Feb 16th, 2023 code sync release

[renatolfc]

We are pleased to announce the release of the latest FarmVibes.AI version, which was made available on GitHub a week ago. We apologize for the delay in sending out this announcement and any inconvenience it may have caused.

Our team has been working tirelessly to deliver new features and enhancements to our software, and this release is no exception.

In this release, we deprecated the update-images command in favor of the update command, which not only updates the FarmVibes.AI images, but also updates the client library in the Python site packages.

We are aware of an issue with old pip versions, which are unable to install our latest version due to our versioning scheme. We are working on a fix, and should be releasing it soon. Please check #48 for tracking it.

As usual, if you already have cloned the repository:

git pull

should fetch the new code. We recommend an update of the library and images with:

bash farmvibes-ai.sh update

Please let us know if you have any issues.

Apart from that, we included the following changes, listed below, from our PR (#44) description:

We also fixed a DNS issue that could happen in some environments, which
prevented workflows from running, because the cluster would end up without
a proper resolver.

This PR fixes:

• Receiving Error: 'spaceeye.split' does not exist in the workflow graph when running visualize dataset notebook of "Crop Segmentation" #35;
• Error in running notebook crop_cycles.ipynb #40;
• Error in running notebook land_degradation.ipynb #41, and
• Received Input Validation Errors while running whatif.ipynb notebook for carbon sequestration. #42.

Further changes are:

Script

• [🎉 NEW] The update command supersedes the update-images command, and
  manages the Python library automatically
• [📈 IMPROVEMENT] We now use automatic retries by default when downloading
  binaries for cluster setup

Client

• [📈 IMPROVEMENT] We now show the total run duration in the monitor header

Workflows

• [🎉 NEW] We added a new workflow that downloads MODIS 8-day surface
  reflectance rasters that intersect with the input geometry and time range
• [🎉 NEW] We added a new workflow to download gNATSGO raster data that
  intersect with the input geometry and time range
• [📈 IMPROVEMENT] We added a step parameter to the chunk_onnx workflow to
  tune the step size of chunking
• [⚒️ FIX] We updated the linear_trend op to define the write_rel_limits
  attribute in RasterChunk

Notebooks

• [⚒️ FIX] We updated the crop_cycle notebook to use the step parameter to
  prevent errors when computing cycles with small regions
• [⚒️ FIX] We fixed the land_degradation notebook with our fix to the
  linear_trend op
• [⚒️ FIX] We updated the scenario.json file to match changes to the
  COMET-Farm API in the carbon whatif notebook
• [⚒️ FIX] We updated the wf_viz_crop_seg.yaml file with the correct sink name for the SpaceEye task
• [📈 IMPROVEMENT] The carbon whatif notebook now uses the monitor, instead of
  block_until_complete
• [📈 IMPROVEMENT] We ported the DeepMC notebook from Tensorflow to PyTorch

Backend

• [📈 IMPROVEMENT] We now add traceparent
  headers to our cloudevents, enabling better inspection of messaging in our
  clusters
• [📈 IMPROVEMENT] We now only deploy services in the cluster after images
  have been downloaded, making cluster creation 20% faster in our local
  tests
• [📈 IMPROVEMENT] We added an in-memory reader for ChipDataset
• [📈 IMPROVEMENT] We centralized hash computation, and improved our
  messaging system to pass cache information in messages, paving the way for
  improved caching mechanisms

Documentation

• [🎉 NEW] We added a new entry to the troubleshooting guide describing what
  to do when your containers run out of space