添加 Box 类及其子类 ItemLocationBox 的实现，包含初始化和解析逻辑

base.py

@@ -0,0 +1,15 @@
+from typing import List
+
+class Box:
+    """
+    Represents a generic ISOBMFF box. This is a simple data container.
+    """
+    def __init__(self, size: int, box_type: str, offset: int, raw_data: bytes):
+        self.size = size
+        self.type = box_type
+        self.offset = offset
+        self.raw_data = raw_data
+        self.children: List['Box'] = [] # Initially empty
+
+    def __repr__(self) -> str:
+        return f"<Box '{self.type}' size={self.size} offset={self.offset}>"

heic_types.py

@@ -0,0 +1,48 @@
+# In heic_types.py
+import struct
+from base import Box  # <-- THIS IS THE KEY CHANGE
+
+class ItemLocation:
+    # ... (rest of this class is unchanged)
+    def __init__(self, item_id, offset, length):
+        self.item_id = item_id
+        self.offset = offset
+        self.length = length
+    def __repr__(self):
+        return f"<ItemLocation ID={self.item_id} offset={self.offset} length={self.length}>"
+
+class ItemLocationBox(Box):
+    # ... (rest of this class is unchanged)
+    def __init__(self, size: int, box_type: str, offset: int, raw_data: bytes):
+        super().__init__(size, box_type, offset, raw_data)
+        self.locations = []
+        self._parse_locations()
+    def _parse_locations(self):
+        # ... (all the parsing logic here is unchanged)
+        stream = self.raw_data
+        version_flags = struct.unpack('>I', stream[:4])[0]
+        sizes = struct.unpack('>H', stream[4:6])[0]
+        offset_size = (sizes >> 12) & 0x0F
+        length_size = (sizes >> 8) & 0x0F
+        base_offset_size = (sizes >> 4) & 0x0F
+        item_count = struct.unpack('>H', stream[6:8])[0]
+        current_pos = 8
+        for _ in range(item_count):
+            item_id = struct.unpack('>H', stream[current_pos : current_pos+2])[0]
+            current_pos += 4 # Skip item_id and construction_method
+            current_pos += 2 # Skip data_reference_index
+            current_pos += base_offset_size
+            extent_count = struct.unpack('>H', stream[current_pos : current_pos+2])[0]
+            current_pos += 2
+            if extent_count > 0:
+                offset = self._read_int(stream, current_pos, offset_size)
+                current_pos += offset_size
+                length = self._read_int(stream, current_pos, length_size)
+                current_pos += length_size
+                self.locations.append(ItemLocation(item_id, offset, length))
+    def _read_int(self, data, pos, size):
+        if size == 1: return data[pos]
+        if size == 2: return struct.unpack('>H', data[pos:pos+2])[0]
+        if size == 4: return struct.unpack('>I', data[pos:pos+4])[0]
+        if size == 8: return struct.unpack('>Q', data[pos:pos+8])[0]
+        return 0

main.py

@@ -1,13 +1,30 @@
+from base import Box
 from parser import parse_boxes
+from heic_types import ItemLocationBox
+from typing import List
+import os
+
+def print_box_tree(boxes: List[Box], indent: int = 0):
+    for box in boxes:
+        print("  " * indent + str(box))
+        
+        if isinstance(box, ItemLocationBox):
+            for loc in box.locations[:5]:
+                print("  " * (indent + 1) + str(loc))
+            if len(box.locations) > 5:
+                print("  " * (indent + 1) + f"... and {len(box.locations) - 5} more locations")
+
+        if box.children:
+            print_box_tree(box.children, indent + 1)
 
 def main():
     # Test with the Apple HEIC file
     print("--- Analyzing Apple HEIC file ---")
     try:
         with open('apple.heic', 'rb') as f:
-            top_level_boxes = parse_boxes(f)
+            file_size = os.fstat(f.fileno()).st_size
-            for box in top_level_boxes:
+            top_level_boxes = parse_boxes(f, file_size)
-                print(box)
+            print_box_tree(top_level_boxes)
     except FileNotFoundError:
         print("Error: apple.heic not found.")
     
@@ -17,9 +34,9 @@ def main():
     print("--- Analyzing Samsung HEIC file ---")
     try:
         with open('samsung.heic', 'rb') as f:
-            top_level_boxes = parse_boxes(f)
+            file_size = os.fstat(f.fileno()).st_size
-            for box in top_level_boxes:
+            top_level_boxes = parse_boxes(f, file_size)
-                print(box)
+            print_box_tree(top_level_boxes)
     except FileNotFoundError:
         print("Error: samsung.heic not found.")
 

parser.py

@@ -1,78 +1,67 @@
 import struct
-
-class Box:
-    """
-    Represents a generic ISOBMFF box.
-    """
-    def __init__(self, size: int, box_type: str, offset: int, raw_data: bytes):
-        self.size = size
-        self.type = box_type
-        self.offset = offset
-        self.raw_data = raw_data
-        self.children = [] # Will be populated for container boxes
-
-    def __repr__(self) -> str:
-        # e.g., <Box 'ftyp' size=36 offset=8>
-        return f"<Box '{self.type}' size={self.size} offset={self.offset}>"
-
-# In parser.py
-
 from typing import List, BinaryIO
+from io import BytesIO
 
-def parse_boxes(stream: BinaryIO) -> List[Box]:
+from base import Box  # Import the base Box class
+from heic_types import ItemLocationBox
+
+# The factory map remains here
+BOX_TYPE_MAP = {
+    'iloc': ItemLocationBox,
+}
+
+CONTAINER_BOXES = {'meta', 'moov', 'trak', 'iprp', 'ipco', 'dinf', 'fiinf', 'ipro'}
+FULL_BOXES = {'meta', 'hdlr', 'pitm', 'iinf', 'iloc'}
+
+def parse_boxes(stream: BinaryIO, max_size: int) -> List[Box]:
     """
-    Parses all top-level boxes from a file stream, handling special size cases.
+    Parses boxes from a file stream up to a maximum size.
+    Now also handles recursive parsing.
     """
     boxes = []
+    start_pos_in_stream = stream.tell()
     
-    stream.seek(0, 2)
+    while stream.tell() - start_pos_in_stream < max_size:
-    file_size = stream.tell()
+        current_offset_in_stream = stream.tell()
-    stream.seek(0)
-
-    while stream.tell() < file_size:
-        current_offset = stream.tell()
         
-        # Read the standard 8-byte header
         header = stream.read(8)
-        if len(header) < 8:
+        if len(header) < 8: break
-            break
 
         size = struct.unpack('>I', header[:4])[0]
-        box_type = header[4:].decode('ascii')
+        box_type = header[4:].decode('ascii', errors='ignore')
         
         header_size = 8
-        content_size = 0
-
-        # --- NEW: Handle special size cases ---
         if size == 1:
-            # The actual size is a 64-bit integer following the type
             largesize_header = stream.read(8)
-            if len(largesize_header) < 8:
+            if len(largesize_header) < 8: break
-                break
             size = struct.unpack('>Q', largesize_header)[0]
             header_size = 16
-            content_size = size - header_size
         elif size == 0:
-            # The box extends to the end of the file
+            size = max_size - (current_offset_in_stream - start_pos_in_stream)
-            content_size = file_size - current_offset - header_size
-            size = content_size + header_size
-        else:
-            # Standard size
-            content_size = size - header_size
-        # --- End of new code ---
 
-        # Read the box's content (raw_data)
+        if size < header_size: break
-        # We need to rewind a bit if we read largesize header to get all content
+            
-        stream.seek(current_offset + header_size)
+        content_size = size - header_size
-        raw_data = stream.read(content_size)
-        if len(raw_data) < content_size:
-            # Avoids errors on truncated files
-            break
-
-        box = Box(size, box_type, current_offset, raw_data)
-        boxes.append(box)
         
-        # Seek to the beginning of the next box
+        stream.seek(current_offset_in_stream + header_size)
-        stream.seek(current_offset + size)
+        raw_data = stream.read(content_size)
+        if len(raw_data) < content_size: break
+
+        # --- Factory Pattern ---
+        box_class = BOX_TYPE_MAP.get(box_type, Box)
+        box = box_class(size, box_type, current_offset_in_stream, raw_data)
+
+        # --- NEW: Recursive parsing logic moved here ---
+        if box.type in CONTAINER_BOXES:
+            child_stream = BytesIO(box.raw_data)
+            parse_size = len(box.raw_data)
+            if box.type in FULL_BOXES:
+                child_stream.read(4) # Skip version/flags
+                parse_size -= 4
+            box.children = parse_boxes(child_stream, parse_size)
+        # --- End of new logic ---
+
+        boxes.append(box)
+        stream.seek(current_offset_in_stream + size)
 
     return boxes