Merge pull request #530 from crocs-muni/fix/try-spacyHEADmain

Move to uv

4 files changed, 16 insertions, 11 deletions

diff --git a/src/sec_certs/cli.py b/src/sec_certs/cli.py
index 6276aa2f..fb5bf802 100644
--- a/src/sec_certs/cli.py
+++ b/src/sec_certs/cli.py
@@ -216,7 +216,7 @@ def main(
             processing_step.run(dset)
 
         end = datetime.now()
-        logger.info(f"The computation took {(end-start)} seconds.")
+        logger.info(f"The computation took {(end - start)} seconds.")
     except Exception as e:
         click.echo(
             f"Unhandled exception: {e}",
diff --git a/src/sec_certs/configuration.py b/src/sec_certs/configuration.py
index 17db9318..c06f872a 100644
--- a/src/sec_certs/configuration.py
+++ b/src/sec_certs/configuration.py
@@ -2,7 +2,7 @@ from __future__ import annotations
 
 import json
 from pathlib import Path
-from typing import Literal, Optional
+from typing import Literal
 
 import yaml
 from pydantic import AnyHttpUrl, Field
@@ -121,7 +121,7 @@ class Configuration(BaseSettings):
         True,
         description="During keyword search, first page usually contains addresses - ignore it.",
     )
-    cc_reference_annotator_dir: Optional[Path] = Field(  # noqa: UP007
+    cc_reference_annotator_dir: Path | None = Field(  # noqa: UP007
         None,
         description="Path to directory with serialized reference annotator model. If set to `null`, tool will search default directory for the given dataset.",
     )
@@ -141,7 +141,7 @@ class Configuration(BaseSettings):
         True,
         description="If true, progress bars will be printed to stdout during computation.",
     )
-    nvd_api_key: Optional[str] = Field(None, description="NVD API key for access to CVEs and CPEs.")  # noqa: UP007
+    nvd_api_key: str | None = Field(None, description="NVD API key for access to CVEs and CPEs.")  # noqa: UP007
     preferred_source_remote_datasets: Literal["sec-certs", "origin"] = Field(
         "sec-certs",
         description="If set to `sec-certs`, will fetch remote datasets from sec-certs.org."
diff --git a/src/sec_certs/model/references_nlp/__init__.py b/src/sec_certs/model/references_nlp/__init__.py
index d4e11e4d..8325c989 100644
--- a/src/sec_certs/model/references_nlp/__init__.py
+++ b/src/sec_certs/model/references_nlp/__init__.py
@@ -9,5 +9,5 @@ try:
 except ImportError as e:
     print(e)
     print(
-        "Requirements for ML annotation of references not met. Please run `pip install sec-certs[nlp]` or install `pip install -r requirements/nlp_requirements.txt."
+        "Requirements for ML annotation of references not met. Please install the 'nlp' extra, for example via: `pip install sec-certs[nlp]`."
     )
diff --git a/src/sec_certs/utils/extract.py b/src/sec_certs/utils/extract.py
index 35361e94..f72fd837 100644
--- a/src/sec_certs/utils/extract.py
+++ b/src/sec_certs/utils/extract.py
@@ -6,7 +6,7 @@ import re
 from collections import Counter
 from enum import Enum
 from pathlib import Path
-from typing import Any
+from typing import Any, no_type_check
 
 import numpy as np
 
@@ -17,7 +17,8 @@ from sec_certs.constants import FILE_ERRORS_STRATEGY, LINE_SEPARATOR, MAX_ALLOWE
 logger = logging.getLogger(__name__)
 
 
-def search_only_headers_anssi(filepath: Path):  # noqa: C901
+@no_type_check
+def search_only_headers_anssi(filepath: Path):  # type: ignore # noqa: C901
     # TODO: Please, refactor me. I reallyyyyyyyyyyyyy need it!!!!!!
     class HEADER_TYPE(Enum):
         HEADER_FULL = 1
@@ -272,7 +273,8 @@ def search_only_headers_anssi(filepath: Path):  # noqa: C901
     return items_found
 
 
-def search_only_headers_bsi(filepath: Path):  # noqa: C901
+@no_type_check
+def search_only_headers_bsi(filepath: Path):  # type: ignore # noqa: C901
     # TODO: Please, refactor me. I reallyyyyyyyyyyyyy need it!!!!!!
     LINE_SEPARATOR_STRICT = " "
     NUM_LINES_TO_INVESTIGATE = 15
@@ -371,7 +373,8 @@ def search_only_headers_bsi(filepath: Path):  # noqa: C901
     return items_found
 
 
-def search_only_headers_nscib(filepath: Path):  # noqa: C901
+@no_type_check
+def search_only_headers_nscib(filepath: Path):  # type: ignore # noqa: C901
     # TODO: Please, refactor me. I reallyyyyyyyyyyyyy need it!!!!!!
     LINE_SEPARATOR_STRICT = " "
     NUM_LINES_TO_INVESTIGATE = 60
@@ -451,7 +454,8 @@ def search_only_headers_nscib(filepath: Path):  # noqa: C901
     return items_found
 
 
-def search_only_headers_niap(filepath: Path):
+@no_type_check
+def search_only_headers_niap(filepath: Path):  # type: ignore # noqa: C901
     # TODO: Please, refactor me. I reallyyyyyyyyyyyyy need it!!!!!!
     LINE_SEPARATOR_STRICT = " "
     NUM_LINES_TO_INVESTIGATE = 15
@@ -502,7 +506,8 @@ def search_only_headers_niap(filepath: Path):
     return items_found
 
 
-def search_only_headers_canada(filepath: Path):  # noqa: C901
+@no_type_check
+def search_only_headers_canada(filepath: Path):  # type: ignore # noqa: C901
     # TODO: Please, refactor me. I reallyyyyyyyyyyyyy need it!!!!!!
     LINE_SEPARATOR_STRICT = " "
     NUM_LINES_TO_INVESTIGATE = 20