Bowser

Quickstart

1. Install uv
2. Run uvx --with bowser-insar --stack-file example-stack.zarr

Install

pip install bowser-insar`

uv add bowser-insar

For local development, clone and create a conda environment:

git clone [email protected]:opera-adt/bowser.git && cd bowser
mamba env create
conda activate bowser-env
pip install .

Quickstart for dolphin

• bowser setup-dolphin is preconfigured to make a JSON file pointing to the outputs inside work_directory of dolphin
• bowser run starts the web server on localhost:

$ bowser setup-dolphin work/
Reading raster metadata: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:00<00:00, 236.13it/s]
Reading raster metadata: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:00<00:00, 95.74it/s]
Reading raster metadata: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:00<00:00, 266.77it/s]
Reading raster metadata: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 100.49it/s]
Reading raster metadata: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 94.88it/s]

$ bowser run
INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
INFO:     Started parent process [99855]
...

Click on the http://127.0.0.1:8000 link to open the map.

Note for running over ssh: you will need to run an ssh command creating a tunnel from your local computer to wherever the bowser server is. For example, if you machine is myserver, you would run in a local terminal

ssh -N -L 8000:localhost:8000 myserver

after starting the web server.

Running in Jupyter

from bowser._widget import make_bowser_widget

# Automatically starts server and creates widget
widget = make_bowser_widget(rasters_file="bowser_rasters.json")
widget

Quickstart for DISP-S1 NetCDFs

Until GDAL puts work into the NetCDF driver, reading remote HDF5/NetCDF files has very poor support. Additionally, there is no standard overview format as there is with GeoTIFFs, so titiler is very unhappy.

We can get around this to some degree by creating VRTs and making external overviews pointing at these VRTs.

1. Download the NetCDFs (hopefully on some server which has fast access to the S3 bucket).
2. Run bowser prepare-disp-s1 to create VRT files pointing in the .nc files, one per dataset we wish to browse.
3. Run bowser setup-disp-s1 to make a JSON file containing all the file metadata, so the browser knows where to look.
4. bowser run

In detail:

1. Download I use s5cmd for faster copying, but the aws CLI tool works too:

mkdir my_disp_files; cd my_disp-files
s5cmd --numworkers 4 cp 's3://opera-bucket/OPERA_L3_DISP-S1_IW_F11115*/OPERA_L3_DISP-S1_IW_F11115*.nc' .

2. Extract VRTS

bowser prepare-disp-s1 -o vrts *.nc

3. Create bowser_rasters.json metadata

bowser setup-disp-s1 vrts

4. Run

bowser run

If this is on a server instead of your laptop, you'll need (on your laptop)

ssh -N -L 8000:localhost:8000 myserver

for whatever port pops up after bowser run

CLI Usage

$ bowser --help
Usage: bowser [OPTIONS] COMMAND [ARGS]...

  CLI for bowser.

Options:
  --help  Show this message and exit.

Commands:
  addo           Add compressed GDAL overviews to files.
  run            Run the web server.
  set-data       Specify what raster data to use.
  setup-dolphin  Set up output data configuration for a dolphin workflow.

To manually specify a raster/set of rasters, use the interactive bowser set-data

Developer Setup

npm is used to manage javascript dependencies.

npm install will install all dependencies.

Making changes to the UI

After making .tsx, HTML or CSS changes, you must run npm run build to build the project:

$ npm run build

> [email protected] build
> tsc && vite build

vite v5.4.19 building for production...
✓ 391 modules transformed.
src/bowser/dist/index.html      0.85 kB │ gzip:   0.51 kB
src/bowser/dist/index.css      25.68 kB │ gzip:   9.13 kB
src/bowser/dist/index.js    1,086.17 kB │ gzip: 241.12 kB

This updates the dist directory, which is served when you run bowser run ....

Currently the dist/ HTML and CSS files are checked in to git for easier deployment for non-javascript users.

License

Copyright 2024, by the California Institute of Technology. ALL RIGHTS RESERVED. United States Government Sponsorship acknowledged. Any commercial use must be negotiated with the Office of Technology Transfer at the California Institute of Technology.

This software may be subject to U.S. export control laws. By accepting this software, the user agrees to comply with all applicable U.S. export laws and regulations. User has the responsibility to obtain export licenses, or other export authority as may be required before exporting such information to foreign countries or providing access to foreign persons.