Introducing new Eynollah line extraction method to QA

[jarmoza]

We will now be integrating the new eynollah line extraction method into the current QA code for the purposes of comparing it to the original watershed line extraction method.

[jarmoza]

Paused while error investigation/error processing for line extraction QA is implemented.

[jarmoza]

Plan for testing/integration of Eynollah line extraction (via meeting with @nvog and @kartikgo on December 18, 2023)

Testing out Eynollah Line Extraction

• Get eynollah working on test example with line_extract_eynollah.sh in code folder
• Downscaling method in shell script

Notes:

• Comment out last line until rest working python3 -u $CODE/eynollah_line_image_extraction.py... (This will be added back in at the end of testing Eynollah).
• Try without downscaling first for baseline processing time then introduce downscaling for comparison (try 25% at first).
• CPU count will probably have to be ~8 per process with 2GB memory each

Integration with QA Line Extraction

• Converting parts of shell script to Python calls like the watershed QA code does
• Downscale via image library and then save images to disk
• New set of QA subprocess methods for eynollah
• Incorporate new 'norm_height' column for watershed line extraction QA

Line Extraction Comparison

• New 'shootout' code introduced to qa-workflow that takes results of 2+ methods and produces comparison output csv

[jarmoza]

Reopening this issue as running Eynollah from QA needs to be reimplemented using the new Eynollah workflow scripts.