GameSentenceMiner 2.15.6__py3-none-any.whl → 2.15.8__py3-none-any.whl

GameSentenceMiner/vad.py

@@ -189,9 +189,10 @@ class WhisperVADProcessor(VADProcessor):
                                                              temperature=0.0)
         voice_activity = []
 
-        logger.debug(json.dumps(result.to_dict(), indent=2))
+        logger.debug(json.dumps(result.to_dict()))
 
         # Process the segments to extract tokens, timestamps, and confidence
+        previous_segment = None
         for i, segment in enumerate(result.segments):
             if len(segment.text) <= 2 and ((i > 1 and segment.start - result.segments[i - 1].end > 1.0) or (i < len(result.segments) - 1 and result.segments[i + 1].start - segment.end > 1.0)):
                 if segment.text in ['えー', 'ん']:
@@ -201,10 +202,18 @@ class WhisperVADProcessor(VADProcessor):
                     logger.info(
                         "Unknown single character segment, not skipping, but logging, please report if this is a mistake: " + segment.text)
                     
-            if segment.no_speech_prob and segment.no_speech_prob > 0.8:
+            if segment.no_speech_prob and segment.no_speech_prob > 0.6:
                 logger.debug(f"Skipping segment with high no_speech_prob: {segment.no_speech_prob} for segment {segment.text} at {segment.start}-{segment.end}")
                 continue
-
+            
+            unique_words = set(word['text'] for word in segment.words)
+            if len(unique_words) <= 1 and len(segment.words) > 1:
+                logger.debug(f"Skipping segment with low unique words: {unique_words} for segment {segment.text} at {segment.start}-{segment.end}")
+                continue
+                
+            if previous_segment and segment.start - previous_segment.end > 5 and segment.no_speech_prob > .3:
+                logger.debug(f"Skipping segment after long pause with high no_speech_prob after: {segment.no_speech_prob} for segment {segment.text} at {segment.start}-{segment.end}")
+                continue
 
             logger.debug(segment.to_dict())
             voice_activity.append({
@@ -225,15 +234,7 @@ class WhisperVADProcessor(VADProcessor):
             #             'confidence': word.probability
             #         })
 
-        # Analyze the detected words to decide whether to use the audio
-        should_use = False
-        unique_words = set(word['text'] for word in voice_activity)
-        if len(unique_words) > 1 or not all(item in ['えー', 'ん'] for item in unique_words):
-            should_use = True
-
-        if not should_use:
-            return None
-
+            previous_segment = segment
         # Return the detected voice activity and the total duration
         return voice_activity
 

{gamesentenceminer-2.15.6.dist-info → gamesentenceminer-2.15.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GameSentenceMiner
-Version: 2.15.6
+Version: 2.15.8
 Summary: A tool for mining sentences from games. Update: Overlay?
 Author-email: Beangate <bpwhelan95@gmail.com>
 License: MIT License

{gamesentenceminer-2.15.6.dist-info → gamesentenceminer-2.15.8.dist-info}/RECORD

@@ -4,7 +4,7 @@ GameSentenceMiner/config_gui.py,sha256=i79PrY2pP8_VKvIL7uoDv5cgHvCCQBIe0mS_YnX2A
 GameSentenceMiner/gametext.py,sha256=fgBgLchezpauWELE9Y5G3kVCLfAneD0X4lJFoI3FYbs,10351
 GameSentenceMiner/gsm.py,sha256=t2GAhMwVEHUzCdqM4tIgAzBUvNmt_Gec515iePacD6k,31945
 GameSentenceMiner/obs.py,sha256=EyAYhaLvMjoeC-3j7fuvkqZN5logFFanPfb8Wn1C6m0,27296
-GameSentenceMiner/vad.py,sha256=VPYXFq0wnj1rzhZjFXDyTVQYHOSQaoap3tH-lMA-CTA,19821
+GameSentenceMiner/vad.py,sha256=BqwvKF0lRB_ys1On_Ua1vvimUFmkHO-D9gBbCkaBaxc,20193
 GameSentenceMiner/ai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 GameSentenceMiner/ai/ai_prompting.py,sha256=41xdBzE88Jlt12A0D-T_cMfLO5j6MSxfniOptpwNZm0,24068
 GameSentenceMiner/assets/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -72,9 +72,9 @@ GameSentenceMiner/web/templates/index.html,sha256=LqXZx7-NE42pXSpHNZ3To680rD-vt9
 GameSentenceMiner/web/templates/text_replacements.html,sha256=tV5c8mCaWSt_vKuUpbdbLAzXZ3ATZeDvQ9PnnAfqY0M,8598
 GameSentenceMiner/web/templates/utility.html,sha256=3flZinKNqUJ7pvrZk6xu__v67z44rXnaK7UTZ303R-8,16946
 GameSentenceMiner/wip/__init___.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-gamesentenceminer-2.15.6.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-gamesentenceminer-2.15.6.dist-info/METADATA,sha256=_UoXIP7jWBvF4lpRp2OLNOhGUpTAXG2cnUKXmbKepyw,7348
-gamesentenceminer-2.15.6.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-gamesentenceminer-2.15.6.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
-gamesentenceminer-2.15.6.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
-gamesentenceminer-2.15.6.dist-info/RECORD,,
+gamesentenceminer-2.15.8.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+gamesentenceminer-2.15.8.dist-info/METADATA,sha256=bXBIX71jQ-PLrLVxpjrANJmu_ET8bA0EjRbgbgBl-AM,7348
+gamesentenceminer-2.15.8.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+gamesentenceminer-2.15.8.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
+gamesentenceminer-2.15.8.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
+gamesentenceminer-2.15.8.dist-info/RECORD,,