Merge pull request #137 from ferstar/master

Restore reader position after retrieving signature bytes

Author: h2non

filetype/types/document.py

@@ -67,8 +67,10 @@ def match_document(self, buf):
             return
 
         # Loop through next 3 files and check if they match
+        # NOTE: OpenOffice/Libreoffice orders ZIP entry differently, so check the 4th file
+        # https://github.com/h2non/filetype/blob/d730d98ad5c990883148485b6fd5adbdd378364a/matchers/document.go#L134
         idx = 0
-        for i in range(3):
+        for i in range(4):
             # Search for next file header
             idx = self.search_signature(buf, idx + 4, 6000)
             if idx == -1:
@@ -110,7 +112,7 @@ def __init__(self):
 
     def match(self, buf):
         if len(buf) > 515 and buf[0:8] == b"\xD0\xCF\x11\xE0\xA1\xB1\x1A\xE1":
-            if buf[512:515] == b"\xEC\xA5\xC1\x00":
+            if buf[512:516] == b"\xEC\xA5\xC1\x00":
                 return True
             if (
                 len(buf) > 2142

filetype/utils.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 
+from io import BufferedIOBase
 # Python 2.7 workaround
 try:
     import pathlib
@@ -48,7 +49,8 @@ def get_bytes(obj):
     returning a sliced bytearray.
 
     Args:
-        obj: path to readable, file-like object(with read() method), bytes, bytearray or memoryview
+        obj: path to readable, file-like object(with read() method), bytes,
+        bytearray or memoryview
 
     Returns:
         First 8192 bytes of the file content as bytearray type.
@@ -71,6 +73,13 @@ def get_bytes(obj):
     if isinstance(obj, pathlib.PurePath):
         return get_signature_bytes(obj)
 
+    if isinstance(obj, BufferedIOBase):
+        start_pos = obj.tell()
+        obj.seek(0)
+        magic_bytes = obj.read(_NUM_SIGNATURE_BYTES)
+        obj.seek(start_pos)  # restore reader position
+        return get_bytes(magic_bytes)
+
     if hasattr(obj, 'read'):
         return get_bytes(obj.read(_NUM_SIGNATURE_BYTES))
 

tests/fixtures/sample_1.doc

@@ -14,8 +14,14 @@
 class TestFileType(unittest.TestCase):
     def test_guess_jpeg(self):
         img_path = FIXTURES + '/sample.jpg'
-        for obj in (img_path, open(img_path, 'rb')):
-            kind = filetype.guess(obj)
+        with open(img_path, 'rb') as fp:
+            for obj in (img_path, fp):
+                kind = filetype.guess(obj)
+                self.assertTrue(kind is not None)
+                self.assertEqual(kind.mime, 'image/jpeg')
+                self.assertEqual(kind.extension, 'jpg')
+            # reset reader position test
+            kind = filetype.guess(fp)
             self.assertTrue(kind is not None)
             self.assertEqual(kind.mime, 'image/jpeg')
             self.assertEqual(kind.extension, 'jpg')
@@ -70,16 +76,18 @@ def test_guess_zstd(self):
             self.assertEqual(kind.extension, 'zst')
 
     def test_guess_doc(self):
-        kind = filetype.guess(FIXTURES + '/sample.doc')
-        self.assertIsNotNone(kind)
-        self.assertEqual(kind.mime, 'application/msword')
-        self.assertEqual(kind.extension, 'doc')
+        for name in 'sample.doc', 'sample_1.doc':
+            kind = filetype.guess(os.path.join(FIXTURES, name))
+            self.assertIsNotNone(kind)
+            self.assertEqual(kind.mime, 'application/msword')
+            self.assertEqual(kind.extension, 'doc')
 
     def test_guess_docx(self):
-        kind = filetype.guess(FIXTURES + '/sample.docx')
-        self.assertTrue(kind is not None)
-        self.assertEqual(kind.mime, 'application/vnd.openxmlformats-officedocument.wordprocessingml.document')
-        self.assertEqual(kind.extension, 'docx')
+        for name in 'sample.docx', 'sample_1.docx':
+            kind = filetype.guess(os.path.join(FIXTURES, name))
+            self.assertTrue(kind is not None)
+            self.assertEqual(kind.mime, 'application/vnd.openxmlformats-officedocument.wordprocessingml.document')
+            self.assertEqual(kind.extension, 'docx')
 
     def test_guess_odt(self):
         kind = filetype.guess(FIXTURES + '/sample.odt')

tests/fixtures/sample_1.docx

@@ -8,8 +8,8 @@ envlist = py{27,35,36,37,38,39}, lint, doc, clean
 skip_missing_interpreters = true
 
 [testenv:test]
-deps = pytest-benchmark
-commands = pytest
+deps = pytest
+commands = pytest --ignore=tests/test_benchmark.py
 
 [testenv:lint]
 basepython = python3