Redesign batch segment source selection

README.md

@@ -162,7 +162,7 @@ Use the batch wrapper to run `zed_svo_grid_to_mp4` over many segment directories
 
 ```bash
 uv run python scripts/zed_batch_svo_grid_to_mp4.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --recursive \
     --jobs 2 \
     --encoder-device auto \
@@ -173,8 +173,8 @@ You can also provide the exact segments to convert:
 
 ```bash
 uv run python scripts/zed_batch_svo_grid_to_mp4.py \
-    --segment-dir <SEGMENT_DIR_A> \
-    --segment-dir <SEGMENT_DIR_B> \
+    --segment <SEGMENT_DIR_A> \
+    --segment <SEGMENT_DIR_B> \
     --jobs 2
 ```
 
@@ -193,7 +193,7 @@ When you suspect a previous run left behind partial MP4 files, opt into `ffprobe
 
 ```bash
 uv run python scripts/zed_batch_svo_grid_to_mp4.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --probe-existing \
     --jobs 2
 ```
@@ -202,7 +202,7 @@ Use `--report-existing` to audit existing outputs without launching conversions.
 
 ```bash
 uv run python scripts/zed_batch_svo_grid_to_mp4.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --report-existing
 ```
 
@@ -236,7 +236,7 @@ Use the wrapper to recurse through a dataset root, run `zed_svo_to_mcap --segmen
 
 ```bash
 uv run python scripts/zed_batch_svo_to_mcap.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --recursive \
     --jobs 2 \
     --cuda-visible-devices GPU-9cc7b26e-90d4-0c49-4d4c-060e528ffba6 \
@@ -258,7 +258,7 @@ Enable per-camera pose export when the segment has valid tracking:
 
 ```bash
 uv run python scripts/zed_batch_svo_to_mcap.py \
-    --segment-dir <SEGMENT_DIR> \
+    --segment <SEGMENT_DIR> \
     --with-pose \
     --pose-config <POSE_CONFIG>
 ```
@@ -275,6 +275,8 @@ Single-source `zed_svo_to_mcap` now writes the one-camera `copy` shape by defaul
 
 For the simple non-GUI path, use `scripts/mcap_rgbd_example.py` and [docs/mcap_recipes.md](./docs/mcap_recipes.md). That helper supports current `bundled` and `copy` MCAPs, and it also accepts the legacy `/camera/*` shape by treating it as a single-camera stream with the literal label `camera`.
 
+For calibration-based depth/RGB mapping, use `scripts/mcap_depth_alignment.py` and [docs/depth_alignment.md](./docs/depth_alignment.md). That helper explains the current affine mapping implied by the exported calibration topics and can export example aligned-depth and overlay PNGs from a chosen MCAP frame.
+
 ### MCAP RGBD Viewer
 
 The repo includes an example RGB+depth viewer at `scripts/mcap_rgbd_viewer.py`. It supports legacy standalone `/camera/*` MCAPs, bundled `/bundle` + `/zedN/*` MCAPs, and `copy` MCAPs with namespaced `/{label}/*` topics and no `/bundle`, including the default single-source output from `zed_svo_to_mcap`.
@@ -322,7 +324,7 @@ That is why the batch wrapper supports mixed pools such as two NVENC workers plu
 
 ```bash
 uv run python scripts/zed_batch_svo_to_mcap.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --recursive \
     --overwrite \
     --hardware-jobs 2 \
@@ -340,7 +342,7 @@ Use `--probe-existing` to validate existing MCAPs before skipping them. Invalid
 
 ```bash
 uv run python scripts/zed_batch_svo_to_mcap.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --probe-existing \
     --jobs 2
 ```
@@ -349,7 +351,7 @@ Use `--report-existing` to audit existing MCAPs without launching conversions:
 
 ```bash
 uv run python scripts/zed_batch_svo_to_mcap.py \
-    <DATASET_ROOT> \
+    --dataset-root <DATASET_ROOT> \
     --report-existing
 ```