Error using db.from_delayed with sparse arrays

[joshua-gould]

File "python3.11/site-packages/dask/bag/core.py", line 1881, in reify
    if len(seq) and isinstance(seq[0], Iterator):
      ^^^^^^^^^^^^^^^^^
  File "python3.11/site-packages/scipy/sparse/_base.py", line 425, in __len__
    raise TypeError("sparse array length is ambiguous; use getnnz()"

Minimal Complete Verifiable Example:

import dask.bag as db
import numpy as np
from dask import delayed
from scipy.sparse import csr_array


def add(x, y):
    return x + y


@delayed
def create_sparse_array_delayed():
    return csr_array(np.random.random((10, 10)))


@delayed
def create_array_delayed():
    return np.random.random((10, 10))


db.from_sequence(
    [csr_array(np.random.random((10, 10))), csr_array(np.random.random((10, 10)))]).fold(
    add).compute()  # works with sparse arrays when created from sequence
db.from_delayed([create_array_delayed(), create_array_delayed()]).fold(add).compute()  # works with numpy arrays
db.from_delayed([create_sparse_array_delayed(), create_sparse_array_delayed()]).fold(add).compute()  # fails

Environment:

• Dask version: 2024.12.0
• Python version: 3.11
• Operating System: Mac
• Install: pip

[phofl]

Thanks for your report. Any advice on how we can make this work without adding scipy as a dependency for bags?

[batcity]

FYI I've created a fix for this bug here: https://github.com/dask/dask/pull/12103

cc: @phofl @joshua-gould