Merge pull request #13 from nestauk/tweaks-package

Tweaks to package

nlp_link/__init__.py

@@ -0,0 +1,19 @@
+import yaml
+from pathlib import Path
+from typing import Optional
+import logging
+
+
+def get_yaml_config(file_path: Path) -> Optional[dict]:
+    """Fetch yaml config and return as dict if it exists."""
+    if file_path.exists():
+        with open(file_path, "rt") as f:
+            return yaml.load(f.read(), Loader=yaml.FullLoader)
+
+
+# Define project base directory
+PROJECT_DIR = Path(__file__).resolve().parents[1]
+
+logger = logging.getLogger(__name__)
+
+soc_mapper_config = get_yaml_config(PROJECT_DIR / "nlp_link/soc_mapper/config.yaml")

nlp_link/soc_mapper/soc_map.py

@@ -21,6 +21,8 @@
 from tqdm import tqdm
 import numpy as np
 
+from nlp_link import soc_mapper_config
+
 from nlp_link.soc_mapper.soc_map_utils import (
     load_job_title_soc,
     process_job_title_soc,
@@ -31,8 +33,7 @@
 
 from nlp_link.linker_utils import load_bert
 
-from utils.utils import (
-    soc_mapper_config,
+from nlp_link.utils.utils import (
     load_s3_json,
     load_local_json,
     save_to_s3,

nlp_link/soc_mapper/soc_map_utils.py

@@ -2,7 +2,7 @@
 
 import re
 
-from utils.utils import soc_mapper_config
+from nlp_link import soc_mapper_config
 
 
 def load_job_title_soc(soc_mapper_config: dict = soc_mapper_config) -> pd.DataFrame():

utils/utils.py → nlp_link/utils/utils.py

@@ -1,27 +1,10 @@
-import yaml
-from pathlib import Path
-from typing import Optional
-import logging
 import boto3
 import json
 from fnmatch import fnmatch
 from decimal import Decimal
 import numpy
 
-
-def get_yaml_config(file_path: Path) -> Optional[dict]:
-    """Fetch yaml config and return as dict if it exists."""
-    if file_path.exists():
-        with open(file_path, "rt") as f:
-            return yaml.load(f.read(), Loader=yaml.FullLoader)
-
-
-# Define project base directory
-PROJECT_DIR = Path(__file__).resolve().parents[1]
-
-logger = logging.getLogger(__name__)
-
-soc_mapper_config = get_yaml_config(PROJECT_DIR / "nlp_link/soc_mapper/config.yaml")
+from nlp_link import logger
 
 
 def get_s3_resource():

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "nlp-link"
-version = "0.1.2"
+version = "0.1.3"
 description = "A python package to semantically link two lists of texts."
 authors = ["Nesta <[email protected]>"]
 readme = "README.md"
@@ -15,10 +15,11 @@ torch = "^1.13.1"
 pytest = "^8.3.2"
 tqdm = "^4.66.4"
 numpy = "^1.26.4"
-s3fs = "^2022.5.0"
 openpyxl = "^3.1.3"
-boto3 = "^1.21.21"
 wasabi = "^1.1.3"
+s3fs = {extras = ["boto3"], version = ">=2023.12.0"}
+boto3 = "*"
+botocore = "*"
 
 [build-system]
 requires = ["poetry-core"]