Update VI COG Metadata attributes

[madhuksridhar]

Required metadata attributes (compiled here) are shown below:

long_name
AREA_OR_POINT  (needed by GeoTIFF?)
HORIZONTAL_CS_NAME
MGRS_TILE_ID
ULX
ULY
NCOLS
NROWS
SPATIAL_RESOLUTION
SENSING_TIME
SATELLITE  
ACCODE
cloud_coverage
spatial_coverage
MEAN_SUN_AZIMUTH_ANGLE
MEAN_SUN_ZENITH_ANGLE
MEAN_VIEW_AZIMUTH_ANGLE
MEAN_VIEW_ZENITH_ANGLE
NBAR_SOLAR_ZENITH
add_offset
scale_factor
HLS-VI_PROCESSING_TIME
_FillValue

All fields except HLS-VI_PROCESSING_TIME, scale_factor, long_name, SATELLITE can be copied from HLS base granule attributes

• SATELLITE will be one of the following : L8, L9, S2A, S2B. The value can be extracted from HLS base cmr xml file. For e.g. PRODUCT_URI for S30 and LANDSAT_PRODUCT_ID for L30
• scale_factor is 0.0001 for all VI indices except TVI. scale_factor for TVI is 1.0

[chuckwondo]

@madhuksridhar:

• There is no MGRS_TILE_ID in the band metadata for either L30 or S30. However, oddly, L30 has a tag SENTINEL2_TILEID (e.g., 06WVS), and S30 has TILE_ID (e.g., TILE_ID=S2B_OPER_MSI_L1C_TL_2BPS_20240507T205106_A037447_T13RCN_N05.10). What does a MGRS_TILE_ID look like and how do I obtain an appropriate value from L30 and S30 metadata?
• You show long_name (with an underscore), but we already have longname (without an underscore). Which one is correct?
• When we read the bands, we divide the values by 10,000, and then after computing each index, we multiply by 10,000, except for TVI. In one of the TVI files in the test fixtures, the range of values is [-15, 18]. Does that make sense?

[madhuksridhar]

• If you split a HLS granule id ( HLS.L30.T06WVS.2024120T211159.v2.0) with period separator, MGRS_TILE_ID is the third entry (e.g. T06WVS)
• longname should be changed to long_name

@bwbaker1 or @zhouqiang06 can you confirm about TVI range?

[chuckwondo]

@madhuksridhar, I again accidentally pushed to main (grrr!), but I think I found the github setting to prevent this, so hopefully I will be blocked from doing this again.

Please see this commit for the changes I made for this issue: d86f978

[madhuksridhar]

looks good.