From 7682f3036c784d6eabeda0407eab4f2a475a1b09 Mon Sep 17 00:00:00 2001
From: Andras Schmelczer <andras@schmelczer.dev>
Date: Sat, 25 Jun 2022 11:05:21 +0200
Subject: [PATCH] Add truecasing

---
 great_ai/src/great_ai/utilities/get_sentences.py | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/great_ai/src/great_ai/utilities/get_sentences.py b/great_ai/src/great_ai/utilities/get_sentences.py
index 87bb172..de0ad25 100644
--- a/great_ai/src/great_ai/utilities/get_sentences.py
+++ b/great_ai/src/great_ai/utilities/get_sentences.py
@@ -1,15 +1,19 @@
-from typing import List
+from typing import List, cast
 
 from segtok.segmenter import split_multi
 
 from .data import sentence_ending_punctuations
 
 
-def get_sentences(text: str, ignore_partial: bool = False) -> List[str]:
+def get_sentences(
+    text: str, ignore_partial: bool = False, true_case: bool = False
+) -> List[str]:
     if text.strip() == "":
         return []
 
-    possible_sentences = [s for s in split_multi(text) if s]
+    possible_sentences = [
+        cast(str, s).strip() for s in split_multi(text) if cast(str, s).strip()
+    ]
 
     if ignore_partial:
         possible_sentences = [
@@ -18,4 +22,9 @@ def get_sentences(text: str, ignore_partial: bool = False) -> List[str]:
             if s[0].isupper() and s[-1] in sentence_ending_punctuations
         ]
 
+    if true_case:
+        possible_sentences = [
+            s[0].lower() + s[1:] for s in possible_sentences  # very crude method
+        ]
+
     return possible_sentences