[Bug]: Biomedical NEN

[darije]

Describe the bug

When running example from Biomedical NEN (https://github.com/flairNLP/flair/blob/master/resources/docs/HUNFLAIR2.md) I get AttributeError: 'csr_matrix' object has no attribute 'A'

To Reproduce

from flair.data import Sentence
from flair.models import EntityMentionLinker
from flair.nn import Classifier

# make a sentence
sentence = Sentence("Behavioral abnormalities in the Fmr1 KO2 Mouse Model of Fragile X Syndrome")

# load biomedical NER tagger + predict entities
tagger = Classifier.load("hunflair2")
tagger.predict(sentence)

# load gene linker and perform normalization
gene_linker = EntityMentionLinker.load("gene-linker")
gene_linker.predict(sentence)

# load disease linker and perform normalization
disease_linker = EntityMentionLinker.load("disease-linker")
disease_linker.predict(sentence)

# load species linker and perform normalization
species_linker = EntityMentionLinker.load("species-linker")
species_linker.predict(sentence)

for entity in sentence.get_labels("link"):
    print(entity)

Expected behavior

Span[0:2]: "Behavioral abnormalities" → MESH:D001523/name=Mental Disorders (197.9467010498047)
Span[4:5]: "Fmr1" → 108684022/name=FRAXA (219.9510040283203)
Span[6:7]: "Mouse" → 10090/name=Mus musculus (213.6201934814453)
Span[9:12]: "Fragile X Syndrome" → MESH:D005600/name=Fragile X Syndrome (193.7115020751953)

Logs and Stack traces

Traceback (most recent call last):
  File "c:\Users\darij\Desktop\FlairNLP\biomedNEN.py", line 14, in <module>
    gene_linker.predict(sentence)
  File "C:\Users\darij\Desktop\FlairNLP\.venv\Lib\site-packages\flair\models\entity_mention_linking.py", line 822, in predict
    candidates = self.candidate_generator.search(entity_mentions=mentions[i : i + batch_size], top_k=top_k)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\darij\Desktop\FlairNLP\.venv\Lib\site-packages\flair\models\entity_mention_linking.py", line 614, in search
    mention_embs = self.embed(entity_mentions)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\darij\Desktop\FlairNLP\.venv\Lib\site-packages\flair\models\entity_mention_linking.py", line 594, in embed
    self.embeddings["sparse"].embed(inputs)
  File "C:\Users\darij\Desktop\FlairNLP\.venv\Lib\site-packages\flair\embeddings\document.py", line 213, in embed
    tfidf_vectors = torch.from_numpy(self.vectorizer.transform(raw_sentences).A)
                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'csr_matrix' object has no attribute 'A'

Screenshots

No response

Additional Context

Windows 11, VS Code, Python venv 3.12

Environment

Versions:

Flair

0.15.0

Pytorch

2.5.1+cu124

Transformers

4.48.1

GPU

True

[alanakbik]

@WangXII can you take a look?

[WangXII]

Pinging @sg-wbi , our expert on the NEN module

[sg-wbi]

The problem is with the installation of dependencies:

scikit-learn>=1.0.2 pulls scipy-1.15.1

Collecting scikit-learn>=1.0.2 (from flair==0.15.0)
  Using cached scikit_learn-1.6.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (18 kB)
Collecting scipy>=1.6.0 (from scikit-learn>=1.0.2->flair==0.15.0)
  Using cached scipy-1.15.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (61 kB)

but scipy-1.14.0 removed the .A attribute.

The simple fix is to change .A to .toarray: PR here.

Or you can try downgrading scipy to 1.13.

[sg-wbi]

I don't think scipy is used much in the library but there's a conflict in the dependencies which prevents seeing scipy updates in the tests.

In testing gensim>=4.2.0 is installed, which removes scipy-1.15.1 and installs scipy-1.13.1 (for whatever reason).

[darije]

@sg-wbi "The simple fix is to change .A to .toarray" - this worked!
Thank you!