Added static file folder (#41)

* added static file folder

* added constants file

scripts/label_all_tokens.py

@@ -1,10 +1,10 @@
 import argparse
 import pickle
-from pathlib import Path
 
 from tqdm.auto import tqdm
 from transformers import AutoTokenizer, PreTrainedTokenizer, PreTrainedTokenizerFast
 
+from delphi.constants import STATIC_ASSETS_DIR
 from delphi.eval import token_labelling
 
 
@@ -35,14 +35,13 @@ def main():
     args = parser.parse_args()
 
     # Access command-line arguments
-    # Directory to save the results
-    SAVE_DIR = Path("src/delphi/eval/")
+
     model_name = args.model_name
 
     print("\n", " LABEL ALL TOKENS ".center(50, "="), "\n")
     print(f"You chose the model: {model_name}\n")
     print(
-        f"The language model will be loaded from Huggingface and its tokenizer used to do two things:\n\t1) Create a list of all tokens in the tokenizer's vocabulary.\n\t2) Label each token with its part of speech, dependency, and named entity recognition tags.\nThe respective results will be saved to files located at: '{SAVE_DIR}'\n"
+        f"The language model will be loaded from Huggingface and its tokenizer used to do two things:\n\t1) Create a list of all tokens in the tokenizer's vocabulary.\n\t2) Label each token with its part of speech, dependency, and named entity recognition tags.\nThe respective results will be saved to files located at: '{STATIC_ASSETS_DIR}'\n"
     )
 
     # ================ (1) =================
@@ -60,8 +59,8 @@ def main():
 
     # Save the list of all tokens to a file
     filename = "all_tokens_list.txt"
-    filepath = SAVE_DIR / filename
-    with open(filepath, "w", encoding="utf-8") as f:
+    filepath = STATIC_ASSETS_DIR.joinpath(filename)
+    with open(f"{filepath}", "w", encoding="utf-8") as f:
         f.write(tokens_str)
 
     print(f"Saved the list of all tokens to:\n\t{filepath}\n")
@@ -88,16 +87,16 @@ def main():
 
     # Save the labelled tokens to a file
     filename = "labelled_token_ids_dict.pkl"
-    filepath = SAVE_DIR / filename
-    with open(filepath, "wb") as f:
+    filepath = STATIC_ASSETS_DIR.joinpath(filename)
+    with open(f"{filepath}", "wb") as f:
         pickle.dump(labelled_token_ids_dict, f)
 
     print(f"Saved the labelled tokens to:\n\t{filepath}\n")
 
     # sanity check that The pickled and the original dict are the same
     print("Sanity check ...", end="")
     # load pickle
-    with open(filepath, "rb") as f:
+    with open(f"{filepath}", "rb") as f:
         pickled = pickle.load(f)
     # compare
     assert labelled_token_ids_dict == pickled

scripts/map_tokens.py

@@ -3,6 +3,7 @@
 import argparse
 import pickle
 
+from delphi.constants import STATIC_ASSETS_DIR
 from delphi.eval.token_map import token_map
 from delphi.eval.utils import load_validation_dataset
 
@@ -18,5 +19,5 @@
 
     mapping = token_map(dataset)
 
-    with open(f"data/{args.output}", "wb") as f:
+    with open(f"{STATIC_ASSETS_DIR}/{args.output}", "wb") as f:
         pickle.dump(mapping, file=f)

setup.py

@@ -5,4 +5,5 @@
     version="0.1",
     packages=find_packages(where="src"),
     package_dir={"": "src"},
+    package_data={"delphi.static": ["*"]},
 )

src/delphi/constants.py

@@ -0,0 +1,3 @@
+from importlib.resources import files
+
+STATIC_ASSETS_DIR = files("delphi.static")