fix: DrHider ZIP expansion — extract archives before obfuscation

deploy/services/drhider.py

@@ -190,6 +190,9 @@ class TwoPassObfuscator:
         """
         self._workdir = tempfile.mkdtemp(prefix="drhider_")
         try:
+            # --- Распаковать ZIP-файлы ---
+            files = self._expand_zips(files)
+
             # --- Проход 1: сбор сущностей ---
             all_texts: Dict[str, str] = {}            # filename → text for LLM NER
             all_docx: Dict[str, object] = {}          # filename → docx Document for replacement
@@ -231,6 +234,43 @@ class TwoPassObfuscator:
             self._mapping.clear()
             self._regex_replacements.clear()
 
+    def _expand_zips(self, files: List[Tuple[str, bytes, str]]) -> List[Tuple[str, bytes, str]]:
+        """Распаковать ZIP-файлы, заменив их содержимым. Остальные файлы — как есть."""
+        result = []
+        for fname, content, ctype in files:
+            if fname.lower().endswith('.zip'):
+                try:
+                    with zipfile.ZipFile(io.BytesIO(content)) as zf:
+                        total_size = sum(info.file_size for info in zf.infolist())
+                        if total_size > 500 * 1024 * 1024:  # 500 MB
+                            log.warning("ZIP too large, skipping expansion: %s", fname)
+                            result.append((fname, content, ctype))
+                            continue
+                        if len(zf.infolist()) > 500:
+                            log.warning("ZIP too many files, skipping: %s", fname)
+                            result.append((fname, content, ctype))
+                            continue
+                        for info in zf.infolist():
+                            if info.is_dir():
+                                continue
+                            # cp437 → utf8 для имён
+                            try:
+                                inner_name = info.filename.encode('cp437').decode('utf-8')
+                            except (UnicodeDecodeError, UnicodeEncodeError):
+                                inner_name = info.filename
+                            # Убрать path traversal
+                            inner_name = os.path.basename(inner_name)
+                            if not inner_name:
+                                continue
+                            inner_data = zf.read(info)
+                            result.append((inner_name, inner_data, ""))
+                except Exception as e:
+                    log.warning("Failed to expand ZIP %s: %s", fname, e)
+                    result.append((fname, content, ctype))
+            else:
+                result.append((fname, content, ctype))
+        return result
+
     # --- Проход 1: обнаружение ---
 
     def _extract_text(self, fname: str, content: bytes, ctype: str) -> Tuple[str, Optional[object]]:

site/services/drhider.py

@@ -190,6 +190,9 @@ class TwoPassObfuscator:
         """
         self._workdir = tempfile.mkdtemp(prefix="drhider_")
         try:
+            # --- Распаковать ZIP-файлы ---
+            files = self._expand_zips(files)
+
             # --- Проход 1: сбор сущностей ---
             all_texts: Dict[str, str] = {}            # filename → text for LLM NER
             all_docx: Dict[str, object] = {}          # filename → docx Document for replacement
@@ -231,6 +234,43 @@ class TwoPassObfuscator:
             self._mapping.clear()
             self._regex_replacements.clear()
 
+    def _expand_zips(self, files: List[Tuple[str, bytes, str]]) -> List[Tuple[str, bytes, str]]:
+        """Распаковать ZIP-файлы, заменив их содержимым. Остальные файлы — как есть."""
+        result = []
+        for fname, content, ctype in files:
+            if fname.lower().endswith('.zip'):
+                try:
+                    with zipfile.ZipFile(io.BytesIO(content)) as zf:
+                        total_size = sum(info.file_size for info in zf.infolist())
+                        if total_size > 500 * 1024 * 1024:  # 500 MB
+                            log.warning("ZIP too large, skipping expansion: %s", fname)
+                            result.append((fname, content, ctype))
+                            continue
+                        if len(zf.infolist()) > 500:
+                            log.warning("ZIP too many files, skipping: %s", fname)
+                            result.append((fname, content, ctype))
+                            continue
+                        for info in zf.infolist():
+                            if info.is_dir():
+                                continue
+                            # cp437 → utf8 для имён
+                            try:
+                                inner_name = info.filename.encode('cp437').decode('utf-8')
+                            except (UnicodeDecodeError, UnicodeEncodeError):
+                                inner_name = info.filename
+                            # Убрать path traversal
+                            inner_name = os.path.basename(inner_name)
+                            if not inner_name:
+                                continue
+                            inner_data = zf.read(info)
+                            result.append((inner_name, inner_data, ""))
+                except Exception as e:
+                    log.warning("Failed to expand ZIP %s: %s", fname, e)
+                    result.append((fname, content, ctype))
+            else:
+                result.append((fname, content, ctype))
+        return result
+
     # --- Проход 1: обнаружение ---
 
     def _extract_text(self, fname: str, content: bytes, ctype: str) -> Tuple[str, Optional[object]]: