chore: fix $ and €

anonymize_file_clear.py

@@ -23,7 +23,8 @@ def load_models():
 
 
 def anonymize_with_cml(text, embeddings_model, tokenizer, cml_ner_model):
-    token_pattern = r"(\b[\w\.\/\-@]+\b|[\s,.!?;:'\"-]+)"
+    token_pattern = r"(\b[\w\.\/\-@]+\b|[\s,.!?;:'\"-]+|\$\d+(?:\.\d+)?|\€\d+(?:\.\d+)?)"
+
     tokens = re.findall(token_pattern, text)
     uuid_map = {}
     processed_tokens = []

files/anonymized_document.txt

@@ -1,11 +1 @@
-Contract parties: 7f2310e8 and 7166a872
-
-Date: c0239016 4f1ec317, e1bfa3a5
-
-Scope: 7166a872 agrees to provide graphic design services to 7f2310e8 for the creation of a company logo.
-
-Amount: 7f2310e8 agrees to pay 7166a872 500 upon completion and delivery of the logo.
-
-Deadline: The logo design must be completed and delivered to 7f2310e8 within 14 days of the contract signing date.
-
-Payment terms: 7166a872's international bank account N: df667faf
+14 days and 2c68f2f2

files/mapping_clear_to_anonymized.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e7916c386aa21445d37ad828a7d0dea44d4669ffd29da3203c51aedb7993ee29
-size 938
+oid sha256:f05e1dd38586c17aed4200febefb058205bd510f9b4b7eba5724fbf05ad78d5d
+size 77

files/mapping_doc_embedding_path.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:165830bc31b69f2222f1f4f9273567e94698cceacdb9853efeae55969a0f61bd
-size 203038
+oid sha256:4b119b5428b6f0740ea7201428cc34ebdf9b893c99613afa40e58dc46586a3ed
+size 16674

files/original_document.txt

@@ -4,8 +4,8 @@ Date: February 06, 2000
 
 Scope: Kate agrees to provide graphic design services to David for the creation of a company logo.
 
-Amount: David agrees to pay Kate $500 upon completion and delivery of the logo.
+Amount: David agrees to pay Kate $1500 upon completion and delivery of the logo.
 
-Deadline: The logo design must be completed and delivered to David within 14 days of the contract signing date.
+Deadline: The logo design must be completed and delivered to David within 14-days of the contract signing date.
 
 Payment terms: Kate's international bank account N°: IL150120690000003111111

files/original_document_uuid_mapping.json

@@ -1,8 +1,3 @@
 {
-    "06": "4f1ec317",
-    "2000": "e1bfa3a5",
-    "David": "7f2310e8",
-    "February": "c0239016",
-    "IL150120690000003111111": "df667faf",
-    "Kate": "7166a872"
+    "$500": "2c68f2f2"
 }