-    ax2.fill_between(df['date'], df['cumulative'],
-                     color='#0066ff', alpha=0.2)
+    ax2.fill_between(df['date'], df['cumulative'], color='#0066ff', alpha=0.2)
     ax2.plot(df['date'], df['cumulative'], color='#0066ff', lw=3)
     ax2.set_title('Bunker Glory: Total Cumulative Growth', color='#0066ff')
-    ax3 = fig1.add_subplot(gs1[2])
-    pivot = df.assign(week=df['date'].dt.isocalendar().week, day=df['date'].dt.day_name()).pivot_table(index='day', columns='week',
-                                                                                                       values='Clones', aggfunc='sum').fillna(0).reindex(['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday'])
+
+    # --- 💡 子图 3 [核心新增]: 信号纯度审计 (Signal Purity) ---
+    ax_p = fig1.add_subplot(gs1[2])
+    # 使用渐变紫色，代表地堡的“骇入感知”
+    ax_p.fill_between(df['date'].tail(45), df['purity'].tail(
+        45), color='#ff00ff', alpha=0.1)
+    ax_p.plot(df['date'].tail(45), df['purity'].tail(45),
+              color='#ff00ff', lw=1.5, label='Signal Purity %')
+    ax_p.set_ylim(0, 105)  # 留点空白放标题
+    ax_p.axhline(y=50, color='#444', linestyle=':', alpha=0.5)  # 50% 警戒线
+    ax_p.set_title(
+        'Neural Cloud Integrity: Human Signal Purity Audit (%)', color='#ff00ff')
+    ax_p.legend(loc='lower left', fontsize='x-small')
+
+    # --- 子图 4: 作战热力图 (Heatmap) ---
+    ax3 = fig1.add_subplot(gs1[3])
+    pivot = df.assign(week=df['date'].dt.isocalendar().week, day=df['date'].dt.day_name()).pivot_table(
+        index='day', columns='week', values='Clones', aggfunc='sum'
+    ).fillna(0).reindex(['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday'])
+
     sns.heatmap(pivot, cmap='Greens', ax=ax3,
                 cbar=False, lw=2, linecolor='#1a1a1a')
-    ax3.set_title('Deployment Matrix: All-Time Activity', color='#33cc33')
+    ax3.set_title(
+        'Global Deployment Heatmap: Strategic Activity Matrix', color='#33cc33')
+
     plt.tight_layout()
     plt.savefig('plots/bunker_main_v7.png', dpi=120)
 

package/.github/workflows/bunker-intelligence.yml

@@ -35,7 +35,7 @@ jobs:
           python-version: '3.10'
 
       - name: Install Dependencies
-        run: pip install matplotlib pandas seaborn scipy numpy requests
+        run: pip install matplotlib pandas seaborn scipy numpy requests scikit-learn
 
       # 💡 步骤 3：运行脚本，注意现在的相对路径！
       # 我们在当前根目录运行，调用 scripts_repo 里的脚本，去处理 data_repo 里的数据
@@ -48,8 +48,7 @@ jobs:
         run: |
           cd data_repo
           # 执行“肉机比”扫描与“军衔”聚类分析
-          python ../scripts_repo/.github/scripts/analyze/bot_detect.py
-          python ../scripts_repo/.github/scripts/analyze/cluster.py
+          python ../scripts_repo/.github/scripts/analyze/deep_intelligence.py
 
       - name: '📈 Step 3: Visualization Rendering'
         run: |

package/bin/autodev.js

@@ -4,9 +4,11 @@ const path = require('path')
 const chalk = require('chalk')
 const pkg = require('../package.json')
 const { program } = require('commander')
+
+const { discover } = require('../src/utils/foundation')
 const { local, matrixEffect, bootSequence } = require('../src/utils/ux')
 
-const dialogs = require('../src/bootstrap/registry')({ dir: path.join(__dirname, '../src/dialogs') })
+const dialogs = discover(path.join(__dirname, '../src/dialogs'))
 const dialog = dialogs[local] ?? dialogs['EN-US']
 
 program

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yorha2b-lab/autodev",
-  "version": "5.4.0",
+  "version": "5.5.1",
   "description": "基于视觉大模型的前端(react+Antd)全自动 CRUD 代码生成器",
   "bin": {
     "autodev": "bin/autodev.js"

package/src/bootstrap/index.js

@@ -27,9 +27,9 @@ module.exports = {
             openAI: require('./openai')(),
             core: require('../utils/core'),
             menus: foundation.getExistingMenus(),
+            handlers: foundation.discover(path.join(__dirname, '../handlers')),
             handlebars: require('./handlebar')({ config, template, ...foundation }),
-            handlers: require('./registry')({ dir: path.join(__dirname, '../handlers') }),
-            prompts: require('./registry')({ dir: path.join(__dirname, `../framework/${template}/prompts`) }),
+            prompts: foundation.discover(path.join(__dirname, `../framework/${template}/prompts`)),
         }
 
         require('./foundation')(accessPoint)

package/src/bootstrap/yorha.js

@@ -5,23 +5,23 @@ module.exports = () => {
 
     const createMember = ({ name }) => ({
         start(text) {
-            const spinner = ora({ text: chalk.cyan(`\n${name} ${text}`) }).start()
+            const spinner = ora({ text: chalk.cyan(`${name} ${text}\n`) }).start()
             return spinner
         },
         fail(spinner, text) {
-            spinner.fail(chalk.red(`\n${name} ${text}`))
+            spinner.fail(chalk.red(`${name} ${text}\n`))
         },
         report(text, color = 'yellow') {
-            console.log(chalk[color](`\n${name} ${text}`))
+            console.log(chalk[color](`${name} ${text}\n`))
         },
         update(spinner, text) {
-            spinner.text = chalk.cyan(`\n${name} ${text}`)
+            spinner.text = chalk.cyan(`${name} ${text}\n`)
         },
         warning(spinner, text) {
-            spinner.warn(chalk.yellow(`\n${name} ${text}`))
+            spinner.warn(chalk.yellow(`${name} ${text}\n`))
         },
         success(spinner, text) {
-            spinner.succeed(chalk.green(`\n${name} ${text}`))
+            spinner.succeed(chalk.green(`${name} ${text}\n`))
         },
     })
 

package/src/framework/react/handlebars/partials/handleBlock.hbs

@@ -38,7 +38,7 @@ const onTabChange = key => {
 {{#if hasExpandable}}
 const onExpand = async (expanded, record) => {
     if (expanded) {
-        setExpandedRowKeys(prev=>[...prev, record[rowKey])
+        setExpandedRowKeys(prev=>[...prev, record[rowKey]])
     } else {
         setExpandedRowKeys(prev=>prev.filter(item => item !== record[rowKey]))
     }

package/src/handlers/page.js

@@ -5,7 +5,7 @@ module.exports = {
     watch: 'screenShot',
     async handle(filePath) {
 
-        const { llm, menus, yorha, dialog, config, prompts, generator, foundation } = require('../bootstrap').get()
+        const { llm, menus, yorha, dialog, config, prompts, template, generator, foundation } = require('../bootstrap').get()
 
         const { page } = prompts
         const { nineS, pod042 } = yorha
@@ -37,7 +37,7 @@ module.exports = {
                 pageConfig = require('../../example/example.json')
             } else {
                 pod042.update(spinner, dialog.pod042.uploadVisualMetadata)
-                pageConfig = await recognizePage({ prompt: page(fileName), filePath })
+                pageConfig = await recognizePage({ prompt: page(fileName), filePath, schema: require(`../framework/${template}/schema/page.json`) })
             }
 
             fs.writeFileSync(path.join(targetDir, 'resource.js'), resource({ pageConfig }))
@@ -62,6 +62,9 @@ module.exports = {
             }
 
         } catch (error) {
+            if (fs.existsSync(targetDir)) {
+                fs.rmSync(targetDir, { recursive: true, force: true })
+            }
             pod042.fail(spinner, dialog.pod042.constructionAborted(error))
         }
     }

package/src/handlers/part.js

@@ -20,7 +20,7 @@ module.exports = {
         try {
             pod042.update(spinner, dialog.pod042.extractingUiMetadata)
 
-            const pageConfig = await recognizePage({ prompt, filePath, taskType: 'part' })
+            const pageConfig = await recognizePage({ prompt, filePath })
 
             const { formItems, dictBlocks, processedColumns } = formatFormItemAndColumns({ pageConfig })
 

package/src/services/llm.js

@@ -32,7 +32,7 @@ const askAI = async ({ model, yorha, dialog, openAI, messages, response_format =
     }
 }
 
-module.exports = ({ config, yorha, openAI, dialog, prompts, template }) => {
+module.exports = ({ config, yorha, openAI, dialog, prompts }) => {
 
     const sharp = require('sharp')
 
@@ -65,19 +65,7 @@ module.exports = ({ config, yorha, openAI, dialog, prompts, template }) => {
                 ]
             })
         },
-        alignResponseFields: async ({ responseStr, resourceStr }) => {
-            return askAI({
-                yorha,
-                openAI,
-                dialog,
-                model: textModel,
-                messages: [
-                    { role: 'system', content: API_DESIGNER },
-                    { role: 'user', content: api({ responseStr, resourceStr }) }
-                ]
-            })
-        },
-        recognizePage: async ({ prompt, filePath, taskType = 'page' }) => {
+        recognizePage: async ({ prompt, filePath, schema }) => {
             const compressedBuffer = await sharp(filePath)
                 .resize(1280, null, { withoutEnlargement: true })
                 .jpeg({ quality: 80 })
@@ -99,7 +87,19 @@ module.exports = ({ config, yorha, openAI, dialog, prompts, template }) => {
                         ]
                     }
                 ],
-                response_format: taskType === 'page' ? { type: 'json_schema', strict: true, json_schema: require(`../framework/${template}/schema/page.json`) } : { type: 'json_object' }
+                response_format: schema ? { strict: true, type: 'json_schema', json_schema: schema } : { type: 'json_object' }
+            })
+        },
+        alignResponseFields: async ({ responseStr, resourceStr }) => {
+            return askAI({
+                yorha,
+                openAI,
+                dialog,
+                model: textModel,
+                messages: [
+                    { role: 'system', content: API_DESIGNER },
+                    { role: 'user', content: api({ responseStr, resourceStr }) }
+                ]
             })
         },
     }

package/src/utils/foundation.js

@@ -11,7 +11,6 @@ const fs = require('fs')
 const path = require('path')
 const stringify = require('json-stringify-pretty-compact')
 
-
 /**
  * 获取项目配置信息
  * 优先读取项目根目录的 config.js，如果不存在则使用默认配置
@@ -63,4 +62,6 @@ const copyTemplateDir = (template, templateSubDir, targetSubDir) => {
 
 const contextStringify = ({ context, indent = 4, maxLength = 200 }) => stringify.default(context, { indent, maxLength })
 
-module.exports = { getConfig, getExistingMenus, copyTemplateDir, contextStringify }
+const discover = dir => Object.fromEntries(fs.readdirSync(dir).filter(file => file.endsWith('.js')).map(file => [path.basename(file, '.js'), require(path.join(dir, file))]))
+
+module.exports = { discover, getConfig, getExistingMenus, copyTemplateDir, contextStringify }

package/.github/scripts/analyze/bot_detect.py

@@ -1,43 +0,0 @@
-import os
-import pandas as pd
-import matplotlib.pyplot as plt
-
-
-def plot_bot_composition():
-    print('🤖 Pod 042: 正在绘制‘肉机成分分析图’...')
-    df_c = pd.read_csv('ghrs-data/clones_ledger.csv')
-    df_v = pd.read_csv('ghrs-data/views_ledger.csv')
-    df = pd.merge(df_c, df_v, on='date', how='inner')
-
-    # 算法：Shadow Ratio > 5 判定为机器
-    df['is_bot'] = (df['clones'] / df['views'].replace(0, 1)) > 5
-    bot_clones = df[df['is_bot']]['clones'].sum()
-    total_clones = df['clones'].sum()
-    human_clones = total_clones - bot_clones
-
-    # 💡 绘图：地堡级甜甜圈图
-    plt.style.use('dark_background')
-    fig, ax = plt.subplots(figsize=(8, 8))
-
-    data = [human_clones, bot_clones]
-    labels = ['Human Commanders', 'Machine Lifeforms']
-    colors = ['#33cc33', '#444444']  # 地堡绿 vs 灰
-
-    wedges, texts, autotexts = ax.pie(data, labels=labels, autopct='%1.1f%%',
-                                      startangle=90, colors=colors, pctdistance=0.85,
-                                      explode=(0.05, 0), textprops={'color': 'w'})
-
-    # 画个中间的黑洞，变成甜甜圈
-    centre_circle = plt.Circle((0, 0), 0.70, fc='#0d1117')
-    fig.gca().add_artist(centre_circle)
-
-    ax.set_title('Operational Composition: Meat vs Machine',
-                 color='#33cc33', fontsize=14)
-    plt.tight_layout()
-    os.makedirs('plots', exist_ok=True)
-    plt.savefig('plots/bot_composition.png', dpi=100)
-    print('✅ 甜甜圈图封存成功。')
-
-
-if __name__ == '__main__':
-    plot_bot_composition()

package/.github/scripts/analyze/cluster.py

@@ -1,40 +0,0 @@
-import pandas as pd
-import seaborn as sns
-import matplotlib.pyplot as plt
-
-
-def plot_commander_profiles():
-    print('🤖 Pod 042: 正在绘制‘指挥官军衔分布图’...')
-    df_c = pd.read_csv('ghrs-data/clones_ledger.csv')
-    df_c['intensity'] = df_c['clones'] / df_c['uniques'].replace(0, 1)
-
-    # 定义军衔逻辑
-    def get_rank(row):
-        if row['intensity'] > 5:
-            return 'Industrial (CI/CD)'
-        if row['intensity'] > 1.5:
-            return 'Veteran (Hardcore)'
-        return 'Recruit (Fresh)'
-
-    df_c['rank'] = df_c.apply(get_rank, axis=1)
-    rank_counts = df_c['rank'].value_counts()
-
-    # 💡 绘图：战力梯度图
-    plt.style.use('dark_background')
-    plt.figure(figsize=(10, 6))
-
-    # 用 seaborn 画出更有质感的柱状图
-    ax = sns.barplot(x=rank_counts.index,
-                     y=rank_counts.values, palette='viridis')
-
-    plt.title('Commander Hierarchy Distribution', color='#0066ff', fontsize=14)
-    plt.ylabel('Node Count', color='#888')
-    plt.grid(axis='y', color='#333', linestyle='--', alpha=0.5)
-
-    plt.tight_layout()
-    plt.savefig('plots/commander_profiles.png', dpi=100)
-    print('✅ 军衔分布图封存成功。')
-
-
-if __name__ == '__main__':
-    plot_commander_profiles()

package/src/bootstrap/registry.js

@@ -1,14 +0,0 @@
-module.exports = ({ dir }) => {
-
-    const fs = require('fs')
-    const path = require('path')
-
-    return Object.fromEntries(
-        fs.readdirSync(dir)
-            .filter(file => file.endsWith('.js'))
-            .map(file => [
-                path.basename(file, '.js'),
-                require(path.join(dir, file))
-            ])
-    )
-}