Fix missing 'rb' mode for opening files

[pdebuyl]

Add a "mode" argument to the calls to generic_open. This was the case before
using this context manager and was dropped by mistake in 78ef550

• Closes Creating scene with SEVIRI HRIT reader fails with UnicodeDecodeError #2073
• Tests added

This was not caught by tests as the tests are done with mocks instead of
data files.

[pdebuyl]

The failing CI error is

 E   ImportError: cannot import name 'Markup' from 'jinja2' (/usr/share/miniconda3/envs/test-environment/lib/python3.9/site-packages/jinja2/__init__.py)

at https://github.com/pytroll/satpy/runs/5766689500?check_suite_focus=true#step:7:219

It appears unrelated to the PR unfortunately.

[mraspaud]

@pdebuyl Thanks a lot for this! We are aware of the problem in the CI, so don't worry about these kinds of problems.

[mraspaud]

Thanks for the fix! do you think you can add a test for this, to avoid possible regressions in the future?

[pdebuyl]

I add a text test and a binary test, both plain and bzipped.

[mraspaud]

@pdebuyl a fix for the failing ci is now available in the main branch. Merging main into your branch should solve these issues.

[codecov]

Codecov Report

Merging #2077 (fc5f776) into main (48b2356) will increase coverage by 0.00%.
The diff coverage is 93.54%.

@@           Coverage Diff           @@
##             main    #2077   +/-   ##
=======================================
  Coverage   93.85%   93.85%           
=======================================
  Files         283      283           
  Lines       42279    42307   +28     
=======================================
+ Hits        39681    39709   +28     
  Misses       2598     2598           

Flag	Coverage Δ
behaviourtests	4.73% <0.00%> (-0.01%)	⬇️
unittests	94.42% <93.54%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
satpy/readers/seviri_l1b_hrit.py	90.14% <0.00%> (ø)
satpy/readers/hrit_base.py	81.87% <100.00%> (ø)
satpy/tests/reader_tests/test_utils.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 48b2356...fc5f776. Read the comment docs.

[coveralls]

Coverage increased (+0.004%) to 94.36% when pulling fc5f776 on pdebuyl:fix_generic_open_mode into 48b2356 on pytroll:main.

[pdebuyl]

Hi, I merged and force-pushed. The tests pass except for codebeat.

[sfinkens]

Thanks @pdebuyl for the fix! The relevant lines are still not covered, but currently I see no way to change this without adding even more mocks...

[sfinkens]

We really need to dynamically write a little test data file and use that for testing to get rid of all the open mocks... I'll put that on my PCW list.

[pdebuyl]

Do you think of having actual HRIT files for testing? Else, you would only test the tests... One solution would be to have a "data" repository that could be used optionally when running the tests.

[sfinkens]

I mean writing a small HRIT file during test setup. We already have the data type definitions for the header and image data, so maybe there's a way to write the header and a small image (like 10 scanlines or so) into a valid HRIT file.

[mraspaud]

I like the idea of creating an HRIT file on the fly for testing :)
@pdebuyl we decided a while ago against having data files in the satpy repo. We support a lot of formats, having test data files would make the repo grow too much.

[pdebuyl]

I meant another repository. But indeed, generating small HRIT segments is also an option.

Is there anything else regarding the current pull request?

[mraspaud]

I think we're good, maybe @gerritholl wants to give it a spin before we merge?

[mraspaud]

I ended up testing it myself, and it works fine. Merging

[pdebuyl]

Thanks, and sorry for putting a bug there!

[mraspaud]

No problem, thanks for fixing it!