`KeyError` on `dataset convert dataset-name`

[AlexVanMechelen]

Description

KeyError: 'STANDARD_SECTION_NAMES' appears when running dataset convert dataset-name on the latest commit 0eb04cd.

Error debug output

┌──[user@packing-box]──[/mnt/share]────────                                                                                                                         ────[10.0.2.100]──[11:55:15]────
$ dataset convert test-pe-upx
00:00:00.452 [INFO] Converting to fileless dataset...
00:00:00.454 [INFO] Size of dataset:     49MB
00:00:00.455 [INFO] Loading features...
  0% ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━   0/100 samples • 0:00:02 • -:--:--
Traceback (most recent call last):
  File "/home/user/.opt/tools/dataset", line 239, in <module>
    getattr(ds, args.command)(**vars(args))
  File "/home/user/.local/lib/python3.11/site-packages/pbox/helpers/items.py", line 308, in _wrapper
    return f(s, *a, **kw)
           ^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/dataset/__init__.py", line 368, in convert
    self._compute_all_features()
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/dataset/__init__.py", line 193, in _compute_all_features
    d = self._compute_features(exe)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/dataset/__init__.py", line 201, in _compute_features
    d.update(exe.data)        # be sure to include the features
             ^^^^^^^^
  File "/usr/lib/python3.11/functools.py", line 1001, in __get__
    val = self.func(instance)
          ^^^^^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/__init__.py", line 131, in data
    return Features(self)
           ^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/tinyscript/preimports/log.py", line 85, in _wrapper
    return f(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/features.py", line 89, in __init__
    self._rawdata = Extractors(exe)
                    ^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/extractors/__init__.py", line 60, in __init__
    self[name] = func(exe)
                 ^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/extractors/__common__.py", line 38, in _wrapper
    return f(parse(str(target)), *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/extractors/__common__.py", line 59, in <lambda>
    standard_sections         = parse_binary(lambda exe: [s.name for s in exe.sections if s.name in \
                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/user/.local/lib/python3.11/site-packages/pbox/core/executable/extractors/__common__.py", line 60, in <listcomp>
    get_data(exe.format.name)['STANDARD_SECTION_NAMES']])
    ~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^
KeyError: 'STANDARD_SECTION_NAMES'

Steps to reproduce

$ git clone git@github.com:packing-box/docker-packing-box.git
$ cd docker-packing-box/
$ docker build -t dhondta/packing-box .
$ docker run -it -h packing-box -v `pwd`:/mnt/share dhondta/packing-box
$ dataset make test-pe-upx -n 100 --format PE --packer upx
$ dataset convert test-pe-upx

[dhondta]

Hi @AlexVanMechelen ! You can retry your last command in verbose mode (dataset -v convert test-pe-upx)and check the latest lines of debug information before the traceback of the error. I expect you to see the path to the target executable for which information extraction fails. You will have a line of this type : pefeats '/home/user/.packing-box/datasets/test-pe-upx/files/[sha256] You can then execute the following command to see the file type and post it on this issue ; file /home/user/.packing-box/datasets/test-pe-upx/files/[sha256]

[AlexVanMechelen]

file /home/user/.packing-box/datasets/test-pe-upx/files/[sha256] output:

PE32+ executable (DLL) (console) x86-64, for MS Windows, 19 sections

[dhondta]

@AlexVanMechelen Please send me the failing sample via email.