sar-availability developers guide

Backend bos sarcat scraper (stand-alone CLI tool)

• https://github.com/aria-jpl/bos_sarcat_scraper

  • https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/bos_sarcat_scraper/bosart_scrape.py

    • This script queries BOS and outputs a json of te result set

    • Inputs

      • start/end time

      • OR

      • Since last ingest time on BOS

    • BOS expects wkt format for query

    • 
      

Front end GUI

• Currently inside tosca repo on https://github.com/hysds/tosca/blob/develop/tosca/templates/facetview-saravail.html

• Front end Faceted search is house on aria-puccini

• To edit the template:

$ ssh aria-puccini #use LDAP creds
$ ssh ariaops@localhost
$ vim /usr/local/ariaops/anaconda2/ops/tosca-sar_avail/tosca/templates/facetview-saravail.html
$ supervisorctl restart tosca-sar_avail

PGE: bos ingest

• https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/docker/job-spec.json.bos_ingest

• Runs hourly on 2-hour

• This is the HySDS PGE wrapper for https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/bos_sarcat_scraper/bosart_scrape.py

PGE: bos scrubbed of older

• https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/docker/job-spec.json.scrub_outdated_bos_acqs

• Runs daily

• Scrubs outdated PLANNED and PREDICTED, after 2-days margin

Cron scripts

• https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/archived_planned-cron.py

• https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/scrub_planned_predicted-cron.py

• https://github.com/aria-jpl/bos_sarcat_scraper/blob/master/create_acquisitions.py

  • Avoids ingesting past PLANNED and PREDICTED

  • Calls ingest from inside te PGE

    • If forgot to delete the folder, then verdi will try to ingest it if the dataset dir is still there.

    • Bug if cannot remove dir and verdi

  • Currently no retries. If job failed due to any reason like ES timeout, then job fails. Currently relying on next scrubber to run. But only get 3-hour sliding window, so about 3 tries.

  • Still need a daily 5-day window back-filler to run bos ingest.

• This runs on b-cluster factotum currently

Other

• Location

• Log files

• Debugging process

• Deployment

• Watchdogs to check on hourly scraper already in. current checks on bos_ingest_:master

• ES on b-cluster

  • Alias for sar-availability: acquisition