Update dirStream.py

src/ms_ovba/Models/Entities/reference.py

@@ -32,16 +32,12 @@ def unpack(data: bytes, endien: str) -> Reference:
         endien_symbol = '<' if endien == 'little' else '>'
         name = ''
         offset = 0
-        id = struct.unpack_from(endien_symbol + "H", data, offset)
+        id, size1 = struct.unpack_from(f"{endien_symbol}HI", data, offset)
         if id == 0x0016:
-            offset += 2
-            size1, = struct.unpack_from(endien_symbol + "I", data, offset)
-            offset += 4
-            format = endien_symbol + size1 + "s"
-            name, = struct.unpack_from(format, data, offset)
-            offset += size1
-            size2 = struct.unpack_from(endien_symbol + "I", data, offset)
-            offset += 4
+            offset += 6
+            format = f"{endien_symbol}{size1}sHI"
+            name, id, size2 = struct.unpack_from(format, data, offset)
+            offset += size1 + 6
             if size2 != size1 * 2:
                 # raise warning
                 pass

src/ms_ovba/Models/Entities/reference_record.py

@@ -22,7 +22,7 @@ def unpack(bytestring: bytes, endien: str) -> ReferenceRecord:
             ReferenceOriginal
         )
         endien_symbol = '<' if endien == 'little' else '>'
-        id = struct.unpack(endien_symbol + "H", bytestring)
+        id, = struct.unpack_from(f"{endien_symbol}H", bytestring, 0)
         ref: ReferenceRecord
         if id == 0x000D:
             ref = ReferenceRegistered.unpack(bytestring, endien)
@@ -33,5 +33,5 @@ def unpack(bytestring: bytes, endien: str) -> ReferenceRecord:
         elif id == 0x0033:
             ref = ReferenceOriginal.unpack(bytestring, endien)
         else:
-            raise Exception("Unknown Reference Type")
+            raise Exception(f"Unknown Reference Type: {id}")
         return ref

src/ms_ovba/Views/dirStream.py

@@ -1,12 +1,14 @@
 import struct
+import warnings
 from ms_ovba_compression.ms_ovba import MsOvba
 from ms_ovba.vbaProject import VbaProject
+from ms_ovba.Models.Entities.reference import Reference
 from ms_ovba.Models.Fields.idSizeField import IdSizeField
 from ms_ovba.Models.Fields.doubleEncodedString import (
     DoubleEncodedString
 )
 from ms_ovba.Models.Fields.packed_data import PackedData
-from typing import List, TypeVar
+from typing import List, TypedDict, TypeVar
 
 
 T = TypeVar('T', bound='DirStream')
@@ -15,6 +17,14 @@
 PackableData = DoubleEncodedString | IdSizeField | PackedData
 
 
+class Parameters(TypedDict):
+    references: list[Reference]
+    modules: list
+    help_context_id: int
+    project_cookie: int
+    codepage_name: str
+
+
 class DirStream():
     """
     The dir stream is compressed on write
@@ -93,3 +103,183 @@ def _load_information(self: T) -> List:
             constants
         ])
         return information
+
+    @staticmethod
+    def is_valid(data: bytes) -> bool:
+        try:
+            DirStream.from_bytes(data)
+        except Exception as e:
+            warnings.warn(str(e), SyntaxWarning)
+            return False
+        return True
+
+    @staticmethod
+    def from_bytes(data: bytes) -> Parameters:
+        """
+        Static validation: Checks if bytes follow the DirStream structure.
+        Expects decompressed bytes.
+        """
+        project_data: Parameters = {
+            "references": [],
+            "modules": [],
+            "help_context_id": 0,
+            "project_cookie": 0,
+            "codepage_name": "cp1252"
+        }
+        offset = 0
+
+        # 1. Check PROJECTSYSKIND (Mandatory first record)
+        # IdSizeField(1, 4, 3) -> ID=1 (2 bytes), Size=4 (4 bytes)
+        record_id, size, value = struct.unpack_from("<HII", data, offset)
+        if record_id != 1 or size != 4 or not (0 <= value <= 3):
+            raise ValueError("Incorrect PROJECTSYSKIND")
+        offset += 10
+
+        record_id, size, value = struct.unpack_from("<HII", data, offset)
+        if record_id == 0x4A:
+            if size != 4:
+                raise ValueError("Incorrect PROJECTCOMPATVERSION")
+            offset += 10
+            record_id, size, value = (struct.unpack_from("<HII", data, offset))
+        if record_id != 2 or size != 4 or value != 0x409:
+            raise ValueError("Incorrect PROJECTLCID")
+        offset += 10
+
+        record_id, size, value = struct.unpack_from("<HII", data, offset)
+        if record_id != 0x14 or size != 4 or value != 0x409:
+            raise ValueError("Incorrect PROJECTLCIDINVOKE")
+        offset += 10
+
+        record_id, size, value = struct.unpack_from("<HIH", data, offset)
+        if record_id != 3 or size != 2:
+            raise ValueError("Incorrect PROJECTCODEPAGE")
+        offset += 8
+
+        record_id, size = struct.unpack_from("<HI", data, offset)
+        value, = struct.unpack_from(f"{size}s", data, offset + 6)
+        if record_id != 4 or not (1 <= size <= 128):
+            raise ValueError("Incorrect PROJECTNAME")
+        offset += 6 + size
+
+        record_id, size = struct.unpack_from("<HI", data, offset)
+        value, r2, s2, v2 = struct.unpack_from(
+            f"<{size}sHI{2*size}s", data, offset + 6)
+        if record_id != 5 or size > 2000 or s2 != 2 * size:
+            raise ValueError(
+                f"Incorrect PROJECTDOCSTRING({record_id}, {size}, {r2}, {s2})")
+        offset += 12 + size * 3
+
+        record_id, size = struct.unpack_from("<HI", data, offset)
+        value, r2, s2, v2 = struct.unpack_from(
+            f"<{size}sHI{size}s", data, offset + 6)
+        if record_id != 6 or size > 260 or s2 != size or value != v2:
+            raise ValueError(
+                f"Incorrect PROJECTHELPFILEPATH({record_id}, {size}, {s2})")
+        offset += 12 + size * 2
+
+        record_id, size, value = struct.unpack_from("<HII", data, offset)
+        if record_id != 7 or size != 4:
+            raise ValueError("Incorrect PROJECTHELPCONTEXT")
+        offset += 10
+
+        record_id, size, value = struct.unpack_from("<HII", data, offset)
+        if record_id != 8 or size != 4 or value != 0:
+            raise ValueError("Incorrect PROJECTLIBFLAGS")
+        offset += 10
+
+        record_id, size, value, v2 = struct.unpack_from("<HIIH", data, offset)
+        if record_id != 9:
+            raise ValueError("Incorrect PROJECTVERSION")
+        offset += 12
+
+        record_id, size = struct.unpack_from("<HI", data, offset)
+        value, r2, s2, v2 = struct.unpack_from(
+            f"<{size}sHI{size}s", data, offset + 6)
+        if record_id != 0x0c or size > 2015 or s2 != 2 * size:
+            raise ValueError("Incorrect PROJECTCONSTANTS")
+        offset += 12 + size * 3
+
+        record_id, size = struct.unpack_from("<HI", data, offset)
+        found_one_reference = False
+        while (record_id != 0x0f or not found_one_reference):
+            found_one_reference = True
+            record_size = 0
+            if record_id == 0x16:
+                record_size = 12 + size * 3
+                record_id, size = struct.unpack_from(
+                    "<HI", data, offset + record_size)
+            match record_id:
+                case 0x2f:
+                    record_size += 6 + size
+                    record_id, size = struct.unpack_from(
+                        "<HI", data, offset + record_size)
+                    if record_id == 0x16:
+                        record_size = 12 + size * 3
+                    record_id, size = struct.unpack_from(
+                        "<HI", data, offset + record_size)
+                    record_size += 6 + size
+                case 0x33:
+                    record_size += 6 + size
+                    record_id, size = struct.unpack_from(
+                        "<HI", data, offset + record_size)
+                    record_size += 6 + size
+                    record_id, size = struct.unpack_from(
+                        "<HI", data, offset + record_size)
+                    if record_id == 0x16:
+                        record_size = 12 + size * 3
+                        record_id, size = struct.unpack_from(
+                            "<HI", data, offset + record_size)
+                        record_size += 6 + size
+                case 0x0d | 0x0e:
+                    record_size += 6 + size
+                case _:
+                    raise ValueError(f"Unknown Reference Type: {record_id}")
+            ref = Reference.unpack(
+                data[offset:offset + record_size], "little")
+            project_data["references"] += [ref]
+            offset += record_size
+            record_id, size = struct.unpack_from("<HI", data, offset)
+
+        if record_id != 0x0f or size != 2:
+            raise ValueError("Expected ModuleRecord")
+        count, record_id, size, cookie = struct.unpack_from(
+            "<HHIH", data, offset)
+        if record_id != 0x13 or size != 2:
+            raise ValueError(f"Incorrect PROJECTCOOKIE({record_id}, {size})")
+        project_data["project_cookie"] = cookie
+        for _ in range(count):
+            module_data = {}
+            record_size = 0
+            record_id, size = struct.unpack_from("<HI", data, offset)
+            record_size += 6
+            value, r2, s2, v2 = struct.unpack_from(
+                f"<{size}sHI{size*2}s", data, offset + record_size)
+            record_size += size * 3 + 6
+            # validate sizes and that values match
+            module_data["name"] = value
+            record_id, size = struct.unpack_from(
+                f"<{size}sHI", data, offset + record_size)
+            record_size += 6
+            value, record_id, size = struct.unpack_from(
+                f"<{size}sHI", data, offset + record_size)
+            module_data["stream_name"] = value
+            record_size += size + 6
+            value, record_id, size = struct.unpack_from(
+                f"<{size}sHI", data, offset + record_size)
+            module_data["docstring"] = value
+            record_size += size + 6
+            project_data["modules"] += [module_data]
+            offset += record_size
+        return project_data
+
+    @staticmethod
+    def is_file_valid(file_path: str) -> bool:
+        """Helper to validate a compressed .bin file on disk."""
+        try:
+            with open(file_path, "rb") as f:
+                compressed = f.read()
+            from ms_ovba_compression.ms_ovba import MsOvba
+            decompressed = MsOvba().decompress(compressed)
+            return DirStream.is_valid(decompressed)
+        except Exception:
+            return False

tests/Unit/Views/test_dir.py

@@ -0,0 +1,52 @@
+import pytest
+from pathlib import Path
+from ms_cfb.ole_file import OleFile
+from ms_ovba.Views.dirStream import DirStream
+from ms_ovba_compression.ms_ovba import MsOvba
+from unittest import mock
+
+
+mock_vbaproject = mock.Mock()
+
+
+@pytest.fixture
+def my_fixture() -> None:
+    # Setup: Runs BEFORE the test
+    # print("\nSetting up...")
+    yield
+    # Teardown: Runs AFTER the test
+    Path("tests/blank/dir.bin").unlink(missing_ok=True)
+
+
+def test_construct() -> None:
+    dir = DirStream(mock_vbaproject)
+    assert isinstance(dir, DirStream)
+
+
+@pytest.mark.usefixtures("my_fixture")
+def test_is_valid() -> None:
+    file = "tests/blank/vbaProject.bin"
+    ole_file = OleFile.create_from_file(file)
+    ole_file.extract_stream('dir', 'tests/blank')
+    with open('tests/blank/dir.bin', 'rb') as f:
+        compressed_data = f.read()
+
+        # Use MsOvba to decompress the stream
+        ms_ovba = MsOvba()
+        decompressed_data = ms_ovba.decompress(compressed_data)
+        assert DirStream.is_valid(decompressed_data)
+
+
+@pytest.mark.usefixtures("my_fixture")
+def test_from_bytes() -> None:
+    expected = {}
+    file = "tests/blank/vbaProject.bin"
+    ole_file = OleFile.create_from_file(file)
+    ole_file.extract_stream('dir', 'tests/blank')
+    with open('tests/blank/dir.bin', 'rb') as f:
+        compressed_data = f.read()
+
+        # Use MsOvba to decompress the stream
+        ms_ovba = MsOvba()
+        decompressed_data = ms_ovba.decompress(compressed_data)
+        assert DirStream.from_bytes(decompressed_data) == expected

tests/blank/dir

@@ -1,14 +1,19 @@
-00000000   01 00 04 00 00 00 03 00 00 00 4A 00 04 00 00 00  ..........J.....
-00000010   03 00 00 00 02 00 04 00 00 00 09 04 00 00 14 00  ................
-00000020   04 00 00 00 09 04 00 00 03 00 02 00 00 00 E4 04  ..............ä.
-00000030   04 00 0A 00 00 00 56 42 41 50 72 6F 6A 65 63 74  ......VBAProject
-00000040   05 00 00 00 00 00 40 00 00 00 00 00 06 00 00 00  ......@.........
-00000050   00 00 3D 00 00 00 00 00 07 00 04 00 00 00 00 00  ..=.............
-00000060   00 00 08 00 04 00 00 00 00 00 00 00 09 00 04 00  ................
-00000070   00 00 57 02 BE 65 11 00 0C 00 00 00 00 00 3C 00  ..W.¾e........<.
-00000080   00 00 00 00 16 00 06 00 00 00 73 74 64 6F 6C 65  ..........stdole References begin at 84
-00000090   3E 00 0C 00 00 00 73 00 74 00 64 00 6F 00 6C 00  >.....s.t.d.o.l.
-000000A0   65 00 0D 00 68 00 00 00 5E 00 00 00 2A 5C 47 7B  e...h...^...*\G{
+01 00 04 00 00 00 03 00 00 00                    ..........
+4A 00 04 00 00 00 03 00 00 00                    J.........
+02 00 04 00 00 00 09 04 00 00                    ..........
+14 00 04 00 00 00 09 04 00 00                    ..........
+03 00 02 00 00 00 E4 04                          ......ä.
+04 00 0A 00 00 00 56 42 41 50 72 6F 6A 65 63 74  ......VBAProject
+05 00 00 00 00 00 40 00 00 00 00 00              ......@.....
+06 00 00 00 00 00 3D 00 00 00 00 00              ......=.....
+07 00 04 00 00 00 00 00 00 00                    ..........
+08 00 04 00 00 00 00 00 00 00                    ..........
+09 00 04 00 00 00 57 02 BE 65 11 00              ......W.¾e..
+0C 00 00 00 00 00 3C 00 00 00 00 00              ......<.....
+
+16 00 06 00 00 00 73 74 64 6F 6C 65 3E 00 0C 00  ......stdole>...
+00 00 73 00 74 00 64 00 6F 00 6C 00 65 00        ..s.t.d.o.l.e.
+0D 00 68 00 00 00 5E 00 00 00 2A 5C 47 7B        ..h...^...*\G{
 000000B0   30 30 30 32 30 34 33 30 2D 30 30 30 30 2D 30 30  00020430-0000-00
 000000C0   30 30 2D 43 30 30 30 2D 30 30 30 30 30 30 30 30  00-C000-00000000
 000000D0   30 30 34 36 7D 23 32 2E 30 23 30 23 43 3A 5C 57  0046}#2.0#0#C:\W