Fix ANSI control sequences in Jupyter notebooks

[nschrader]

I suggest this fix for #176.

1. it checks if stdout is a TTY before printing private ANSI control sequences (DECTCEM 1&2)
2. I also corrected the EL sequence, which was missing the 0 for its canonical form (^[K is a special case of ^[0K)
3. I added a CR before the EL (which was already partly the case, but cluttered around the code).
4. some clean-up in the code calling the foregoing functions

Other control sequences seem to work fine in Jupyter. Tests:

• Jupyter notebook
  
• Terminal
  

[pavdmyt]

Hi @nschrader

Thanks for your PR!
I'd suggest to squash your commits and write a meaningful commit message. Also, seems like Travis CI integration has failed in some way, so I have to enable automatic PR testing again.

[pavdmyt]

Running tests manually (make test) on your branch, there are following failures:

    1236 passed
      13 failed
         - tests/test_in_out.py:92 test_hide_show['non-ascii str']
         - tests/test_in_out.py:78 test_write['non-ascii str']
         - tests/test_in_out.py:78 test_write['non-ascii unicode str']
         - tests/test_in_out.py:228 test_write_non_str_objects[foo-foo]
         - tests/test_in_out.py:228 test_write_non_str_objects[obj1-{'cat': 'meow'}]
         - tests/test_in_out.py:92 test_hide_show['non-ascii unicode str']
         - tests/test_in_out.py:228 test_write_non_str_objects[23-23]
         - tests/test_in_out.py:78 test_write['ascii str']
         - tests/test_in_out.py:228 test_write_non_str_objects[obj3-['foo', 'bar', "'", 23]]
         - tests/test_in_out.py:92 test_hide_show['empty']
         - tests/test_in_out.py:92 test_hide_show['ascii str']
         - tests/test_in_out.py:157 test_spinner_hiding_with_context_manager
         - tests/test_in_out.py:181 test_spinner_nested_hiding_with_context_manager
     714 skipped

You have to ensure that tests are working correctly and add a specific test cases for you fix.

[pavdmyt]

Please address the above comments.

[nschrader]

Thanks for your review, Pavel. Couldn't you simply "squash and merge" this PR on GitHub? So I don't need to force-push anything.

I didn't see the test suite, thanks for the hint! Should be all good now, just needed to adjust for new EL sequence.

[pavdmyt]

Looks good 👍
Just a minor item to clarify.

[pavdmyt]

I'm not sure I understand this part. Why this extra filling is needed?

[pavdmyt]

LGTM

[pavdmyt]

@nschrader once again, thanks for your contribution!

I'll include this into the next yaspin release.

[nschrader]

@pavdmyt sorry for the mess, but I was testing again this morning and noticed some more problems which Jupyter. Turns out that ^[K isn't actually clearing the line, rather just printing 3 white spaces. That's why ^[0K "fixed" some issues (well, just overrode one more character). I added some more commits, reverting the EL sequence and overriding the current line with white spaces, the only thing that reliably works in Jupyter.

I'd suggest that you revert this PR and I open a new one, so there is a clean trace.