Session Handoff — Why Dachshunds: download final 3, fix misfile, complete the library

Where it started

Continuation of the Why Dachshunds build. This stretch: download the last 3 videos (#99, #180, #250), then add music+logo to just those 3, and make the tracker accurate. A misfiling bug surfaced and was fixed mid-task.

Decisions locked + what shipped

• Last 3 videos downloaded + finished. #99, #180, #250 pulled (Round 5 staging), then music (10%/5s) + logo (bottom-left 200px) applied to just those 3, with backups in each folder's _originals-no-music/ and _pre-logo/.
• Misfile fixed. The by-ID filer (scripts/dachshund/build_file_plan_by_id.py) fell back to its --batch 6 arg and put all 3 in Batch 6. Corrected: #99 moved to Batch 2 - 80 Sondra Voice, #180 to Batch 4 - 60 Sondra Voice (per tracker Batch column); #250 correctly in Batch 6. Known tool gap: build_file_plan_by_id should derive batch per-number from the tracker, not rely on --batch.
• Tracker is now accurate -> Why_Dachshunds_MASTER_Tracker_v21.xlsx. Backfilled 130 stale rows; all 250 now Created=Y + Downloaded=Y, every number confirmed physically on disk (250/250).
• Library is COMPLETE. All 250 made, downloaded, filed in correct batch folders, music + logo on all 257 files (incl. a/b variants).

Key files for next session

• C:\Users\srver\AIOS-Workspace\aios-starter-kit\PROJECTS\YT - Why Dachshunds Channel\Why_Dachshunds_MASTER_Tracker_v21.xlsx — current master (untracked / Backblaze-only).
• C:\Users\srver\AIOS-Workspace\aios-starter-kit\scripts\dachshund\add_music.py, add_logo.py, build_file_plan_by_id.py — toolkit. NOTE: add_music/add_logo process a WHOLE folder and source from backups — never run them on a full batch folder to fix a single file (it would strip logos off the rest); process single files inline instead.
• C:\Users\srver\AIOS-Workspace\aios-starter-kit\Sondra Brain\wiki\projects\why-dachshunds-channel.md — production state + locked specs.
• C:\Users\srver\AIOS-Workspace\aios-starter-kit\outputs\thumbnails\why-dachshunds-01\ — generated thumbnails (with-text + *-NOTEXT.png) + gallery (also live at https://sturdy-lodge-shk2.here.now/).
• Memory: C:\Users\srver\.claude\projects\c--Users-srver-AIOS-Workspace-aios-starter-kit\memory\feedback_ceo_dashboard_placement.md (touched earlier today, not this stretch).

Running state

• Background processes: none (logo batch bj8u8f1xx completed: 257/257, 0 fails).
• Dev servers / ports: none
• Open worktrees / branches: none (on main)

Verification — how to confirm things still work

• python -X utf8 -c "import openpyxl; wb=openpyxl.load_workbook(r'PROJECTS/YT - Why Dachshunds Channel/Why_Dachshunds_MASTER_Tracker_v21.xlsx',data_only=True); ws=wb['Master Tracker']; rows=list(ws.iter_rows(values_only=True)); idx={h:i for i,h in enumerate(rows[0])}; print([r[idx['#']] for r in rows[1:] if r[idx['#']] and str(r[idx['Final Downloaded to Drive?']] or '').strip().upper()!='Y'])" — expect [].
• ffprobe audio check on Batch 2/.../99 - .mp4, Batch 4/.../180 - .mp4, Batch 6/.../250 - *.mp4 — each returns audio.

Deferred + open questions

• Open: Final music track — all 257 use the coffee time.wav test bed; re-run add_music.py per folder from _originals-no-music/ when chosen (never stacks).
• Open: Channel thumbnail font — text-free bases ready (*-NOTEXT.png); next is a Pillow text-overlay script with a locked font/color/outline.
• Open: Winning thumbnail concept (Concept 2 "In His Head" recommended).
• Deferred: Batch 1 Claymation subfolder got neither music nor logo (only Animated videos).
• Deferred: fix build_file_plan_by_id.py to always derive batch from the tracker per-number.

Pick up here

Library is complete; next is the thumbnail text-overlay system — build a Pillow script and show Sondra 3-4 font options on a text-free base to lock the channel signature.