plot_network: color-code network/matrix plot based on temp/perp baselines by yuankailiu · Pull Request #693 · insarlab/MintPy

yuankailiu · 2021-11-10T00:35:07Z

Description of proposed changes

Here are two proposed changes:

1). The displacement map in tsview.py needs to be projected onto a specific Cartopy projection (e.g., ccrs.PlateCarree()) to properly overlay the coastline. This is added. But I am not sure whether there should be tunable projections. For small regions (for most of the cases), I think it is fine.

2). In 'plot_network.py', Other than coherence, I personally think it would be good to also allow the lines (connecting acquisitions) to be colored based on their temporal or spatial baselines. This is just for network design displaying purposes. You can reject that as well.

Reminders

Pass Codacy code review (green)
Pass Circle CI test (green)
Make sure that your code follows our style. Use the other functions/files as a basis.
If modifying functionality, describe changes to function behavior and arguments in a comment below the function declaration.

yunjunz

Thank you @yuankailiu, the two features sound fine to me.

Here are my suggestions:

For the plot_network.py, I tried to run plot_network.py inputs/ifgramStack.h5 --tb and it does seem to work. Could you double-check, and add an example usage as well?
For the tsview.py, it's more complicated, as the script needs to support files in radar coordinates as well. Below is how view.py handles this, which eventually translated into the inps.map_proj_obj variable. I would recommend: packing this checking into a sub-function in view.py or utils/plot.py, and call it from tsview.py. If you want to go for this, please consider submitting it in another PR, as it requires much more changes.

MintPy/mintpy/view.py

Lines 310 to 345 in 92e00ad

    
           # Map Projection via cartopy 
        
           # cartopy requires that: 
        
           #   1. file is geocoded AND 
        
           #   2. file coordinates are in the unit of degrees / meters AND 
        
           #   3. set to display in geo-coordinates 
        
           # use cartopy (by initiating inps.map_proj_obj) ONLY IF: 
        
           #   1. show fancy lat/lon label via --lalo-label OR 
        
           #   2. show coastline via --coastline 
        
           inps.map_proj_obj = None 
        
           inps.coord_unit = metadata.get('Y_UNIT', 'degrees').lower() 
        
           if (inps.geo_box 
        
                   and inps.coord_unit.startswith(('deg', 'meter')) 
        
                   and inps.fig_coord == 'geo' 
        
                   and (inps.lalo_label or inps.coastline)): 
        
               # get projection name from the data coord unit 
        
               # https://scitools.org.uk/cartopy/docs/latest/crs/projections.html 
        
               msg = 'initiate cartopy map projection: ' 
        
               if inps.coord_unit.startswith('deg'): 
        
                   inps.map_proj_obj = ccrs.PlateCarree() 
        
                   vprint(msg + 'PlateCarree') 
        
               elif inps.coord_unit.startswith('meter'): 
        
                   if 'UTM_ZONE' in metadata.keys(): 
        
                       utm_zone = metadata['UTM_ZONE'] 
        
                       inps.map_proj_obj = ccrs.UTM(utm_zone) 
        
                       vprint(msg + f'UTM zone {utm_zone}') 
        
                       # check --lalo-label (works for PlateCarree only) 
        
                       if inps.lalo_label: 
        
                           raise ValueError('--lalo-label is NOT supported for projection: UTM') 
        
                   else: 
        
                       print('WARNING: Un-recognized coordinate unit: {}'.format(inps.coord_unit)) 
        
                       print('    Switch to the native Y/X and continue to plot') 
        
                       inps.fig_coord = 'radar'

yuankailiu · 2022-01-06T23:45:08Z

Hi Yunjun, thanks for reviewing this.

plot_network.py and plot.py have been updated, with an example added in plot_network.py -h. Now it is working after my testing.

I have also changed the utils1.py a bit in the get_residual_std() function. I make it to return the std_file as the get_residual_rms() function.

As for tsview.py, I will do it later in another PR. Thanks!

mintpy/utils/utils1.py

mintpy/timeseries_rms.py

mintpy/plot_network.py

yunjunz

Thank you @yuankailiu for the updates. Could you check the few minor suggestions I listed above?

After that, the code is clean enough and I could work on the idea below (unless you are willing to work that, in that case let me know), which requires quite some refactoring:

Your implementation makes me think that we could use a more generic/flexible --dset dsName option: if dsName is a 3D matrix, then use maskFIle to calculate the spatial average (which is what currently do); if dsName is a 1D matrix, e.g. bperp, treat it as the calculated spatial average directly, and all the following code should be the same.

+ Can color the network lines with tbase (days) using --tbase option + Can color the network lines with pbase (days) using --pbase option (use plot_network.py -h to see the explanations and an example) utils1: + get_residual_std(): also return `std_file` variable

yunjunz

Looks all good to me.

yunjunz self-requested a review November 11, 2021 03:43

yunjunz requested changes Nov 11, 2021

View reviewed changes

yuankailiu force-pushed the dev branch 3 times, most recently from a17b51c to d0cd424 Compare January 6, 2022 23:37

yuankailiu changed the title ~~Map projection in tsview.py + Network plot line coloring based on baselines~~ plot_network: Network plot lines coloring based on baselines Jan 6, 2022

yuankailiu force-pushed the dev branch from d0cd424 to 7e78abd Compare January 6, 2022 23:40

yunjunz reviewed Jan 7, 2022

View reviewed changes

mintpy/utils/utils1.py Show resolved Hide resolved

yunjunz reviewed Jan 7, 2022

View reviewed changes

mintpy/timeseries_rms.py Outdated Show resolved Hide resolved

yunjunz reviewed Jan 12, 2022

View reviewed changes

mintpy/plot_network.py Outdated Show resolved Hide resolved

yunjunz requested changes Jan 12, 2022

View reviewed changes

yuankailiu added 2 commits January 20, 2022 14:44

few minor suggestions from the reviewer in PR#693

1ff802c

yuankailiu force-pushed the dev branch from 7e78abd to 1ff802c Compare January 20, 2022 23:05

yuankailiu requested a review from yunjunz January 20, 2022 23:08

yunjunz added 3 commits January 20, 2022 16:30

Update utils1.py

f9e06b6

merge --tb/tbv/pb/pbv into -d/-v

7cc638e

codacy

27b2043

yunjunz changed the title ~~plot_network: Network plot lines coloring based on baselines~~ plot_network: color-code network/matrix plot based on temp/perp baselines Jan 21, 2022

yunjunz approved these changes Jan 21, 2022

View reviewed changes

yunjunz merged commit 1727c7d into insarlab:main Jan 21, 2022

yuankailiu deleted the dev branch January 21, 2022 03:47

yunjunz mentioned this pull request Jan 24, 2022

bugfix in load_data.py & plot_network.py #729

Merged

2 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

plot_network: color-code network/matrix plot based on temp/perp baselines#693

plot_network: color-code network/matrix plot based on temp/perp baselines#693
yunjunz merged 5 commits intoinsarlab:mainfrom
yuankailiu:dev

yuankailiu commented Nov 10, 2021 •

edited by yunjunz

Loading

Uh oh!

yunjunz left a comment

Uh oh!

yuankailiu commented Jan 6, 2022

Uh oh!

Uh oh!

Uh oh!

Uh oh!

yunjunz left a comment

Uh oh!

yunjunz left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	# Map Projection via cartopy
	# cartopy requires that:
	# 1. file is geocoded AND
	# 2. file coordinates are in the unit of degrees / meters AND
	# 3. set to display in geo-coordinates
	# use cartopy (by initiating inps.map_proj_obj) ONLY IF:
	# 1. show fancy lat/lon label via --lalo-label OR
	# 2. show coastline via --coastline
	inps.map_proj_obj = None
	inps.coord_unit = metadata.get('Y_UNIT', 'degrees').lower()
	if (inps.geo_box
	and inps.coord_unit.startswith(('deg', 'meter'))
	and inps.fig_coord == 'geo'
	and (inps.lalo_label or inps.coastline)):

	# get projection name from the data coord unit
	# https://scitools.org.uk/cartopy/docs/latest/crs/projections.html
	msg = 'initiate cartopy map projection: '
	if inps.coord_unit.startswith('deg'):
	inps.map_proj_obj = ccrs.PlateCarree()
	vprint(msg + 'PlateCarree')

	elif inps.coord_unit.startswith('meter'):
	if 'UTM_ZONE' in metadata.keys():
	utm_zone = metadata['UTM_ZONE']
	inps.map_proj_obj = ccrs.UTM(utm_zone)
	vprint(msg + f'UTM zone {utm_zone}')

	# check --lalo-label (works for PlateCarree only)
	if inps.lalo_label:
	raise ValueError('--lalo-label is NOT supported for projection: UTM')

	else:
	print('WARNING: Un-recognized coordinate unit: {}'.format(inps.coord_unit))
	print(' Switch to the native Y/X and continue to plot')
	inps.fig_coord = 'radar'

Conversation

yuankailiu commented Nov 10, 2021 • edited by yunjunz Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

yunjunz left a comment

Choose a reason for hiding this comment

Uh oh!

yuankailiu commented Jan 6, 2022

Uh oh!

Uh oh!

Uh oh!

Uh oh!

yunjunz left a comment

Choose a reason for hiding this comment

Uh oh!

yunjunz left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

yuankailiu commented Nov 10, 2021 •

edited by yunjunz

Loading