feat: implementar Fase 2 del roadmap (Modo Replica Exacta / Mirroring)

windows-agent/agent/scanner.py

@@ -45,12 +45,28 @@ def is_file_stable(filepath: Path, min_stable_seconds: int = 60, sample_interval
     except Exception:
         return False
 
+import platform
+import stat
+
 class DirectoryScanner:
     """Scans Windows source paths for files matching specific backup patterns."""
 
-    def __init__(self, source_path: str, file_patterns: str = "*.bak,*.mdf", min_stable_seconds: int = 60):
+    def __init__(
+        self,
+        source_path: str,
+        file_patterns: str = "*.bak,*.mdf",
+        exclude_patterns: str = "",
+        skip_hidden: bool = False,
+        skip_system: bool = False,
+        skip_readonly: bool = False,
+        min_stable_seconds: int = 60
+    ):
         self.source_path = Path(source_path).resolve()
         self.patterns = [p.strip() for p in file_patterns.split(",") if p.strip()]
+        self.exclude_patterns = [p.strip() for p in exclude_patterns.split(",") if p.strip()]
+        self.skip_hidden = skip_hidden
+        self.skip_system = skip_system
+        self.skip_readonly = skip_readonly
         self.min_stable_seconds = min_stable_seconds
 
     def scan(self) -> List[Path]:
@@ -61,15 +77,20 @@ class DirectoryScanner:
         matched_files: List[Path] = []
 
         if self.source_path.is_file():
-            if self._matches_patterns(self.source_path.name):
-                matched_files.append(self.source_path)
+            if self._matches_patterns(self.source_path.name) and not self._matches_exclude(self.source_path.name):
+                if not self._should_skip_by_attributes(self.source_path):
+                    matched_files.append(self.source_path)
             return matched_files
 
-        for root, _, files in os.walk(self.source_path):
+        for root, dirs, files in os.walk(self.source_path):
+            # Prune excluded directories from search in-place
+            dirs[:] = [d for d in dirs if not self._matches_exclude(d)]
+            
             for file in files:
-                if self._matches_patterns(file):
+                if self._matches_patterns(file) and not self._matches_exclude(file):
                     full_path = Path(root) / file
-                    matched_files.append(full_path)
+                    if not self._should_skip_by_attributes(full_path):
+                        matched_files.append(full_path)
 
         return matched_files
 
@@ -80,3 +101,29 @@ class DirectoryScanner:
             if fnmatch.fnmatch(filename.lower(), pattern.lower()):
                 return True
         return False
+
+    def _matches_exclude(self, filename: str) -> bool:
+        for pattern in self.exclude_patterns:
+            if fnmatch.fnmatch(filename.lower(), pattern.lower()):
+                return True
+        return False
+
+    def _should_skip_by_attributes(self, filepath: Path) -> bool:
+        if platform.system() != "Windows":
+            return False
+            
+        try:
+            st = filepath.stat()
+            attrs = getattr(st, "st_file_attributes", 0)
+            if not attrs:
+                return False
+                
+            if self.skip_readonly and (attrs & 0x01): # FILE_ATTRIBUTE_READONLY
+                return True
+            if self.skip_hidden and (attrs & 0x02):   # FILE_ATTRIBUTE_HIDDEN
+                return True
+            if self.skip_system and (attrs & 0x04):   # FILE_ATTRIBUTE_SYSTEM
+                return True
+        except Exception:
+            pass
+        return False