Add JMA HRIT readers for MTSAT-1R/2

[sfinkens]

• Add method to compute geostationary earth mask

• Add JMA HRIT readers for MTSAT-1R/2

  • Add YAML definitions for MTSAT-1R/2
  • Update YAML definition for Himawari-8
    • Rename hrit_jma.yaml -> hrit_ahi.yaml
    • Remove radiance datasets because it is not supported
    • Add counts datasets
  • Modify existing JMA HRIT reader:
    • Add support for half disk scans
    • Extract platform name from the file (needs to be improved for
      Himawari-9)
    • Mask space pixels

• Tests added

• Tests passed

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

• Fully documented

• Add example notebook

[coveralls]

Coverage increased (+0.2%) to 74.813% when pulling 19a304c on sfinkens:feature-mtsat into ed9bd67 on pytroll:master.

[codecov]

Codecov Report

Merging #524 into master will increase coverage by 0.19%.
The diff coverage is 99.37%.

@@            Coverage Diff             @@
##           master     #524      +/-   ##
==========================================
+ Coverage   74.62%   74.81%   +0.19%     
==========================================
  Files         136      136              
  Lines       18342    18482     +140     
==========================================
+ Hits        13687    13827     +140     
  Misses       4655     4655

Impacted Files	Coverage Δ
satpy/tests/reader_tests/test_hrit_jma.py	96.58% <100%> (+6.1%)	⬆️
satpy/readers/utils.py	71.11% <100%> (+2.43%)	⬆️
satpy/tests/reader_tests/test_utils.py	75.89% <100%> (+4.01%)	⬆️
satpy/readers/hrit_jma.py	98.43% <97.95%> (+0.76%)	⬆️

---

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 ed9bd67...19a304c. Read the comment docs.

[sfinkens]

@djhoese Ready to merge as far as I'm concerned. Do you know what's wrong with the AppVeyor build?

[sfinkens]

Ah no, we wanted to wait for JMA's feedback regarding Himawari-9

[mraspaud]

LGTM

[djhoese]

@sfinkens Yes, I do. I actually spent my weekend trying to figure out what to do with the appveyor tests. You can see the results in this issue and the ones linked to it: conda-forge/gdal-feedstock#241

Bottom line, we have to wait for a new build of gdal and possibly the geotiff library before things will start working. Hopefully if some of the core conda-forge folks have time today we'll see new builds today or tomorrow (USA/Brazil time zones).

Also keep in mind based on #527 that these reader names will likely change in before the SatPy 0.11 release. If this is a problem let's discuss it on slack and find a reasonable solution.

[mraspaud]

so it will become ahi_hrit and mtsat_imager_hrit, right ?

[djhoese]

I think mtsat2_imager_hrit and jami_hrit by the current sensor configuration I see in those files. See my most recent complications on the naming issue. Not sure how I feel about sensors that require separates or what to do when the sensor doesn't have a unique name like "imager" (old GOES also have "imager").

[sfinkens]

Since underscores separate the fields I would prefer mtsat-imager_hrit

[sfinkens]

Regarding the TODOs in the code: I will check that as soon as I have feedback from JMA and have the whole dataset.