Add Sentinel-1D (S1D) support

[mgovorcin]

This PR extends s1-reader to support S1D, following the same pattern as the S1C support added in #144.

Changes

s1_burst_slc.py

• relative_orbit_number: add S1D branch with orbit_number_offset = 42,
  derived from the first available S1D acquisition (abs_orbit=2389 → rel=73,
  track 73 confirmed via manifest.safe). Note: offset will be validated against
  the calibrated orbit in July 2026.
• polyfit: fix numpy 0-d array TypeError when calling %f % on
  np.sqrt(...) — add .item() to extract the Python scalar. Required for
  numpy >= 2.0. already addressed in Fix numpy >= 2.0 TypeError in polyfit #150

s1_info.py

• Extend glob pattern S1[AB]_IW* → S1[ABCD]_IW* so the s1_info CLI
  discovers S1D SAFE files in a directory.

Testing

Verified with S1D IW SLC acquired 2026-03-13 and 2026-04-17 (track 73, Croatia).
load_bursts() returns correct burst IDs, relative_orbit_number returns 73, and bursts geocode successfully with compass. Note: that relative orbit offset will need to be update in July, once they finish with calibration.

This if for ESA wiki page: "Sentinel-1D is currently repeating Sentinel-1C's ground track with a 1-day delay to help calibrate its integration.In July 2026, a maneuver phase will move 1C into its final configuration, re-establishing the standard 6-day revisit frequency between 1C and 1D."

S1D GSLC results browse image generated with COMPASS:

Successfully generated a single-day S1C × S1D interferogram over Croatia (track 73, 27 bursts, full IW frame) to validate S1D burst loading and geocoding compatibility with compass/isce3:

Inputs:
(1-day temporal baseline)

• Reference: S1C_IW_SLC__1SDV_20260416T164125_20260416T164152_007244_00EAE8_2D6E (RESORB)
• Secondary: S1D_IW_SLC__1SDV_20260417T164135_20260417T164202_002389_003E83_9EFF (RESORB)
• DEM: Copernicus GLO-30
• Geocoding: compass, 5×10m grid, bbox [15.58, 45.55, 19.33, 47.41]

1-day interferogram from S1C and S1D: