Create mask_functions.py

legalnlp/mask_functions.py

@@ -0,0 +1,161 @@
+import re
+
+def mask_email(txt):
+
+    """
+        Finds an email pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the email pattern
+        Returns
+        -----------
+        str
+            masked email string as ' [email] '
+        list
+            list with the found pattern(s)
+    """
+
+    
+    pattern=r'[^\s]+@[^\s]+'
+    sub=' [email] '
+
+    return re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+
+def mask_url(txt):
+
+    """
+        Finds an url pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the url pattern
+        Returns
+        -----------
+        str
+            masked url string as ' [url] '
+        list
+            list with the found pattern(s)
+    """
+        
+    pattern='http\S+'
+    pattern2='www\S+'
+    sub=' [url] '
+
+    txt, find = re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+    txt, find2 = re.sub(pattern2, sub, txt, flags=re.I), re.findall(pattern2, txt, flags=re.I)
+
+    return txt, find+find2
+
+def mask_oab(txt):
+
+    """
+        Finds an OAB (which stands for Order of Attorneys of Brazil) pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the OAB pattern
+        Returns
+        -----------
+        str
+            masked OAB string as ' [oab] '
+        list
+            list with the found pattern(s)
+    """
+    
+    find=[]
+    pattern='OAB\s?[:-]?\s?\d+\s?/?\s?[A-Z]?[A-Z]?'
+    pattern2='OAB\s?/?\s?[A-Z]?[A-Z]?\s?[:-]?\s?\d+'
+    sub=' [oab] '
+
+    txt, find = re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+    txt, find2 = re.sub(pattern2, sub, txt, flags=re.I), re.findall(pattern2, txt, flags=re.I)
+
+    return txt, find+find2
+
+def mask_data(txt):
+
+    """
+        Finds a date-format pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the date
+        Returns
+        -----------
+        str
+            masked date string as ' [data] '
+        list
+            list with the found pattern(s)
+    """
+
+    
+    pattern="\d{2}\s?\/\s?\d{2}\s?\/\s?\d{4}"
+    sub=' [data] '
+
+    return re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+
+def mask_processo(txt, num=15):
+
+    """
+        Finds a lawsuit number pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the lawsuit number pattern
+        Returns
+        -----------
+        str
+            masked lawsuit number string as ' [processo] '
+        list
+            list with the found pattern(s)
+    """
+        
+    pattern="\d{"+str(num)+",}" #consideramos números com mais de 15 dígitos como sendo o número de um processo
+    sub=' [processo] '
+
+    return re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+
+def mask_numero(txt):
+
+    """
+        Finds a number pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the number pattern
+        Returns
+        -----------
+        str
+            masked number string as ' [numero] '
+        list
+            list with the found pattern(s)
+            
+    """
+        
+    pattern="\d+"
+    sub=' [numero] '
+
+    return re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)
+
+def mask_valor(txt):
+
+    """
+        Finds a value pattern and then masks it.
+        Parameters
+        -----------
+        txt:  str
+            A piece of text containing the value pattern
+        Returns
+        -----------
+        str
+            masked value string as ' [valor] '
+        list
+            list with the found pattern(s)
+    """
+        
+
+    pattern="R\s?\$\s?\d+[.,]?\d+[.,]?\d+"
+    sub=' [valor] '
+
+    return re.sub(pattern, sub, txt, flags=re.I), re.findall(pattern, txt, flags=re.I)