Scrubbing

[KrishPatel13]

Presidio is not capable for Scrubbing Credit-Card numbers and Social Security Numbers (ssn), which is strange (it should do it). See Test file and the attached test session text file for reference.
Besides that, it is working for the below fields:

• Email
• Phone Number
• Date of Birth
• Address
• Driving Licence
• Passport
• National ID
• Routing Number
• Bank Account Number

[KrishPatel13]

My Test Session Transcript:

pytest .\test_scrub.py 
====================================================== test session starts =======================================================
platform win32 -- Python 3.9.7, pytest-7.1.3, pluggy-1.0.0
rootdir: C:\Users\Krish Patel\OneDrive - University of Toronto\Desktop\MLDSAI Inc\PAT
plugins: anyio-3.6.2, hypothesis-6.62.0, typeguard-2.13.3
collected 14 items
​
test_scrub.py ....F..F.....F                                                                                                [100%]
​
============================================================ FAILURES ============================================================ 
_____________________________________________________ test_scrub_credit_card _____________________________________________________ 
​
    def test_scrub_credit_card():
        # Test scrubbing of credit card number
>       assert scrub("My credit card number is 1234-5678-9012-3456 and ") == "My credit card number is ****-****-****-**** and "   
E       AssertionError: assert 'My credit ca...012-3456 and ' == 'My credit ca...***-**** and '
E         - My credit card number is ****-****-****-**** and 
E         + My credit card number is 1234-5678-9012-3456 and
​
test_scrub.py:24: AssertionError
------------------------------------------------------- Captured log call -------------------------------------------------------- 
WARNING  presidio-analyzer:nlp_engine_provider.py:109 configuration file C:\Users\Krish Patel\AppData\Roaming\Python\Python39\site-packages\conf\default.yaml not found.  Using default config: {'nlp_engine_name': 'spacy', 'models': [{'lang_code': 'en', 'model_name': 'en_core_web_lg'}]}.
_________________________________________________________ test_scrub_ssn _________________________________________________________ 
​
    def test_scrub_ssn():
        # Test scrubbing of social security number
>       assert scrub("My social security number is 123-45-6789") == "My social security number is ***********"
E       AssertionError: assert 'My social se...s 123-45-6789' == 'My social se...s ***********'
E         - My social security number is ***********
E         + My social security number is 123-45-6789
​
test_scrub.py:36: AssertionError
------------------------------------------------------- Captured log call -------------------------------------------------------- 
WARNING  presidio-analyzer:nlp_engine_provider.py:109 configuration file C:\Users\Krish Patel\AppData\Roaming\Python\Python39\site-packages\conf\default.yaml not found.  Using default config: {'nlp_engine_name': 'spacy', 'models': [{'lang_code': 'en', 'model_name': 'en_core_web_lg'}]}.
____________________________________________________ test_scrub_all_together _____________________________________________________ 
​
    def test_scrub_all_together():
        # Text with all PII/PHI types
        text_with_pii_phi = "John Smith's email is johnsmith@example.com and his phone number is 555-123-4567. His credit card number is 1234-5678-9012-3456 and his social security number is 123-45-6789. He was born on 01/01/1980."
>       assert scrub(text_with_pii_phi) == "****'s email is ***@***.*** and his phone number is ***-***-****. His credit card number is ****-****-****-**** and his social security number is ***-**-****. He was born on **/**/****."
E       assert '************...n **********.' == "****'s email...n **/**/****."
E         - ****'s email is ***@***.*** and his phone number is ***-***-****. His credit card number is ****-****-****-**** and his social security number is ***-**-****. He was born on **/**/****.
E         + ************ email is ********************* and his phone number is ************. His credit card number is 1234-5678-9012-3456 and his social security number is 123-45-6789. He was born on **********.
​
test_scrub.py:61: AssertionError
------------------------------------------------------- Captured log call -------------------------------------------------------- 
WARNING  presidio-analyzer:nlp_engine_provider.py:109 configuration file C:\Users\Krish Patel\AppData\Roaming\Python\Python39\site-packages\conf\default.yaml not found.  Using default config: {'nlp_engine_name': 'spacy', 'models': [{'lang_code': 'en', 'model_name': 'en_core_web_lg'}]}.
==================================================== short test summary info ===================================================== 
FAILED test_scrub.py::test_scrub_credit_card - AssertionError: assert 'My credit ca...012-3456 and ' == 'My credit ca...***-****...
FAILED test_scrub.py::test_scrub_ssn - AssertionError: assert 'My social se...s 123-45-6789' == 'My social se...s ***********'     
FAILED test_scrub.py::test_scrub_all_together - assert '************...n **********.' == "****'s email...n **/**/****."
================================================= 3 failed, 11 passed in 19.99s ==================================================

[abrichr]

Excellent work @KrishPatel13 !

It looks like the scrubbing is working, and the test needs to be modified:

assert '************...n **********.' == "****'s email...n **/**/****."

[KrishPatel13]

My next step will be Implementing scrubbing for images

[KrishPatel13]

Reviewed and Changed!

[KrishPatel13]

@abrichr Re-review is requested!

Thank you for your patience!

[KrishPatel13]

Reviewed

[abrichr]

Thank you @KrishPatel13 ! Can you please merge the latest from main?

[abrichr]

Please remove unrelated changes 🙏

[KrishPatel13]

microsoft/presidio#1079

I was not able to record so I finally had to investigate deep and found that after adding simplest code of Presidio of initializing the engines it gave the errors. So I had to raise the issue on their GitHubpage.

[KrishPatel13]

microsoft/presidio#1079

I was not able to record so I finally had to investigate deep and found that after adding simplest code of Presidio of initializing the engines it gave the errors. So I had to raise the issue on their GitHub page.

Resolved:
Issue Description:
If I pressed Ctrl + C after 3 "starting". It gave errors. I even did some typing and then pressed Ctrl + C, got the error.

BUT,

when I waited for at least 10-15 sec idle after the 3 starting and then press ctrl + C, it recorded successfully.

[KrishPatel13]

@abrichr We can close this WITHOUT merging once the below is merged! :)

#211
Ready for review! :)

[abrichr]

Looks great @KrishPatel13 ! Can you please look into modifying configure_logging to scrub data before it is logged? If it's a lot of work we can leave it as a TODO for now.

[abrichr]

Closing in favor of #211