test: add manual test for mask from segmentation

manual_tests/segmentation_from_api.py

@@ -0,0 +1,86 @@
+import logging
+import os
+
+from cloudvolume import CloudVolume
+from cryoet_data_portal import Annotation, Client
+
+from cryoet_data_portal_neuroglancer.io import load_omezarr_data
+from cryoet_data_portal_neuroglancer.precompute.mesh import (
+    clean_mesh_folder,
+    generate_multiresolution_mesh_from_segmentation,
+    generate_single_resolution_mesh,
+)
+from cryoet_data_portal_neuroglancer.precompute.segmentation_mask import encode_segmentation
+from cryoet_data_portal_neuroglancer.utils import determine_size_of_non_zero_bounding_box
+
+logging.basicConfig(level=logging.DEBUG, force=True)
+
+zarr_path = "102-membrane-1.0_segmentationmask.zarr"
+output_path = "102-membrane-1.0_segmentationmask_encoded"
+resolution = 1.0  # Non 1.0 is a harder test
+
+
+def grab_annotation():
+    client = Client()
+    if not os.path.exists(zarr_path):
+        annotation = Annotation.get_by_id(client, 2480)
+        annotation.download(format="zarr")
+
+
+def make_precomputed_segmentation():
+
+    encode_segmentation(
+        zarr_path,
+        output_path,
+        (resolution, resolution, resolution),
+        include_mesh=False,
+        convert_non_zero_to=1,
+        delete_existing=False,
+    )
+
+
+def make_multi_res_mesh():
+    dask_data = load_omezarr_data(zarr_path)
+    mesh_shape = determine_size_of_non_zero_bounding_box(dask_data)
+    del dask_data
+    generate_multiresolution_mesh_from_segmentation(
+        output_path,
+        "mesh",
+        max_lod=2,
+        mesh_shape=mesh_shape,
+        max_simplification_error=0,
+    )
+    clean_mesh_folder(output_path, "mesh")
+
+
+def make_single_res_mesh():
+    """Likely to fail due to memory constraints"""
+    dask_data = load_omezarr_data(zarr_path)
+    generate_single_resolution_mesh(
+        dask_data,
+        output_path,
+        "single_mesh",
+        (resolution, resolution, resolution),
+    )
+
+
+def serve_files():
+    # Multi res mesh
+    cv = CloudVolume(f"file://{output_path}")
+    cv.viewer(port=1337)
+
+    # Single res mesh
+    # cv = CloudVolume(f"file://{output_path}", mesh_dir="single_mesh")
+    # cv.viewer(port=1338)
+
+
+def main():
+    grab_annotation()
+    make_precomputed_segmentation()
+    make_multi_res_mesh()
+    # make_single_res_mesh()
+    serve_files()
+
+
+if __name__ == "__main__":
+    main()