refractoring and more tests

multilang_translator/main.py

@@ -11,6 +11,9 @@ from examples import custom_style_2
 
 
 import os
+from copy import copy
+import time
+import logging as log
 from .translator import llm_translator
 from .text_to_speech import text_to_speech, resample
 from .backend_controller.broadcaster_config import broadcaster_config
@@ -20,60 +23,71 @@ from .backend_controller.broadcaster_copy_files import copy_to_broadcaster
 from .encode import encode_lc3
 
 ANNOUNCEMENT_DIR = os.path.join(os.path.dirname(__file__), 'announcements')
-N_MAX_BIS = 4
-SAMPLING_RATE = int(8e3)
+SAMPLING_RATE = int(16e3)
 FRAME_DUR_MS = 10
-BPS = int(16e3)
-FILENAMES = {
-    "de": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_de",
-    "en": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_en",
-    "fr": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_fr",
-    "es": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_es",
-    "it": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_it",
+BPS = int(32e3) # TODO test 16khz 16kbps
+CONFIG = {
+    "de": {
+        "file": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_de",
+        "tts": 'de_DE-kerstin-low',
+    },
+    "en": {
+        "file": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_en",
+        "tts": 'en_US-lessac-medium'
+    },
+    "fr": {
+        "file": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_fr",
+        "tts": 'fr_FR-siwis-medium'
+    },
+    "es": {
+        "file": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_es",
+        "tts": 'es_ES-sharvard-medium'
+    },
+    "it": {
+        "file": f"{ANNOUNCEMENT_DIR}/announcement_{SAMPLING_RATE//1000}_{FRAME_DUR_MS}_{BPS//1000}_it",
+        "tts": 'it_IT-paola-medium'
+    }
 }
 
 os.makedirs(ANNOUNCEMENT_DIR, exist_ok=True)
 
 def synthesize_resample_encode(text, tts_model, output_file):
-    text_to_speech.synthesize(text, tts_model, output_file)
+    audio_dur = text_to_speech.synthesize(text, tts_model, output_file)
     resample.resample(output_file, output_file, target_rate=SAMPLING_RATE)
     encode_lc3.encode_lc3(output_file, bps=BPS, frame_dur_ms=FRAME_DUR_MS)
+    return audio_dur
 
 def translate_from_german_and_encode(text_de):
-    file = FILENAMES['de']
-    synthesize_resample_encode(text_de, 'de_DE-kerstin-low', f'{file}.wav')
+    config = copy(CONFIG)
+    base_lang = "de"
 
-    text_en = llm_translator.translator_de_en(text_de)
-    file = FILENAMES['en']
-    synthesize_resample_encode(text_en, 'en_US-lessac-medium', f'{file}.wav')
-
-    text_fr = llm_translator.translator_de_fr(text_de)
-    file = FILENAMES['fr']
-    synthesize_resample_encode(text_fr, 'fr_FR-siwis-medium', f'{file}.wav')
-
-    text_es = llm_translator.translator_de_es(text_de)
-    file = FILENAMES['es']
-    synthesize_resample_encode(text_es, 'es_ES-sharvard-medium', f'{file}.wav')
-
-    text_it = llm_translator.translator_de_it(text_de)
-    file = FILENAMES['it']
-    synthesize_resample_encode(text_it, 'it_IT-paola-medium', f'{file}.wav')
+    file = config[base_lang]["file"]
+    audio_dur_s [base_lang] = synthesize_resample_encode(text_de, config['de']["tts"], f'{file}.wav')
+    
+    del config[base_lang]
 
+    audio_dur_s = {}
+    for key, val in config.items():
+        text = llm_translator.translate_de_to_x(key, text_de)
+        file = val['file']
+        audio_dur_s[key] = synthesize_resample_encode(text, val['tts'], f'{file}.wav')
+    return audio_dur_s
 
 def announcement_from_german_text(text_de):
 
-    translate_from_german_and_encode(text_de)
-
+    audio_durs = translate_from_german_and_encode(text_de)
     # Transfer the files to broadcaster memory
     start = time.time()
-    for val in FILENAMES.values():
-        copy_to_broadcaster(f'{val}.lc3')
+    for val in CONFIG.values():
+        copy_to_broadcaster(f'{val["file"]}.lc3')
     log.info("Transfering files to broadcaster took %s s", round(time.time() - start, 3))
 
     # Instruct the broadcaster to stream the files
-    for i, val in enumerate(list(FILENAMES.values())[:N_MAX_BIS]):
-        time.sleep(1)
-        broadcaster_play_file(i, f'{os.path.basename(val)}.lc3')
+    for i, d in enumerate(list(CONFIG.items())):
+        key, val = d
+        broadcaster_play_file(i, f'{os.path.basename(val["file"])}.lc3')
+        time.sleep(audio_durs[key] + 1)
+
     log.info("Starting all broadcasts %s s", round(time.time() - start, 3))
    
 
@@ -92,15 +106,3 @@ def announcement_from_german_text(text_de):
 
 # answers = prompt(questions, style=custom_style_2)
 # pprint(answers)
-
-if __name__ == '__main__':
-    import time
-    from translator import test_content
-    import logging as log
-    log.basicConfig(level=log.INFO)
-
-    start= time.time()
-    #broadcaster_config()
-
-    announcement_from_german_text(test_content.TESTSENTENCE_DE_RAINBOW)
-    print("Generating and starting the announcement took", time.time() - start)