Change unnecessary warning messages to debug

[djhoese]

There were a lot of warnings that always show up when doing certain types of processing. It was important when features were first developed, but as things have stabilized these types of messages are more for satpy developers rather than for users. If a user is curious why their product wasn't generated they will still enable debug logging to get the most information as possible.

Additionally there are a couple small changes in this PR:

• Remove conda-forge workaround in travis for conda-forge/defaults incompatibility. Uses channel priority instead

• Remove unnecessary 'abi' sensor in abi_l1b.yaml. The sensor is already added in the code.

• Other small code style changes

• Tests added

• Tests passed

• Passes git diff origin/master -- "*py" | flake8 --diff

[coveralls]

Coverage decreased (-0.02%) to 74.236% when pulling ae30c4f on djhoese:bugfix-abi-warnings into 6fc2669 on pytroll:master.

[codecov]

Codecov Report

Merging #504 into master will increase coverage by <.01%.
The diff coverage is 87.5%.

@@            Coverage Diff             @@
##           master     #504      +/-   ##
==========================================
+ Coverage   74.25%   74.25%   +<.01%     
==========================================
  Files         137      137              
  Lines       18218    18219       +1     
==========================================
+ Hits        13528    13529       +1     
  Misses       4690     4690

Impacted Files	Coverage Δ
satpy/tests/test_enhancements.py	98.55% <ø> (ø)	⬆️
satpy/readers/abi_l1b.py	96.66% <100%> (+0.02%)	⬆️
satpy/scene.py	84.03% <85.71%> (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 6fc2669...ae30c4f. Read the comment docs.

[mraspaud]

I'm ok with PRs fixing multiple small things, but please describe them in the description. The description on this PR makes it sound like there is only one change.

[djhoese]

I completely agree. I honestly don't remember clicking submit on this. I wasn't done with it. I also didn't plan for this to include so many small things but I ran git status and ended up having a lot of little things hanging around in my local repository.

[djhoese]

With the changes currently in this PR (masking out negative reflectances):

Previous results by clipping to 0 and 1 in crefl enhancement per band (note the bright cyan on the terminator):

Changing it to be as described and setting any pixels with negative reflectances to black (looks like there are still blue pixels, just less of them):

Final conclusion: After looking closely at all images the bright cyan pixels are present in all of them. They are just harder to see in some of them or are in different quantities. Note this is all with the crefl enhancement but there are bright cyan/blue pixels in the cira stretch to.