6. Run the processing chain#

You can drive the chain either through the run notebook or directly from a shell. The notebook is the recommended way for a first run; the shell approach is better suited to batch processing once you understand the chain.

Option A: from the notebook#

Open notebooks/2_BPS_Run.ipynb and edit cell 0 with at minimum:

INPUT_FOLDER      = "/home/yourname/bps-work/data/run_001"
PROCESSOR_VERSION = "04.43"
MISSION_PHASE     = "TOMOGRAPHIC"   # or "INTERFEROMETRIC"

The notebook executes the six cells described below (6.1 → 6.6), each of which can also be run on its own from a shell.

Option B: from a shell#

# 1. Activate the BPS environment
bps443                                   # or:  conda activate bps-dev  (source path)

cd ~/bps-work/BPS/docs/tutorials/run-bps-locally

# 2. Generate JobOrder XMLs for L1F
python scripts/JOBuilder.py L1F 04.43 ~/bps-work/data/run_001

# 3. Run framing on each JobOrder
for jo in ~/bps-work/data/run_001/JobOrder_L1F_*.xml; do
  bps_l1_framing_processor "$jo"
done

# 4. Generate and run L1 (chain mode, picks up EOF frames from L1F output)
python scripts/JOBuilder.py L1_chain 04.43 ~/bps-work/data/run_001
for jo in ~/bps-work/data/run_001/JobOrder_L1_*.xml; do
  bps_l1_processor "$jo"
done

# 5. Generate and run STA (Tomographic phase)
python scripts/JOBuilder.py STA_chain 04.43 ~/bps-work/data/run_001 TOMOGRAPHIC
for jo in ~/bps-work/data/run_001/JobOrder_STA_*.xml; do
  bps_stack_processor "$jo"
done

# 6. Generate and run L2A
python scripts/JOBuilder.py L2A_chain 04.43 ~/bps-work/data/run_001
for jo in ~/bps-work/data/run_001/JobOrder_L2A_*.xml; do
  bps_l2a_processor "$jo"
done

What each step does#

6.1 Check environment#

Verifies that the biofetch conda environment exists, creating it from scripts/biofetch.yml if missing. biofetch is the lightweight env used by JOBuilder.py; it does not require the full BPS install.

6.2 Check input folder#

Sanity-check that INPUT_FOLDER/Inputs/ exists and is not empty. If you see ⚠️ Inputs/ not found, go back to step 4.

6.3 L1F chain (framing)#

Reads the raw ISP packets (RAW_0S), determines frame boundaries, and produces framed L1 virtual products (EOF files, one per frame). After L1F completes, inspect the generated frames_* folder and choose which frames to process downstream:

SELECTED_FRAMES = []              # empty = process ALL frames
SELECTED_FRAMES = ["306", "307"]  # specific frame IDs

6.4 L1 chain (SLC focusing)#

Focuses each selected frame into a Single Look Complex (SLC) product (SCS__1S). Additional AUX files (AUX_INS, AUX_PP1) are resolved by JOBuilder.py in this priority order:

  1. INPUT_FOLDER/Inputs/ (if placed manually)

  2. AUX_USER_DIR (from config.ini)

  3. AUX_DEFAULT_DIR (from config.ini)

Output goes to INPUT_FOLDER/OUTPUT_L1/.

6.5 STA chain (stack processor)#

Stacks multiple SLC products from different repeat cycles over the same frame and produces coherence/interferometric stack products (STA__1S).

Warning

STA requires multiple acquisitions over the same frame. If you only downloaded a single repeat cycle, STA will either fail or produce a degenerate result. Use a full major cycle (7 repeat passes) for the complete chain.

The MISSION_PHASE variable selects TOMOGRAPHIC or INTERFEROMETRIC. Output goes to INPUT_FOLDER/OUTPUT_STA_<frame>_V<version>/.

6.6 L2A chain (biomass estimation)#

Reads the STA stack products and estimates L2A products per frame. Requires the AUX_PP2_2A_* file in the AUX directory configured in config.ini. Output goes to INPUT_FOLDER/OUTPUT_L2A_<frame>_V<version>/.