Add basic unit tests #215
Add basic unit tests #215
Conversation
…es Add tests to validate the output of the method `image_to_data`
nok
force-pushed
the
feature/expand_testing
branch
from
d5d61a7
to
6cbbea3
Compare
|
Evidently, Tesseract v3.04.01 with Leptonica has trouble with gif images.
|
|
@johnthagen please review whenever you have some spare time and let me know if everything looks good for the tests. |
…sing `isort` Make method `predict` private by adding a leading underscore `_predict`
…t__` to `pytesseract.pytesseract`
tests/test_pytesseract.py
Outdated
|
|
||
|
|
||
| @pytest.mark.parametrize('test_file', [ | ||
| # https://github.com/tesseract-ocr/tesseract/issues/2558 |
There was a problem hiding this comment.
Can these tests be turned on for tesseract 4.0 / bionic?
There was a problem hiding this comment.
It also depends on the behaviour and version of Leptonica. On my system I have the newest versions of Tesseract and Leptonica and it doesn't work. Nevertheless I will test it with the images of Travis.
JFYI, we can skip tests with this decorator:
@pytest.mark.skipif(
TESSERACT_VERSION[0] < 4,
reason='requires tesseract >= 4'
)
# ...
tests/test_pytesseract.py
Outdated
| # os.path.join(DATA_DIR, 'test.bmp'), | ||
| # os.path.join(DATA_DIR, 'test.gif'), | ||
| os.path.join(DATA_DIR, 'test.jpg'), | ||
| Image.open(os.path.join(DATA_DIR, 'test.jpg')), |
There was a problem hiding this comment.
Could probably simplify the logic if the Image.open() and os.path.join was moved into the test itself, and the only things parameterized was the file itself?
There was a problem hiding this comment.
I will split this tests in two different tests. You're right, I mixed it.
There was a problem hiding this comment.
gif images don't work with Tesseract 3, so I skip these cases.
bmp images don't work with any version of Tesseract.
There was a problem hiding this comment.
Hmm, interesting. According to the Leptonica docs - Image I/O, there is support for BMP and GIF. Leptonica is the underlying library that tesseract utilize.
There was a problem hiding this comment.
Also, as far as I can see from the Travis CI report, the gif test passes on bionic.
So indeed, the problem is the version of Leptopnica (leptonica-1.73 - xenial vs leptonica-1.75.3 - bionic which is also a bit old - Feb 16, 2018).
…ting # Please enter a commit message to explain why this merge is necessary, # especially if it merges an updated upstream into a topic branch. # # Lines starting with '#' will be ignored, and an empty message aborts # the commit.
…esseract.tesseract_cmd` at the end of the test
…t and ignore bitmaps generally
|
Thanks for your review and feedback @int3l and @johnthagen. In general Tesseract 3/4 can't handle bitmap images. But Tesseract 4 can handle gif images. The tests handle these circumstances. |
|
Once again, @nok and @johnthagen thank you for the contributions and the time spent on the tasks. |
@int3l It makes fun and I learned something new 😄. Yes, please add me to the list of contributors, thanks! |
Hello,
I added some unit tests to cover all common methods. For that I added more test data.
Finally I wasn't able to use a bitmap file for the tests, because tesseract (leptonica) failed.
But this bug is related to an open issue: tesseract-ocr/tesseract#2558.
Nevertheless the other tests passed successfully (Python 2.7, Python 3.6 and Python 3.7).
Screenshot: