Configuration Overview

The Hybrid Science Data System (HySDS) interface (i.e. dist-s1 run_sas) assumes all the requisite OPERA RTC-S1 data is localized to the runtime environment and a runconfig.yml has been provided with relevant local paths. To run this entrypoint is simply:

dist-s1 run_sas --run_config_path run_config.yml

For a complete example (in which a run_config is generated in the process) use:

dist-s1 run_sas_prep --mgrs_tile_id '11SLT' \
    --post_date '2025-01-21' \
    --track_number 71 \
    --dst_dir '../../notebooks/los-angeles' \
    --memory_strategy 'high' \
    --low_confidence_alert_threshold 3.5 \
    --high_confidence_alert_threshold 5.5 \
    --apply_water_mask true \
    --device 'cpu' \
    --product_dst_dir '../../notebooks/los-angeles' \
    --model_source 'transformer_original' \
    --use_date_encoding false \
    --model_dtype 'float32' \
    --n_workers_for_norm_param_estimation 4 \
    --batch_size_for_norm_param_estimation 32 \
    --stride_for_norm_param_estimation 4 \
    --algo_config_path algo_config.yml \
    --run_config_path run_config.yml && \
dist-s1 run_sas --run_config_path run_config.yml

Sample yml configs can be found here:

• run_config.yml
• algo_config.yml

Note: There is a field algo_config_path in run_config.yml so that the <algo_config_path>.yml is loaded within RunConfigData data model to specify an appropriate AlgoConfigData instance that is in line with the algo_config.yml. We have provided tables to the relevant fields for the configuration ymls here:

• RunConfig Parameter Table
• AlgoConfig Parameter Table

A RunConfig and AlgoConfig expose all the different parameters that can be specified for the generation of a DIST-S1 product. The parameters are written as gdal tags to the final DIST-S1 product for provenance and can be read through rasterio in any of the final COG layers:

import rasterio

with rasterio.open(<PATH_TO_A_DIST_S1_LAYER>) as ds:
    tags = ds.tags()