iris.util.broadcast_to_shape does not work with lazy data

[bouweandela]

🐛 Bug Report

How To Reproduce

Steps to reproduce the behaviour:

Run the following code:

>>> import dask.array as da
>>> from iris.util import broadcast_to_shape
>>> a = da.array([[1, 2, 3], [4, 5, 6]])
>>> b = broadcast_to_shape(a, (5, 2, 6, 3), (1, 3))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/bandela/mambaforge/envs/test/lib/python3.11/site-packages/iris/util.py", line 98, in broadcast_to_shape
    array_view = _broadcast_helper(array)
                 ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/bandela/mambaforge/envs/test/lib/python3.11/site-packages/iris/util.py", line 95, in _broadcast_helper
    strides[dim] = a.strides[idim]
                   ^^^^^^^^^
AttributeError: 'Array' object has no attribute 'strides'

Expected behaviour

The function documentation mentions that "This function maintains laziness when called; it does not realise data.", so I would expect the function to return a Dask array instead of crash when following the example from the docstring with a dask array as input.

Screenshots

Environment

• OS & Version: Ubuntu 23.04
• Iris Version: 3.5

$ mamba env export 

gives

name: test
channels:
  - conda-forge
dependencies:
  - _libgcc_mutex=0.1=conda_forge
  - _openmp_mutex=4.5=2_gnu
  - antlr-python-runtime=4.7.2=py311h38be061_1003
  - brotli=1.0.9=h166bdaf_8
  - brotli-bin=1.0.9=h166bdaf_8
  - brotlipy=0.7.0=py311hd4cff14_1005
  - bzip2=1.0.8=h7f98852_4
  - c-ares=1.18.1=h7f98852_0
  - ca-certificates=2022.12.7=ha878542_0
  - cartopy=0.21.1=py311hd88b842_1
  - certifi=2022.12.7=pyhd8ed1ab_0
  - cf-units=3.2.0=py311hcb2cf0a_0
  - cffi=1.15.1=py311h409f033_3
  - cftime=1.6.2=py311h4c7f6c3_1
  - charset-normalizer=3.1.0=pyhd8ed1ab_0
  - click=8.1.3=unix_pyhd8ed1ab_2
  - cloudpickle=2.2.1=pyhd8ed1ab_0
  - contourpy=1.0.7=py311ha3edf6b_0
  - cryptography=40.0.2=py311h9b4c7bb_0
  - curl=8.0.1=h588be90_0
  - cycler=0.11.0=pyhd8ed1ab_0
  - dask-core=2023.4.1=pyhd8ed1ab_0
  - expat=2.5.0=hcb278e6_1
  - fonttools=4.39.3=py311h2582759_0
  - freetype=2.12.1=hca18f0e_1
  - fsspec=2023.4.0=pyh1a96a4e_0
  - geos=3.11.2=hcb278e6_0
  - hdf4=4.2.15=h9772cbc_5
  - hdf5=1.12.2=nompi_h4df4325_101
  - icu=72.1=hcb278e6_0
  - idna=3.4=pyhd8ed1ab_0
  - importlib-metadata=6.6.0=pyha770c72_0
  - importlib_metadata=6.6.0=hd8ed1ab_0
  - iris=3.5.0=pyhd8ed1ab_0
  - jinja2=3.1.2=pyhd8ed1ab_1
  - jpeg=9e=h0b41bf4_3
  - keyutils=1.6.1=h166bdaf_0
  - kiwisolver=1.4.4=py311h4dd048b_1
  - krb5=1.20.1=h81ceb04_0
  - lcms2=2.15=hfd0df8a_0
  - ld_impl_linux-64=2.40=h41732ed_0
  - lerc=4.0.0=h27087fc_0
  - libaec=1.0.6=hcb278e6_1
  - libblas=3.9.0=16_linux64_openblas
  - libbrotlicommon=1.0.9=h166bdaf_8
  - libbrotlidec=1.0.9=h166bdaf_8
  - libbrotlienc=1.0.9=h166bdaf_8
  - libcblas=3.9.0=16_linux64_openblas
  - libcurl=8.0.1=h588be90_0
  - libdeflate=1.17=h0b41bf4_0
  - libedit=3.1.20191231=he28a2e2_2
  - libev=4.33=h516909a_1
  - libexpat=2.5.0=hcb278e6_1
  - libffi=3.4.2=h7f98852_5
  - libgcc-ng=12.2.0=h65d4601_19
  - libgfortran-ng=12.2.0=h69a702a_19
  - libgfortran5=12.2.0=h337968e_19
  - libgomp=12.2.0=h65d4601_19
  - libiconv=1.17=h166bdaf_0
  - liblapack=3.9.0=16_linux64_openblas
  - libnetcdf=4.8.1=nompi_h261ec11_106
  - libnghttp2=1.52.0=h61bc06f_0
  - libnsl=2.0.0=h7f98852_0
  - libopenblas=0.3.21=pthreads_h78a6416_3
  - libpng=1.6.39=h753d276_0
  - libsqlite=3.40.0=h753d276_1
  - libssh2=1.10.0=hf14f497_3
  - libstdcxx-ng=12.2.0=h46fd767_19
  - libtiff=4.5.0=h6adf6a1_2
  - libuuid=2.38.1=h0b41bf4_0
  - libwebp-base=1.3.0=h0b41bf4_0
  - libxcb=1.13=h7f98852_1004
  - libxml2=2.10.4=hfdac1af_0
  - libzip=1.9.2=hc929e4a_1
  - libzlib=1.2.13=h166bdaf_4
  - locket=1.0.0=pyhd8ed1ab_0
  - markupsafe=2.1.2=py311h2582759_0
  - matplotlib-base=3.7.1=py311h8597a09_0
  - munkres=1.1.4=pyh9f0ad1d_0
  - ncurses=6.3=h27087fc_1
  - netcdf4=1.6.2=nompi_py311hc6fcf29_100
  - numpy=1.24.3=py311h64a7726_0
  - openjpeg=2.5.0=hfec8fc6_2
  - openssl=3.1.0=hd590300_2
  - packaging=23.1=pyhd8ed1ab_0
  - partd=1.4.0=pyhd8ed1ab_0
  - pillow=9.4.0=py311h50def17_1
  - pip=23.1.2=pyhd8ed1ab_0
  - platformdirs=3.5.0=pyhd8ed1ab_0
  - pooch=1.7.0=pyha770c72_3
  - proj=9.2.0=h8ffa02c_0
  - pthread-stubs=0.4=h36c2ea0_1001
  - pycparser=2.21=pyhd8ed1ab_0
  - pyopenssl=23.1.1=pyhd8ed1ab_0
  - pyparsing=3.0.9=pyhd8ed1ab_0
  - pyproj=3.5.0=py311h1850bce_1
  - pyshp=2.3.1=pyhd8ed1ab_0
  - pysocks=1.7.1=pyha2e5f31_6
  - python=3.11.3=h2755cc3_0_cpython
  - python-dateutil=2.8.2=pyhd8ed1ab_0
  - python-xxhash=3.2.0=py311h2582759_0
  - python_abi=3.11=3_cp311
  - pyyaml=6.0=py311hd4cff14_5
  - readline=8.2=h8228510_1
  - requests=2.29.0=pyhd8ed1ab_0
  - scipy=1.10.1=py311h8e6699e_0
  - setuptools=67.7.2=pyhd8ed1ab_0
  - shapely=2.0.1=py311h54d622a_1
  - six=1.16.0=pyh6c4a22f_0
  - sqlite=3.40.0=h4ff8645_1
  - tk=8.6.12=h27826a3_0
  - toolz=0.12.0=pyhd8ed1ab_0
  - typing-extensions=4.5.0=hd8ed1ab_0
  - typing_extensions=4.5.0=pyha770c72_0
  - tzdata=2023c=h71feb2d_0
  - udunits2=2.2.28=hc3e0081_0
  - urllib3=1.26.15=pyhd8ed1ab_0
  - wheel=0.40.0=pyhd8ed1ab_0
  - xorg-libxau=1.0.9=h7f98852_0
  - xorg-libxdmcp=1.1.3=h7f98852_0
  - xxhash=0.8.1=h0b41bf4_0
  - xz=5.2.6=h166bdaf_0
  - yaml=0.2.5=h7f98852_2
  - zipp=3.15.0=pyhd8ed1ab_0
  - zlib=1.2.13=h166bdaf_4
  - zstd=1.5.2=h3eb15da_6

Additional context

Click to expand this section...

Please add additional verbose information in this section e.g., code, output, tracebacks, screenshots etc

[trexfeathers]

@ESadek-MO do you have the code that you used to test laziness on this function?

[ESadek-MO]

Unfortunately, due to a lack of foresight, I do not. There's every chance some of these feel through the gap and were incorrectly documented.
I'm currently off, but will try and replicate the test when I return.

[rcomer]

This function is properly old, and I think might even predate numpy.broadcast_to. It definitely predates Iris' use of dask. I suggest the best thing here would be to simplify the code to use both numpy.broadcast_to and dask.array.broadcast_to.