Error in sc.tl.umap

[OnlyBelter]

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of scanpy.
• (optional) I have confirmed this bug exists on the master branch of scanpy.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Minimal code sample (that we can copy&paste without having any data)

I am learning the example of Integrating data using ingest and BBKNN. When I run code

sc.tl.umap(adata_ref)

I get

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-13-0e548e19df3a> in <module>
      1 sc.pp.pca(adata_ref)
      2 sc.pp.neighbors(adata_ref)
----> 3 sc.tl.umap(adata_ref)

~/miniconda3/envs/tf/lib/python3.6/site-packages/scanpy/tools/_umap.py in umap(adata, min_dist, spread, n_components, maxiter, alpha, gamma, negative_sample_rate, init_pos, random_state, a, b, copy, method, neighbors_key)
    171             neigh_params.get('metric', 'euclidean'),
    172             neigh_params.get('metric_kwds', {}),
--> 173             verbose=settings.verbosity > 3,
    174         )
    175     elif method == 'rapids':

TypeError: simplicial_set_embedding() missing 3 required positional arguments: 'densmap', 'densmap_kwds', and 'output_dens'

Versions

scanpy==1.6.0 anndata==0.7.5 umap==0.5.0 numpy==1.19.5 scipy==1.5.4 pandas==1.1.5 scikit-learn==0.24.0 statsmodels==0.12.1 python-igraph==0.8.3 leidenalg==0.8.3

[OnlyBelter]

Since umap-learn updated to version 0.5.0 from 0.4.6, the interface may have changed.

[scottgigante]

+1. Seems like pinning the version to <0.5 for now would solve the problem.

[scottgigante]

Looks like this was solved on master last month. Any chance we could get a bugfix release?

[ivirshup]

So we've just put out a release of 1.7.0rc1, and will be releasing it proper soon.
I'm looking at making a 1.6.1 release where the only change is pinning umap, but there are some CI issues (largely tests failing due to Matplotlib outputs changing slightly).

[ndrsfel]

Thank you for the release update! I just had to
pip install git+https://github.com/theislab/scanpy.git@1.7.0rc1

[ivirshup]

I've just pushed 1.6.1 (pinning umap), but you can get the release candidate for 1.7.0 with pip install "scanpy==1.7.0rc1"

[savitaj]

Has this error been fixed? Facing the same issue for scanpy version 1.7.2

[OnlyBelter]

Version 1.8.0 works well.

[brianpenghe]

I tried downgrading umap-learn to 0.4.6 but then sc.pp.neighbors won't work. I've been using scanpy 1.5.0

[brianpenghe]

OK I computed the neighbors using umap-learn 0.5.1 and then downgraded to 0.4.6 for UMAP. Not elegant but so far so good.

[brianpenghe]

Updating to Scanpy 1.8.2 worked!

[dajsfiles]

Sorry for opening this thread again, but I think I've run into the same problem. Here's my code and error:

mat_all = sc.read_loom(filename="RSV.loom")
sc.pp.pca(mat_all)
sc.pp.neighbors(mat_all)
sc.tl.umap(mat_all)

The error message:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
/tmp/31048.tmpdir/ipykernel_3245/2128514342.py in <module>
      3 sc.pp.pca(mat_all)
      4 sc.pp.neighbors(mat_all)
----> 5 sc.tl.umap(mat_all)
      6 sc.pl.tsne(mat_all, color="cluster",legend_loc="on data",
      7            size=20, save=True)

/storage1/fs1/leyao.wang/Active/conda/envs/velocyto3.9/lib/python3.9/site-packages/scanpy/tools/_umap.py in umap(adata, min_dist, spread, n_components, maxiter, alpha, gamma, negative_sample_rate, init_pos, random_state, a, b, copy, method, neighbors_key)
    192         default_epochs = 500 if neighbors['connectivities'].shape[0] <= 10000 else 200
    193         n_epochs = default_epochs if maxiter is None else maxiter
--> 194         X_umap = simplicial_set_embedding(
    195             X,
    196             neighbors['connectivities'].tocoo(),

TypeError: simplicial_set_embedding() missing 3 required positional arguments: 'densmap', 'densmap_kwds', and 'output_dens'

And the versions I've been running:
anndata 0.7.8
asttokens 2.0.5
bcrypt 3.2.0
Bottleneck 1.3.2
brotlipy 0.7.0
cached-property 1.5.2
certifi 2021.10.8
cffi 1.15.0
charset-normalizer 2.0.12
chart-studio 1.1.0
click 8.0.4
cmake 3.22.2
colorama 0.4.4
conda 4.11.0
conda-package-handling 1.7.3
cryptography 36.0.1
cycler 0.11.0
Cython 0.29.20
devtools 0.8.0
dunamai 1.9.0
executing 0.8.2
fa2 0.3.5
Fabric 1.6.1
fonttools 4.29.1
get_version 3.5.4
h5py 3.6.0
idna 3.3
igraph 0.9.9
install 1.3.5
joblib 1.1.0
kiwisolver 1.3.2
legacy-api-wrap 1.2
llvmlite 0.38.0
loom 0.0.18
loompy 3.0.6
mamba 0.15.3
matplotlib 3.5.1
mkl-fft 1.3.1
mkl-random 1.2.2
mkl-service 2.4.0
MulticoreTSNE 0.1
natsort 8.1.0
networkx 2.6.3
numba 0.55.1
numexpr 2.8.1
numpy 1.21.2
numpy-groupies 0.9.14
opt-einsum 3.3.0
packaging 21.3
pandas 1.4.1
paramiko 2.9.2
patsy 0.5.2
Pillow 9.0.1
pip 21.2.4
plotly 5.6.0
pycosat 0.6.3
pycparser 2.21
PyNaCl 1.5.0
pynndescent 0.5.6
pyOpenSSL 22.0.0
pyparsing 3.0.7
PyQt5 5.12.3
PyQt5_sip 4.19.18
PyQtChart 5.12
PyQtWebEngine 5.12.1
pyro-api 0.1.2
pyro-ppl 1.8.0
pysam 0.18.0
PySocks 1.7.1
python-dateutil 2.8.2
pytz 2021.3
requests 2.27.1
retrying 1.3.3
ruamel-yaml-conda 0.15.80
scanpy 1.7.0rc1
scikit-learn 1.0.2
scipy 1.7.3
seaborn 0.11.2
setuptools 58.0.4
sinfo 0.3.4
six 1.16.0
statsmodels 0.13.2
stdlib-list 0.8.0
tables 3.7.0
tenacity 8.0.1
texttable 1.6.4
threadpoolctl 3.1.0
torch 1.10.2
tornado 6.1
tqdm 4.62.3
umap-learn 0.4.6
unicodedata2 14.0.0
urllib3 1.26.8
velocyto 0.17.17
wheel 0.37.1
xlrd 1.2.0

If anyone can help me resolve this that would be great. Thanks!

[dajsfiles]

Just updated to scanpy 1.8.2, but the problem persists.

[YubinXie]

scanpy 1.9.1 has this issue too.

[tdefa]

udpating umap-learn work for me

pip install umap-learn==0.5.3