Add a new program to create plots for the entire LRAUV Deployment

[MBARIMike]

This new program will combine the processed data from all the log_files that make up a Deployment, as specified in the .dlist file. The .dlist file is in the YYYY-formatted directory in the directory tree for the log_files. A suitable name for this new program is lrauv_deployment_plots.py. The .png file(s) will be produced in the same directory as the .dlist file and be named using the "Deployment name:" string listed in the .dlist file with spaces and slashes replaced by underscores. Along with the .png files (multiple .pngs if biolume or planktivore data exist) the new program will also write a .html file containing a link the .png file and links to the individual log_file images and data that make up the deployment.

[MBARIMike]

The implementation plan created by me promting Claude Sonnet 4.6:

Plan: lrauv_deployment_plots.py (Issue #148)

TL;DR: New standalone lrauv_deployment_plots.py that reads a .dlist file, concatenates all _1S.nc files in the deployment, generates aggregated PNG plots and an HTML summary. CreateProducts is extended minimally to accept a pre-assembled dataset, configurable output dir, and plot name stem.

---

Phase 1 — Extend CreateProducts (create_products.py)

1. Add three optional params to __init__: ds: xr.Dataset | None = None, output_dir: Path | None = None, plot_name_stem: str | None = None

   • Store as self.ds, self.output_dir, self.plot_name_stem

2. Modify _open_ds(): add early return if self.ds is already set (skip file loading)

3. In each of the three plot methods (plot_2column, plot_biolume_2column, plot_planktivore_2column):

   • Output directory: use self.output_dir if set, else the existing netcdfs_dir = Path(BASE_LRAUV_PATH, ...) logic
   • Filename stem: use self.plot_name_stem if set, else Path(self.log_file).stem

Relevant references:

• CreateProducts.__init__ — src/data/create_products.py ~L93
• _open_ds() — src/data/create_products.py ~L199
• plot_2column() / save logic — src/data/create_products.py ~L1748-L1756
• plot_biolume_2column() / save — ~L1882-L1890
• plot_planktivore_2column() / save — ~L2018-L2027

---

Phase 2 — New lrauv_deployment_plots.py

4. Create src/data/lrauv_deployment_plots.py with a DeploymentPlotter class and CLI.

   CLI args:

   • --dlist (required): path to .dlist file, relative to BASE_LRAUV_PATH or absolute
   • -v / --verbose (optional)

5. Parse .dlist file to get deployment name (reuse get_deployment_name() from utils.py, or parse directly). Convert spaces → underscores for filenames.

6. Derive deployment directory from dlist path:
   {vehicle}/missionlogs/{YYYY}/{YYYYMMDD_YYYYMMDD}/

7. Enumerate all *_1S.nc files across all subdirectories of the deployment dir (sorted by time/path).

8. Concatenate with xr.open_mfdataset(files, combine='by_coords') (fallback: manual xr.concat([xr.open_dataset(f) for f in files], dim='time', join='outer')).

9. Instantiate CreateProducts with combined dataset:

   cp = CreateProducts(
       log_file=first_log_file,   # needed for _is_lrauv() and instrument detection
       ds=combined_ds,
       output_dir=dlist_dir,
       plot_name_stem=deployment_name,
       verbose=...,
   )

10. Call plot methods conditionally:

    • Always: cp.plot_2column()
    • If biolume variables present in combined_ds: cp.plot_biolume_2column()
    • If planktivore variables present: cp.plot_planktivore_2column()

---

Phase 3 — HTML file generation

11. Write {deployment_name}.html to the same directory as the .png files (YYYY/YYYYMMDD_YYYYMMDD/) containing:
    • H1: Deployment name
    • Inline <img> tags or links for each deployment PNG produced
    • Table of log_files with columns:
      • Log file stem (identifier)
      • Links to per-log-file PNGs (local relative paths)
      • Links to OPeNDAP data: http://dods.mbari.org/opendap/data/lrauv/{log_file}{.dmr.html | .html} depending on .nc4 vs .nc extension

---

Relevant files

• src/data/create_products.py — extend with ds/output_dir/plot_name_stem params
• src/data/lrauv_deployment_plots.py — new file to create
• src/data/utils.py — reuse get_deployment_name() for dlist parsing
• src/data/nc42netcdfs.py — BASE_LRAUV_PATH, LRAUV_OPENDAP_BASE constants
• src/data/resample.py — FREQ, LRAUV_OPENDAP_BASE constants

---

Verification

1. Run python lrauv_deployment_plots.py --dlist tethys/missionlogs/2012/20120908_20120920.dlist -v 1
2. Verify PNG files written to data/lrauv_data/tethys/missionlogs/2012/20120908_20120920/ named CANON_september2012_2column_cmocean.png etc.
3. Verify HTML written alongside PNGs in YYYY/YYYYMMDD_YYYYMMDD/
4. Verify individual plot_2column() still works for a single log_file (regression test)
5. Run existing test suite: pytest src/data/test_process_lrauv.py

---

Decisions

• PNG/HTML output: deployment PNGs and HTML all go in YYYY/YYYYMMDD_YYYYMMDD/
• Concat strategy: try xr.open_mfdataset(files, combine='by_coords') first; on any error fall back to xr.concat([xr.open_dataset(f) for f in files], dim='time', join='outer')
• log_file for CreateProducts: pass first log_file for _is_lrauv(); use join='outer' concat so the combined dataset includes ALL variables from ALL log_files (NaN-filled where absent) — biolume/planktivore plot guards will correctly fire if any log_file in the deployment had those variables
• Missing _1S.nc: warn and skip

[MBARIMike]

It might be nice to have some kind of notification option to email or send to Slack messages with the newly created images and links to data that are in the .html files.

[MBARIMike]

So far, the combined deployment plot looks like:

With a corresponding .html file available at:

https://dods.mbari.org/data/lrauv/ahi/missionlogs/2025/20251022_20251024/Ahi_38_i2map_1S_2column_cmocean.html