Home

To-do List:

(High priority and/or Should be easy)

General

• Build out the wiki and document each script in detail
• Have single.py grab data since previous trigger in case system crash misses a run?
• Benchmark tests
• Multiple station handling
• Publish! (SRL or JStatSoft)

Initial setup

• Option to initially populate with nearby previously cataloged events (would also need to know to not count these twice)
• Eventually make setup interactive, rather than config file? (or make building of config file interactive)

Triggering

• Removing outright bad triggers (e.g., spikes, calibration pulses), set aside borderline cases: kurtosis complete in time and frequency domains, lots of bad triggers still getting through - in progress
• Test (incorporate?) peak matching for faster initial correlation and alignment when trigger is late?
• Error handling when encountering a gap or no data in waveserver/IRIS data so script completes instead of just choking
• Complains about updating id when no new events are found in the time chunk?

Correlation

• Correlation between cores (may help with preventing leftovers?)
• Correlation of core with full cluster - Alicia in progress
• Realign clusters to core
• Handling of extremely large families (cross-correlation in parallel, smarter clustering?? use stack? subfamilies? kernelized least squares hashing?)

Tables

• Option for manual removal of clusters or correlated noise

OPTICS / Clustering

• Clustering optimization (IncOPTICS? also cleaning up that code in general)

Plotting

• Cluster stacks and clearer waveform plots
• Occurrence Timeline (family number vs. time, color by correlation to core or amplitude)
• Use timing to produce stacks on multiple stations from a single station run
• Interactive plotting to further investigate a single cluster

Recently Implemented

• Support for different sampling rates (was originally hard-coded to 100 Hz)
• Removed some hard coding in plotting, added padding
• Option to designate where waveform data come from: IRIS or waveserver
• Setup of configuration via config file
• Orphan expiration
• Store last ID and trigger time in header to pick up where last run left off
• Write control program that can be run as a cron job