fix intersecting words logic (#60)

* fix intersecting words logic

* update version

Author: joeferraratonic

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "tonic-textual"
-version = "3.10.4"
+version = "3.10.5"
 description = "Wrappers around the Tonic Textual API"
 authors = ["Adam Kamor <[email protected]>", "Joe Ferrara <[email protected]>", "Ander Steele <[email protected]>", "Ethan Philpott <[email protected]>", "Lyon Van Voorhis <[email protected]>", "Kirill Medvedev <[email protected]>", "Travis Matthews <[email protected]>"]
 license = "MIT"

tonic_textual/__init__.py

@@ -1 +1 @@
-__version__ = "3.10.4"
+__version__ = "3.10.5"

tonic_textual/helpers/redact_audio_file_helper.py

@@ -119,16 +119,23 @@ def get_intervals_to_redact(
     )
     output_intervals = []
     for span in de_identify_results:
-        start = span.start
-        end = span.end
+        span_start = span.start
+        span_end = span.end
         intersecting_words: List[TranscriptionWord] = []
         for word_obj in enriched_transcript_words:
             word_start = word_obj.char_start
             word_end = word_obj.char_end
-            if start < word_end and word_start < end:
+            # beep a word if it overlaps with the found span
+            # this beeps entire word when span is part of a word
+            if word_start < span_end and word_start >= span_start:
                 intersecting_words.append(word_obj)
-            if word_start > end: # done
+            elif word_end >= span_start and word_end < span_end:
+                intersecting_words.append(word_obj)
+            elif word_start > span_end: # done
                 break
+        # if fail to find intersecting words continue
+        if len(intersecting_words) == 0:
+            continue
         # unecessary if transcript_words is sorted but cheap
         span_time_start = min([word_obj.start for word_obj in intersecting_words])
         span_time_end = max([word_obj.end for word_obj in intersecting_words])