Add Plumbing tests using Allen's own MapMyCells example files

Fetched from https://brain-map.org/bkp/analyze/mapmycells/files (verified
both URLs resolve): a 1k-cell human MTG example and a 10k-cell whole-mouse-
brain example, both small enough for a fast/cheap smoke test. Staged at
gs://pd-test-storage-public/MapMyCells/input/plumbing/{human,mouse}/,
matching the <Pipeline>/input/{plumbing,scientific}/ layout other WARP
pipelines use.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Author: rsc3

pipelines/wdl/mapmycells/MapMyCells.changelog.md

@@ -10,3 +10,4 @@
 * Repointed `test_inputs/Plumbing/test_human.json` and `test_mouse.json` at real, existing gene-expression h5ads borrowed from the scANVI pipeline's own test set (`10k_PBMC_gex.h5ad`, `Mouse_Hippocampus_AIT_plumbing_gex.h5ad`) instead of `query.h5ad`/`query_mouse.h5ad`, which never existed; MapMyCells only consumes gene expression, so only the gex file from each scANVI test case is reused, not the ATAC or reference h5ads
 * Added a Citing section to the README crediting the Allen Institute's `cell_type_mapper` (source code + the MapMyCells preprint, Daniel et al. 2026, bioRxiv doi:10.64898/2026.03.06.710160)
 * Moved `test_human.json`/`test_mouse.json` from `test_inputs/Plumbing/` to `test_inputs/Scientific/` -- these become the Scientific tests; new, smaller Plumbing tests to follow separately. No truth re-seed needed for this move: `UpdateTestInputs.py` derives the GCS truth tier from the input filename/content, not from which `test_inputs/` subdirectory the file lives in, and both files' already-seeded truth already happens to sit at the paths this heuristic computes for them (`test_mouse` under `.../truth/plumbing/...`, `test_human` under `.../truth/scientific/...` -- the latter was the "known issue" from the previous entry, now moot since `test_human` actually is a Scientific test)
+* Added new `test_inputs/Plumbing/test_human.json` and `test_mouse.json` using Allen Institute's own MapMyCells example query files (https://brain-map.org/bkp/analyze/mapmycells/files: 1k-cell human MTG example, 10k-cell whole-mouse-brain example) staged at `gs://pd-test-storage-public/MapMyCells/input/plumbing/{human,mouse}/`, matching the `<Pipeline>/input/{plumbing,scientific}/` convention other WARP pipelines use for test data

pipelines/wdl/mapmycells/README.md

@@ -38,8 +38,8 @@ The task generates two output files:
 ## Usage
 
 You can find example inputs in the `test_inputs/` folder:
-* `test_inputs/Scientific/test_human.json` — Maps using `"Human_MTG"`
-* `test_inputs/Scientific/test_mouse.json` — Maps using `"Mouse_WMB"`
+* `test_inputs/Plumbing/test_human.json` / `test_inputs/Plumbing/test_mouse.json` — small smoke tests, using Allen Institute's own [MapMyCells example files](https://brain-map.org/bkp/analyze/mapmycells/files) (1k and 10k cells respectively)
+* `test_inputs/Scientific/test_human.json` / `test_inputs/Scientific/test_mouse.json` — full-scale runs
 
 The Human_MTG and Mouse_WMB reference-atlas assets are baked into the pipeline's docker image (see [warp-tools/3rd-party-tools/mapmycells](https://github.com/broadinstitute/warp-tools/tree/develop/3rd-party-tools/mapmycells)), so you do not need to download the reference atlases or pass them into the workflow yourself. Simply specify the `reference_atlas` and provide your `query_h5ad`!
 

pipelines/wdl/mapmycells/test_inputs/Plumbing/test_human.json

@@ -0,0 +1,7 @@
+{
+  "MapMyCells.input_id": "human_mtg_example",
+  "MapMyCells.query_h5ad": "gs://pd-test-storage-public/MapMyCells/input/plumbing/human/10xhumanmtgsea-adccn20230505_example_1kcells_36601genes.h5ad",
+  "MapMyCells.reference_atlas": "Human_MTG",
+  "MapMyCells.cpu": 2,
+  "MapMyCells.memory_gb": 8
+}

pipelines/wdl/mapmycells/test_inputs/Plumbing/test_mouse.json

@@ -0,0 +1,7 @@
+{
+  "MapMyCells.input_id": "mouse_wmb_example",
+  "MapMyCells.query_h5ad": "gs://pd-test-storage-public/MapMyCells/input/plumbing/mouse/wholemousebrain_ccn20230722_example_10kcells_550genes.h5ad",
+  "MapMyCells.reference_atlas": "Mouse_WMB",
+  "MapMyCells.cpu": 2,
+  "MapMyCells.memory_gb": 8
+}