FY-4A REGC data got wrong area_def

[sgxl]

Hi, thanks for maintaining satpy.

Describe the bug
When I read the Level-1 REGC（not full-disk data） files of FY-4A AGRI sensor using the satpy package, wrong area_def was gotten.

To Reproduce

print(f'The version of satpy: {satpy.__version__}')

print(f'Input filenames:')
for file in file_list_fy4a:
    print(f'-- {os.path.basename(file)}')

# Reading datas with satpy
reader_fy4a = satpy.Scene(filenames=file_list_fy4a, reader='agri_fy4a_l1')
reader_fy4a.load(['C12', 'C13'])
# get the area_def of the scene
area_fy4a = reader_fy4a.coarsest_area()
lons_fy4a, lats_fy4a = area_fy4a.get_lonlats()

# plot the distribution of Latitudes of each pixel of the scene
plt.imshow(lats_fy4a, cmap='rainbow')
plt.colorbar(shrink=0.5, label='Latitude')
plt.tight_layout()
plt.show()

Expected behavior
The REGC file is a small observation area of FY-4A scan mechanism, the area_def of the scene should cover China region （in the Northern Hemisphere）.

But what I got is a wrong area_def in the southern hemisphere, I think there may be some code wrong while reading the REGC file of FY-4A, the area should be a subset the geostationary projection, and the indices for the subset is set in the attributes of HDF file of FY-4A. (sorry that I haven't check the source code yet because I am not familiar with the code structure for satpy now)

Actual results

Screenshots
Obviously, the figure beblow shows wrong latitude value distribution of each pixel. (detatils are in the "Expected behavior" section above, thanks)

I am not certained that whether it is a bug or my coding fault.

[djhoese]

Here is where the area definition is created for this reader and the Begin/End Line Number attributes are read:

satpy/satpy/readers/fy4_base.py

Lines 216 to 217 in d1f98a2

	begin_cols = float(self.file_content["/attr/Begin Pixel Number"])
	end_lines = float(self.file_content["/attr/End Line Number"])

However, you say you're using Satpy 0.42.2 which means you are missing a lot of changes to Satpy. This includes this:

#2533

which may include fixes for some of the stuff you're seeing. Could you try updating and see if it fixes things for you?

[sgxl]

Thank you for your response. I will update my satpy package and try again. I apologize for not checking the posted issues carefully.

Here is where the area definition is created for this reader and the Begin/End Line Number attributes are read:

satpy/satpy/readers/fy4_base.py

Lines 216 to 217 in d1f98a2

	begin_cols = float(self.file_content["/attr/Begin Pixel Number"])
	end_lines = float(self.file_content["/attr/End Line Number"])

However, you say you're using Satpy 0.42.2 which means you are missing a lot of changes to Satpy. This includes this:

#2533

which may include fixes for some of the stuff you're seeing. Could you try updating and see if it fixes things for you?