Import spacy model as module

app/gui.py

@@ -8,7 +8,7 @@ import gradio as gr
 import joblib
 
 if TYPE_CHECKING:
-    from sklearn.pipeline import Pipeline
+    from sklearn.base import BaseEstimator
 
 __all__ = ["launch_gui"]
 
@@ -19,7 +19,7 @@ NEGATIVE_LABEL = "Negative 😤"
 
 
 @lru_cache(maxsize=1)
-def load_model() -> Pipeline:
+def load_model() -> BaseEstimator:
     """Load the trained model and cache it."""
     model_path = os.environ.get("MODEL_PATH", None)
     if model_path is None:

app/model.py

@@ -2,8 +2,8 @@ from __future__ import annotations
 
 import warnings
 
+import en_core_web_sm
 import numpy as np
-import spacy
 from joblib import Memory
 from sklearn.base import BaseEstimator, TransformerMixin
 from sklearn.feature_extraction.text import TfidfVectorizer
@@ -16,7 +16,7 @@ from app.constants import CACHE_DIR
 
 __all__ = ["create_model", "train_model", "evaluate_model"]
 
-nlp = spacy.load("en_core_web_sm", disable=["tok2vec", "parser", "ner"])
+nlp = en_core_web_sm.load(disable=["tok2vec", "parser", "ner"])
 
 
 class TextTokenizer(BaseEstimator, TransformerMixin):