Skip to content

Snapshots

What a snapshot captures

A snapshot saves the current processing parameters and quality metrics for an event at a point in time. Specifically, it stores:

  • All event-level parameters: the time window, bandpass filter settings, and Min CC (min_cc) — see Aligning with ICCS for what each controls
  • Per-seismogram parameters for every seismogram in the event: the current t1 pick, select flag, and flip flag
  • Quality metrics, if available at snapshot time:
  • ICCS CC per seismogram (always present once the event has been opened)
  • MCCC metrics per seismogram and the global RMSE (present only if MCCC has been run)

The seismogram waveform data itself is not copied — snapshots are lightweight. They capture where you are in the parameter space, not the data.

This works because the CC seismograms and context seismograms that ICCS operates on are entirely deterministic: given the original waveform data and a set of parameters, they are always reconstructed identically. Restoring a snapshot therefore restores the exact state of the ICCS instance — there is nothing lost by not saving the derived arrays.

If seismograms are added to the project after a snapshot was taken, they have no entry in that snapshot. When previewing or rolling back, those seismograms are included using their current live parameters — the snapshot's event-level parameters (window, filter, Min CC) still apply to them.

Snapshots are per-event. Each event maintains its own list.

When to take a snapshot

Take a snapshot before making changes you might want to undo:

  • After importing data, before any processing — a clean baseline to return to
  • After initial alignment looks good, before tightening parameters further
  • Before trying an experimental configuration (different window, filter, etc.)
  • Before running MCCC

Snapshots are cheap. Taking one costs almost nothing, and having a rollback point available is worth it.

Creating a snapshot

aimbat snapshot create <ID>                        # no comment
aimbat snapshot create <ID> "after bandpass 1–3Hz" # with comment

snapshot create                        # no comment
snapshot create "after bandpass 1–3Hz" # with comment

The comment is optional but useful for identifying the snapshot later.

Press n to open the snapshot comment dialog, optionally enter a comment, and confirm. The new snapshot appears immediately in the Snapshots tab.

Click New Snapshot in the Processing tab. A dialog lets you enter an optional comment.

Listing snapshots

aimbat snapshot list <ID>  # for a specific event
aimbat snapshot list all   # across all events

snapshot list <ID>  # uses the current event context
snapshot list all   # across all events

The table shows the snapshot ID, date and time, comment, and number of seismograms captured.

Snapshots for the current event are listed in the Snapshots tab. Switch events using the event switcher (e) to see another event's snapshots.

The Snapshots tab lists all snapshots for the selected event.

Inspecting a snapshot

Before rolling back, it can be useful to see what a snapshot contains.

aimbat snapshot details <SNAPSHOT_ID>          # view saved event parameters
aimbat snapshot preview <SNAPSHOT_ID>          # view stack plot
aimbat snapshot preview --matrix <SNAPSHOT_ID> # view matrix image

snapshot details <SNAPSHOT_ID>
snapshot preview <SNAPSHOT_ID>
snapshot preview --matrix <SNAPSHOT_ID>

details shows the event-level parameters (window, filter, min_cc) as they were when the snapshot was taken. preview builds the ICCS stack from the snapshot's parameters and displays it — without modifying anything in the database.

Press Enter on a snapshot row in the Snapshots tab to open the action menu. Options include:

  • Show details — displays the saved event parameters
  • Preview stack — opens the stack plot built from the snapshot
  • Preview matrix image — opens the matrix image

Both preview options support the context (c) and all seismograms (a) toggles in the action menu before launching.

Select a snapshot in the Snapshots tab — its stack and matrix image are shown in the right panel in read-only mode.

Rolling back

Rolling back restores the snapshot's parameters as the current live values. This overwrites the current event and seismogram parameters for this event.

aimbat snapshot rollback <SNAPSHOT_ID>

snapshot rollback <SNAPSHOT_ID>

Press Enter on a snapshot row and choose Rollback to this snapshot. A confirmation dialog appears before any changes are made.

Select a snapshot and click Rollback to this.

After rolling back, the event's parameters are exactly as they were when the snapshot was taken. Any ICCS runs or parameter changes made after that snapshot are undone. The snapshot itself is not deleted — you can roll back to it again.

If the snapshot contains MCCC quality data, the live quality metrics are restored from the best matching snapshot: the one whose parameter hash matches the restored state and that has the most recent MCCC data. In practice this is the snapshot you rolled back to, but if that snapshot predates any MCCC run, the most recent snapshot with the same parameters and MCCC data is used instead.

Deleting a snapshot

aimbat snapshot delete <SNAPSHOT_ID>

snapshot delete <SNAPSHOT_ID>

Press Enter on a snapshot row and choose Delete snapshot. A confirmation dialog appears.

Select a snapshot and click Delete.

Deletion is permanent. The snapshot cannot be recovered after deletion.

Exporting snapshot data

For archiving or scripting purposes, snapshot data can be exported to JSON:

aimbat snapshot dump

snapshot dump

The output is a JSON object with five keys, all cross-referenced by snapshot_id:

Key Contents Always present?
snapshots Snapshot metadata (ID, time, comment, hash) Yes
event_parameters Event parameter snapshots Yes
seismogram_parameters Per-seismogram parameter snapshots Yes
event_quality Event quality snapshots (MCCC RMSE) Only if MCCC has been run
seismogram_quality Per-seismogram quality snapshots (ICCS CC, MCCC metrics) Only if quality metrics exist

Saving results

Any snapshot can be exported as a structured JSON document containing the frozen t1 picks, ICCS correlation coefficients, and — if MCCC was run before the snapshot — per-seismogram MCCC quality metrics and the event-level RMSE. MCCC does not need to have been run; the export is useful at any stage of processing.

This is the primary format for passing AIMBAT picks into downstream tools such as tomographic inversion codes. See Exporting Results for full details on the output format and how to work with it.

aimbat snapshot results <SNAPSHOT_ID>                    # print to stdout
aimbat snapshot results <SNAPSHOT_ID> --output out.json  # save to file

snapshot results <SNAPSHOT_ID>
snapshot results <SNAPSHOT_ID> --output out.json

Press Enter on a snapshot row in the Snapshots tab and choose Save results to JSON. A file-picker dialog opens; the suggested filename is results_<short_id>.json. Confirm to write the file.

Pass --alias to use camelCase field names in the output.

Snapshot notes

Each snapshot can carry a freeform Markdown note — useful for recording observations, decisions, or links to external references at the time the snapshot was taken.

aimbat snapshot note read <SNAPSHOT_ID>  # display the note
aimbat snapshot note edit <SNAPSHOT_ID>  # open in $EDITOR and save on exit

snapshot note read <SNAPSHOT_ID>
snapshot note edit <SNAPSHOT_ID>

If no note exists yet, read prints (no note) and edit opens an empty buffer. The note is saved when you close the editor without error.

Snapshot quality statistics

A summary of quality metrics across all snapshots for an event can be viewed without opening individual snapshot records:

aimbat snapshot quality list <ID>  # for a specific event
aimbat snapshot quality list all   # across all events
aimbat snapshot quality dump       # raw JSON export

snapshot quality list <ID>  # for a specific event
snapshot quality list all   # across all events
snapshot quality dump       # raw JSON export

The table shows per-snapshot aggregated ICCS correlation coefficients and, where MCCC has been run, MCCC metrics (mean, SEM) and the global RMSE. This makes it easy to compare the quality evolution across snapshots without having to export each one individually.