report_pdf: split waveform vs histogram layouts (BW PDF iteration)

Reviewed against real Blastware Event Report PDFs (uploaded to
example-events/pdfsnstuff/) for K558LLB7.V20H (histogram) and
K558LLB8.0E0W (waveform).  Each event type has its own layout because
BW's printouts genuinely differ:

  Waveform header:   Date/Time, Trigger Source, Range, Sample Rate
  Histogram header:  Start, Finish, Intervals At Size, Range, Sample Rate
                     (no trigger field — histograms aren't triggered)

  Waveform stats:    PPV, ZC Freq, Time (Rel. to Trig),
                     Peak Acceleration, Peak Displacement, Sensor Check
  Histogram stats:   PPV, ZC Freq, Date, Time (of peak), Sensor Check

  Waveform plot:     4-channel stacked line, x-axis in SECONDS,
                     trigger triangle + window markers, symmetric Y
                     for geo, zero-anchored mic, "0.0" baseline label
                     on right edge per BW convention
  Histogram plot:    4-channel stacked bars, Y-axis 0-to-peak only
                     (never negative — peaks are magnitudes), 0.0
                     baseline at the bottom

  Waveform footer:   USBM chart placeholder upper-right;
                     "Time X sec/div   Amplitude Geo: Y in/s/div   Mic: 0.001 psi(L)/div"
                     "Trigger = ▶━━◀"
  Histogram footer:  No USBM chart; same scale-info footer with
                     interval-size as the time unit

Other fixes from the first-pass screenshot review:
  - Channel labels (MicL/Long/Vert/Tran) no longer cut off (wider
    left margin)
  - Histogram bars rise from zero baseline (abs of any signed values)
  - ISO timestamp "2026-05-16T22:33:50" → "22:33:50 May 16, 2026"
    matching BW's display format

Known gaps (separate work):
  - Histogram codec returns per-block granularity (~200 bars for
    BW's 4-interval display).  XML-driven data source is the planned
    fix; the structured BW XML has the per-interval aggregates.
  - USBM RI8507 / OSMRE compliance chart still placeholder

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -8,7 +8,12 @@ All notable changes to seismo-relay are documented here.
 
 ### Added
 
-- **Event Report PDF generation** — `GET /db/events/{id}/report.pdf` returns a single-page letter-portrait PDF for any event with waveform data on disk.  Covers every field a Blastware Event Report includes: header metadata (date/time, trigger source, range, sample rate, project/client/operator/location, serial+firmware, battery, calibration, file name), microphone block (PSPL in dB(L) + psi, ZC freq, channel test), per-channel stats table (PPV / ZC Freq / Time of Peak / Peak Accel / Peak Disp / Sensor Check), Peak Vector Sum, and the 4-channel waveform plot stacked Instantel-style (MicL top → Tran bottom, shared time axis, trigger marker, symmetric Y on geo channels, zero-anchored on mic).  Histogram events render as per-interval bar charts instead of waveform plots.  USBM RI8507 / OSMRE compliance chart still stubbed — separate work item.  Backed by matplotlib (vector PDF output, no headless-browser dep); new `sfm/report_pdf.py` does data assembly + rendering.  **Visual layout is approximate** until reference PDFs land at `docs/reference/instantel/` to iterate against.
+- **Event Report PDF generation** — `GET /db/events/{id}/report.pdf` returns a single-page letter-portrait PDF for any event with waveform data on disk.  Covers every field a Blastware Event Report includes: header metadata (date/time, trigger source, range, sample rate, project/client/operator/location, serial+firmware, battery, calibration, file name), microphone block (PSPL in dB(L) + psi, ZC freq, channel test), per-channel stats table (rows differ for waveform vs histogram), Peak Vector Sum, and the 4-channel plot.  Iterated against real Blastware reference PDFs (uploaded to `example-events/pdfsnstuff/`):
+   - **Waveform layout**: header shows Date/Time, Trigger Source, Range, Sample Rate; stats table has PPV / ZC Freq / Time (Rel. to Trig) / Peak Accel / Peak Disp / Sensor Check; bottom plot is 4-channel line waveform (MicL top → Tran bottom), shared time axis in seconds, dashed trigger line + triangle marker at t=0, symmetric Y on geo channels, zero-anchored on mic, "0.0" baseline label on right per BW convention; footer shows `Time X sec/div   Amplitude Geo: Y in/s/div   Mic: 0.001 psi(L)/div` and the trigger window `▶━━◀` marker.  USBM RI8507/OSMRE compliance chart placeholder upper-right.
+   - **Histogram layout**: header shows Start / Finish / Intervals At Size / Range / Sample Rate (no Trigger Source — histograms aren't triggered); NO USBM chart; stats table has PPV / ZC Freq / Date / Time / Sensor Check; bottom plot is per-interval bar chart, Y-axis 0-to-peak (never negative), 0.0 baseline at the bottom; footer shows `Time INTERVAL_SIZE /div   Amplitude Geo: Y in/s/div   Mic: 0.001 psi(L)/div`.
+   - Backed by matplotlib (vector PDF, no headless-browser dep).  Adds matplotlib>=3.8 to deps.
+   - **Known gap**: histogram codec returns per-block granularity (~200 bars for a 4-interval event) instead of BW's per-interval aggregation.  Visual difference vs BW's 4-bar display.  XML-driven data source (parsing the structured `_XML.XML` files BW also exports) is the planned fix; that route also resolves the bw_ascii_report PPV-miss bug.
+   - **Stubbed**: USBM RI8507 / OSMRE compliance chart curves (separate work item; requires coding the regulatory piecewise functions).
 - **"Download PDF" button** in the event modal's footer — triggers the new endpoint; opens in a new tab so the browser handles save-or-display + surfaces any 404 / server errors visibly.
 
 - **SFM webapp now opens to Database view by default** and the History table is fully interactive.  Click any column header to sort ascending / descending (timestamp, serial, per-channel PPV, PVS, mic dB(L), project, client, record type, key — all sortable).  Click any event row to open the event modal, which now renders a **4-channel waveform plot inline** (MicL / Long / Vert / Tran stacked, Instantel-printout order) alongside the existing sidecar review fields.  Headers are sticky so the columns stay visible while scrolling long event lists.  No more "where is the viewer" — pick a unit from the filter dropdown, scan the table, click the event, see the waveform.