Version 1.27 (#98)

- Adding new verbose output to command line with `-v` or `--verbose`
- Adding #92 include py.typed in sdist (thanks to Nicholas Bollweg - bollwyvl)
- Adding #93 Improve PDF file detection, fix json description (thanks to Péter - peterekepeter)
- Fixing #96 #86 stream does not work properly on opened small files (thanks to Felipe Lema and Andy - NebularNerd)
- Removing expected invalid WinZip signature

---------

Co-authored-by: Nicholas Bollweg <[email protected]>
Co-authored-by: Péter <[email protected]>
Co-authored-by: Andy <[email protected]>

Author: 4 people

.pre-commit-config.yaml

@@ -24,28 +24,33 @@ repos:
       exclude: ^test/resources/
     - id: trailing-whitespace
       args: [--markdown-linebreak-ext=md]
-      exclude: ^test/resources/
+      exclude: |
+            (?x)^(
+                ^test/resources/.+|
+                ^puremagic/magic_data.json
+            )$
     - id: check-executables-have-shebangs
     - id: end-of-file-fixer
       exclude: ^test/resources/.+
 
+
 -   repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.4.6
+    rev: v0.5.7
     hooks:
       - id: ruff
 
 -   repo: https://github.com/ambv/black
-    rev: 24.4.2
+    rev: 24.8.0
     hooks:
     - id: black
 
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: 'v1.10.0'
+    rev: 'v1.11.1'
     hooks:
     - id: mypy
 
 -   repo: https://github.com/tox-dev/pyproject-fmt
-    rev: 2.1.3
+    rev: 2.2.1
     hooks:
       - id: pyproject-fmt
 

AUTHORS.rst

@@ -20,4 +20,7 @@ A big thank you to everyone that has helped!
 - Andy (NebularNerd)
 - Raphaël Vinot (Rafiot)
 - Sebastian Kreft (sk-)
-- William Bonnaventure  (Aztorius)
+- William Bonnaventure (Aztorius)
+- Nicholas Bollweg (bollwyvl)
+- Péter (peterekepeter)
+- mara004

CHANGELOG.md

@@ -1,6 +1,15 @@
 Changelog
 =========
 
+Version 1.27
+------------
+
+- Adding new verbose output to command line with `-v` or `--verbose`
+- Adding #92 include py.typed in sdist (thanks to Nicholas Bollweg - bollwyvl)
+- Adding #93 Improve PDF file detection, fix json description (thanks to Péter - peterekepeter)
+- Fixing #96 #86 stream does not work properly on opened small files (thanks to Felipe Lema and Andy - NebularNerd)
+- Removing expected invalid WinZip signature
+
 Version 1.26
 ------------
 

MANIFEST.in

@@ -1,4 +1,5 @@
 include puremagic/*.json
+include puremagic/py.typed
 include LICENSE
 include AUTHORS.rst
 include CHANGELOG.md

README.rst

@@ -35,7 +35,7 @@ Disadvantages:
 Compatibility
 ~~~~~~~~~~~~~
 
--  Python 3.8+
+-  Python 3.7+
 
 Using github ci to run continuous integration tests on listed platforms.
 
@@ -174,7 +174,7 @@ http://www.garykessler.net/library/file_sigs.html
 
 Freedesktop.org
 
-For use of their shared-mime-info file (even if they do use XML, blea), available at:
+For use of their shared-mime-info file, available at:
 https://cgit.freedesktop.org/xdg/shared-mime-info/
 
 License

puremagic/magic_data.json

@@ -20,14 +20,15 @@
     ["", 0, ".crt", "text/plain", "X.509 Certificate"],
     ["", 0, ".reg", "", "Windows Registry File"],
     ["", 0, ".md", "text/plain", "Markdown File"],
-    ["", 0, ".json", "application/json", "Markdown File"],
+    ["", 0, ".json", "application/json", "JSON File"],
     ["", 0, ".rst", "text/plain", "Restructured Text File"],
     ["", 0, ".cfg", "text/plain", "Configuration File"],
     ["", 0, ".flake8", "text/plain", "Flake 8 Configuration File"],
     ["", 0, ".coveragerc", "text/plain", "Coverage File"],
     ["", 0, ".c", "text/x-csrc", "C Code File"],
     ["", 0, ".cc", "text/x-csrc", "C Code File"],
     ["", 0, ".h", "text/x-csrc", "C Header File"],
+    ["", 0, ".pdf", "application/pdf", "Adobe Portable Document Format file"],
     ["", 0, ".stl", "model/stl", "stereolithography CAD software"],
     ["", 0, ".srt", "application/x-subrip", "SubRip subtitles"],
     ["", 0, ".obj", "", "Relocatable object code"],
@@ -63,9 +64,13 @@
     ["", 0, ".b6i", "", "BlindWrite 6 Image File"],
     ["", 0, ".cl2", "", "Adaptec Easy CD/DVD Creator image file"],
     ["", 0, ".cl3", "", "Adaptec Easy CD/DVD Creator image file"],
-    ["", 0, ".cl4", "", "Adaptec Easy CD/DVD Creator image file"]
+    ["", 0, ".cl4", "", "Adaptec Easy CD/DVD Creator image file"],
+    ["", 0, ".vba", "", "Visual Basic Script"],
+    ["", 0, "README", "text/plain", "README File"]
   ],
   "multi-part": {
+    "7b22": [["227d", -2, ".json", "application/json", "JSON File"]],
+    "7b": [["22", -1, ".json", "application/json", "JSON File"]],
     "464f524d": [
       ["494c424d", 8, ".iff", "image/x-ilbm", "IFF Interleaved Bitmap Image"],
       ["38535658", 8, ".iff", "audio/x-8svx", "IFF 8-Bit Sampled Voice"],
@@ -515,7 +520,7 @@
 	],
 	"73696262" : [
 		["72686c62", 8, ".uif", "", "MagicISO Disk Image (Encrypted)"]
-	]	
+	]
   },
   "footers": [
     ["54525545564953494f4e2d5846494c452e00", -18, ".tga", "image/tga", "Truevision Targa Graphic file"],
@@ -525,9 +530,19 @@
     ["3c2f7376673e", -6, ".svg", "image/svg+xml", "Scalable Vector Graphics Image"],
     ["6b6f6c79", -512, ".dmg", "application/x-apple-diskimage", "MacOS X image file"],
     ["4e45524f", -8, ".nrg", "", "Nero Disk Image (Version 1)"],
-    ["4e455235", -12, ".nrg", "", "Nero Disk Image (Version 2)"]	
+    ["4e455235", -12, ".nrg", "", "Nero Disk Image (Version 2)"]
   ],
   "headers": [
+    ["595556344d504547",0, ".y4m", "video/x-yuv4mpeg", "YUV4MPEG2 video file"],
+    ["3c68746d6c", 0, ".html", "text/html", "HTML File"],
+    ["424c5545", 0, ".bvr", "", "Blue Iris Video File"],
+    ["2d2d2d2d2d424547494e20504b4353372d2d2d2d2d", 0, ".p7b", "", "PKCS 7 Certificate File" ],
+    ["7b22", 0, ".json", "application/json", "JSON File"],
+    ["7b", 0, ".json", "application/json", "JSON File"],
+    ["50755454592d557365722d4b65792d46696c65", 0, ".ppk", "", "PuTTY User Key File"],
+    ["2d2d2d2d20424547494e2053534832205055424c4943204b4559202d2d2d2d", 0, "", "", "SSH Public Key"],
+    ["2d2d2d2d424547494e", 0, "", "", "Key or Cert File"],
+    ["2d2d2d2d20424547494e", 0, "", "", "Key or Cert File"],
     ["30313233343536373839", 0, ".puremagic_multi_footer", "text/ascii", "TESTFILE"],
     ["ff0a", 0, ".jxl", "image/jxl", "JPEG XL image (Raw stream)"],
     ["0000000c4a584c200d0a870a", 0, ".jxl", "image/jxl", "JPEG XL image (ISOBMFF container)"],
@@ -580,7 +595,7 @@
     ["425a68", 0, ".bzip2", "application/x-bzip2", "BZIP2 Compressed Archive file"],
     ["664c614300000022", 0, ".flac", "audio/flac", "Free Lossless Audio Codec file"],
     ["434f5744", 0, ".vmdk", "application/octet-stream", "VMware Sparse Extent Image file"],
-    ["23204469736b2044657363726970746f7246696c65", 0, ".vmdk", "application/octet-stream", "VMware Image Descriptor File"],	
+    ["23204469736b2044657363726970746f7246696c65", 0, ".vmdk", "application/octet-stream", "VMware Image Descriptor File"],
     ["4b444d56", 0, ".vmdk", "application/octet-stream", "VMware Virtual Single Disk file"],
     ["e310000100000000", 0, ".info", "", "Amiga icon"],
     ["5468697320697320", 0, ".info", "", "GNU Info Reader file"],
@@ -745,6 +760,7 @@
     ["000100004d534953414d204461746162617365", 0, ".mny", "application/x-msmoney", "Microsoft Money file"],
     ["000100005374616e64617264204a6574204442", 0, ".mdb", "application/x-msaccess", "Microsoft Access file"],
     ["25504446", 0, ".pdf", "application/pdf", "Adobe Portable Document Format file"],
+    ["0d0a25504446", 0, ".pdf", "application/pdf", "Adobe Portable Document Format file"],
     ["a0461df0", 512, ".ppt", "application/vnd.ms-powerpoint", "Microsoft Office PowerPoint Presentation file"],
     ["cf11e0a1b11ae100", 0, ".doc", "application/msword", "Perfect Office Document file"],
     ["d0cf11e0a1b11ae1", 0, ".doc", "application/msword", "Microsoft Office Document file"],
@@ -926,7 +942,6 @@
     ["564350434830", 0, ".pch", "", "Visual C PreCompiled header"],
     ["554641c6d2c1", 0, ".ufa", "", "UFA compressed archive"],
     ["ac9ebd8f0000", 0, ".qdf", "", "Quicken data"],
-    ["57696e5a6970", 29152, ".zip", "application/zip", "WinZip compressed archive"],
     ["504147454455", 0, ".dmp", "", "Windows memory dump"],
     ["4d444d5093a7", 0, ".dmp", "", "Windows dump file"],
     ["458600000600", 0, ".qbb", "", "QuickBooks backup"],
@@ -2006,7 +2021,7 @@
     ["5b436c6f6e6543445d", 0, ".ccd", "", "CloneCD Control File"],
     ["ffffffffffffffffffffffff", 0, ".sub", "", "CloneCD Sub Channel File"],
     ["00ffffffffffffffffffff", 0, ".img", "", "CloneCD Image File"],
-    ["f7fff9fffdfffbfff6fff7fff7fff5fff8fff7fff5fff0fffcfffafffafff7fff8fff6fff7fff7fff2fff2fff8", 0, ".img", "", "CloneCD Image File"],	
+    ["f7fff9fffdfffbfff6fff7fff7fff5fff8fff7fff5fff0fffcfffafffafff7fff8fff6fff7fff7fff2fff2fff8", 0, ".img", "", "CloneCD Image File"],
     ["ffffffffffffffffffffffff", 0, ".b5i", "", "BlindWrite 5 Image File"],
     ["425754352053545245414d205349474e", 0, ".b5t", "", "BlindWrite 5 Stream File"],
     ["425754352053545245414d205349474e", 0, ".b6t", "", "BlindWrite 6 Stream File"],

puremagic/main.py

@@ -21,7 +21,7 @@
 from itertools import chain
 
 __author__ = "Chris Griffith"
-__version__ = "1.26"
+__version__ = "1.27"
 __all__ = [
     "magic_file",
     "magic_string",
@@ -114,6 +114,9 @@ def _max_lengths() -> tuple[int, int]:
     return max_header_length, max_footer_length
 
 
+max_head, max_foot = _max_lengths()
+
+
 def _confidence(matches, ext=None) -> list[PureMagicWithConfidence]:
     """Rough confidence based on string length and file extension"""
     results = []
@@ -133,7 +136,7 @@ def _confidence(matches, ext=None) -> list[PureMagicWithConfidence]:
     if not results:
         raise PureError("Could not identify file")
 
-    return sorted(results, key=lambda x: (x.confidence, x.byte_match), reverse=True)
+    return sorted(results, key=lambda x: (x.confidence, len(x.byte_match)), reverse=True)
 
 
 def _identify_all(header: bytes, footer: bytes, ext=None) -> list[PureMagicWithConfidence]:
@@ -205,7 +208,6 @@ def _magic(header: bytes, footer: bytes, mime: bool, ext=None) -> str:
 
 def _file_details(filename: os.PathLike | str) -> tuple[bytes, bytes]:
     """Grab the start and end of the file"""
-    max_head, max_foot = _max_lengths()
     with open(filename, "rb") as fin:
         head = fin.read(max_head)
         try:
@@ -218,15 +220,17 @@ def _file_details(filename: os.PathLike | str) -> tuple[bytes, bytes]:
 
 def _string_details(string):
     """Grab the start and end of the string"""
-    max_head, max_foot = _max_lengths()
     return string[:max_head], string[-max_foot:]
 
 
 def _stream_details(stream):
     """Grab the start and end of the stream"""
-    max_head, max_foot = _max_lengths()
     head = stream.read(max_head)
-    stream.seek(-max_foot, os.SEEK_END)
+    try:
+        stream.seek(-max_foot, os.SEEK_END)
+    except OSError:
+        # File is smaller than the max_foot size, jump to beginning
+        stream.seek(0)
     foot = stream.read()
     stream.seek(0)
     return head, foot
@@ -374,6 +378,7 @@ def command_line_entry(*args):
         dest="mime",
         help="Return the mime type instead of file type",
     )
+    parser.add_argument("-v", "--v", action="store_true", dest="verbose", help="Print verbose output")
     parser.add_argument("files", nargs="+")
     args = parser.parse_args(args if args else sys.argv[1:])
 
@@ -385,6 +390,21 @@ def command_line_entry(*args):
             print(f"'{fn}' : {from_file(fn, args.mime)}")
         except PureError:
             print(f"'{fn}' : could not be Identified")
+            continue
+        if args.verbose:
+            matches = magic_file(fn)
+            print(f"Total Possible Matches: {len(matches)}")
+            for i, result in enumerate(matches):
+                if i == 0:
+                    print("\n\tBest Match")
+                else:
+                    print(f"\tAlertnative Match #{i}")
+                print(f"\tName: {result.name}")
+                print(f"\tConfidence: {int(result.confidence * 100)}%")
+                print(f"\tExtension: {result.extension}")
+                print(f"\tMime Type: {result.mime_type}")
+                print(f"\tByte Match: {result.byte_match}")
+                print(f"\tOffset: {result.offset}\n")
 
 
 imghdr_bug_for_bug = {  # Special cases where imghdr is probably incorrect.
@@ -444,5 +464,5 @@ def what(file: os.PathLike | str | None, h: bytes | None = None, imghdr_strict:
     return imghdr_exts.get(ext, ext)
 
 
-if __name__ == "__main__":
+if __name__ == "__main__":  # pragma: no cover
     command_line_entry()

test/test_common_extensions.py

@@ -9,6 +9,7 @@
 import puremagic
 
 LOCAL_DIR = os.path.realpath(os.path.dirname(__file__))
+RESOUCE_DIR = os.path.join(LOCAL_DIR, "resources")
 IMAGE_DIR = os.path.join(LOCAL_DIR, "resources", "images")
 VIDEO_DIR = os.path.join(LOCAL_DIR, "resources", "video")
 AUDIO_DIR = os.path.join(LOCAL_DIR, "resources", "audio")
@@ -19,6 +20,14 @@
 TGA_FILE = os.path.join(IMAGE_DIR, "test.tga")
 
 
+class MockBytesIO(BytesIO):
+
+    def seek(self, offset, whence=0):
+        if offset < 0:
+            raise OSError("Invalid seek position")
+        return super().seek(offset, whence)
+
+
 class TestMagic(unittest.TestCase):
     def setUp(self):
         self.mp4magic = b"\x00\x00\x00\x1c\x66\x74\x79\x70\x4d\x53\x4e\
@@ -127,6 +136,10 @@ def test_magic_stream(self):
         self.assertEqual(result[0].extension, ".tga")
         self.assertRaises(ValueError, puremagic.magic_stream, BytesIO(b""))
 
+    def test_small_stream_error(self):
+        ext = puremagic.from_stream(MockBytesIO(b"#!/usr/bin/env python"))
+        self.assertEqual(ext, ".py")
+
     def test_mime(self):
         """Identify mime type"""
         self.assertEqual(puremagic.from_file(TGA_FILE, True), "image/tga")
@@ -171,7 +184,9 @@ def test_cmd_options(self):
         """Test CLI options"""
         from puremagic.main import command_line_entry
 
-        command_line_entry(__file__, "test.py")
+        command_line_entry(__file__, os.path.join(AUDIO_DIR, "test.mp3"), "-v")
+        command_line_entry(__file__, "DOES NOT EXIST FILE")
+        command_line_entry(__file__, os.path.join(RESOUCE_DIR, "fake_file"), "-v")
 
     def test_bad_magic_input(self):
         """Test bad magic input"""