IntelOwl
IntelOwl copied to clipboard
intelowlproject
[Analyzer] Zippy
Name
Zippy
Link
https://github.com/thinkst/zippy
Type of analyzer
file analyzer
Why should we use it
Detect AI-generated text [relatively] quickly via compression ratios
Possible implementation
tool written in python, we could integrate it as a library
hello @mlodic , Since the zippy repo provides us a script that is used as a cli command and is not a pre-built library, should we directly implement the zippy script, wrap it up with desired functionalities, into a single analyzer file and use it as a new analyzer?
Also I'd like to work on this issue pls :)
I think you could install it in the same way that they suggest to install it here and then call the functionality expressed here in the analyzer file (leverage the Zippy class).
Even if the best thing would be to help them to create a packet in pypi and leverage the pypi installation then :P
okay so i did some things here and there:
- installed zippy via cloning and executing the commands in the guide.
- I first tried using zippy by importing it with some random values, and it worked just fine
- I took some references from the other implemented file analyzers and the contribution doc and coded an analyzer with that info.
- Now that I'm trying to integrate it using the admin panel, I have to add a new py module, it throws an error. This probably has something to do with the import zippy I'm doing. Can't exactly find a lot of information about how to proceed in such cases.
- I did take some refernce from other analyzers but couldnt exactly figure out a good suitable way for implementing relative imports in Analyzers.
- This is the code I wrote to implement the analyzer:
# This file is a part of IntelOwl https://github.com/intelowlproject/IntelOwl
# See the file 'LICENSE' for copying permission.
from zippy import EnsembledZippy
from api_app.analyzers_manager.classes import FileAnalyzer
from api_app.analyzers_manager.exceptions import (
AnalyzerRunException,
)
from api_app.analyzers_manager.observable_analyzers.yaraify import YARAify
from tests.mock_utils import patch
class ZippyAnalyser(FileAnalyzer):
"""
Tells if a file is written by HUMAN or AI
"""
def run(self):
binary_data=self.read_file_bytes()
text_data = binary_data.decode('utf-8')
try:
response=(EnsembledZippy().run_on_text_chunked(text_data))
except Exception as e:
print("Zippy not imported correctly")
raise AnalyzerRunException
print(response)
return response.json()
@classmethod
def _monkeypatch_zippy(cls):
patches = [
patch(
'zippy.EnsembledZippy.run_on_text_chunked',
return_value={'mocked_response': 'AI 0.62739394'}
),
]
return super()._monkeypatch(patches=patches)
Can you guide me to some reference as to why and how can I solve the error(4) in question.
hey, thanks for your efforts :) please open a PR and set it as "draft". In this way, we can help you to solve such kind of problems better, without having to copy/pasting things here which complicates things further. ty