refactor: 优化 chardet 依赖配置，仅保留在 HTML reader 中

- 从 pdf/docx/xlsx/pptx reader 的依赖列表中移除 chardet
- 保留 chardet 在 html reader 的依赖配置中（唯一实际使用方）
- 更新 README.md 文档，移除不必要的 chardet 依赖说明
- 简化测试命令，移除非 HTML reader 测试中的 chardet

scripts/config.py

@@ -30,8 +30,7 @@ DEPENDENCIES = {
                 "unstructured[pdf]",
                 "markitdown[pdf]",
                 "pypdf",
-                "markdownify",
-                "chardet"
+                "markdownify"
             ]
         },
         "Darwin-x86_64": {
@@ -42,8 +41,7 @@ DEPENDENCIES = {
                 "numpy<2",
                 "markitdown[pdf]",
                 "pypdf",
-                "markdownify",
-                "chardet"
+                "markdownify"
             ]
         }
     },
@@ -56,8 +54,7 @@ DEPENDENCIES = {
                 "markitdown[docx]",
                 "pypandoc-binary",
                 "python-docx",
-                "markdownify",
-                "chardet"
+                "markdownify"
             ]
         }
     },
@@ -69,8 +66,7 @@ DEPENDENCIES = {
                 "unstructured[xlsx]",
                 "markitdown[xlsx]",
                 "pandas",
-                "tabulate",
-                "chardet"
+                "tabulate"
             ]
         }
     },
@@ -82,8 +78,7 @@ DEPENDENCIES = {
                 "unstructured[pptx]",
                 "markitdown[pptx]",
                 "python-pptx",
-                "markdownify",
-                "chardet"
+                "markdownify"
             ]
         }
     },