refactor: centralize configuration management with type-safe schema

Major changes:
- Add dataclass-based configuration schema with validation (config/schema.py)
- Create centralized config loader and default values (config/defaults.py, config/loader.py)
- Remove duplicate database_config.py, merge into unified structure
- Consolidate browser settings into ERP config
- Add batch_size parameter support to extractor

Bug fixes:
- Fix settings save error by updating config paths (browser.* → erp.*)
- Fix batch_size not being applied in data extraction

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

gui/data_extraction_tab.py

@@ -114,7 +114,7 @@ class DataExtractionTab(ttk.Frame):
         self.verbose_var = tk.BooleanVar(value=self.config.get('extraction.verbose', True))
         ttk.Checkbutton(options_group, text="详细日志", variable=self.verbose_var).grid(row=0, column=0, sticky="w", padx=5)
 
-        self.headless_var = tk.BooleanVar(value=self.config.get('browser.headless', True))
+        self.headless_var = tk.BooleanVar(value=self.config.get('erp.headless', True))
         ttk.Checkbutton(options_group, text="无头模式 (不显示浏览器)", variable=self.headless_var).grid(row=0, column=1, sticky="w", padx=5)
 
         # 进度显示
@@ -200,7 +200,8 @@ class DataExtractionTab(ttk.Frame):
                 username=self.config.get('erp.username'),
                 password=self.config.get('erp.password'),
                 headless=self.headless_var.get(),
-                verbose=self.verbose_var.get()
+                verbose=self.verbose_var.get(),
+                batch_size=self.config.get('extraction.batch_size', 100)
             )
 
             # 创建实时输出流，每次写入立即更新 GUI