Update en_core_web_sm version (#64)

* update en_core_web_sm version

* set en-core-web-sm version in pyproject.toml

* Update module_textdeid.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* raise error if en_core_web_sm_model not found

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: qsodia

Makefile

@@ -24,7 +24,6 @@ deps: deps-python deps-deid deps-electron
 deps-python:
 	uv --version || (echo "uv is not installed. Please install uv and try again." && exit 1)
 	uv sync
-	uv run python -m spacy download en_core_web_sm
 
 deps-deid:
 	cd deid && npm install

module_textdeid.py

@@ -51,7 +51,17 @@
 def create_nlp_engine():
     if getattr(sys, 'frozen', False):
         bundle_dir = os.path.abspath(os.path.dirname(sys.executable))
-        model_path = os.path.join(bundle_dir, '_internal', 'en_core_web_sm', 'en_core_web_sm-3.7.1')
+        models_base_dir = os.path.join(bundle_dir, '_internal', 'en_core_web_sm')
+        model_subdir = None
+        if os.path.isdir(models_base_dir):
+            for entry in os.listdir(models_base_dir):
+                full_path = os.path.join(models_base_dir, entry)
+                if entry.startswith('en_core_web_sm-') and os.path.isdir(full_path):
+                    model_subdir = full_path
+                    break
+        if model_subdir is None:
+            raise FileNotFoundError(f"en_core_web_sm model not found in {models_base_dir}")
+        model_path = model_subdir
         import spacy
         from presidio_analyzer.nlp_engine import SpacyNlpEngine
         nlp_engine = SpacyNlpEngine(models=[{"lang_code": "en", "model_name": model_path}])

pyproject.toml

@@ -36,9 +36,11 @@ dependencies = [
     "ruamel-yaml-clib>=0.2.15",
     "six>=1.17.0",
     "spacy>=3.8.11",
+    "en-core-web-sm @ https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl",
     "tenacity>=9.1.4",
     "typing-extensions>=4.15.0",
     "tzdata>=2025.3",
     "tzlocal>=5.3.1",
     "zipp>=3.23.0",
 ]
+