From 2ae269d939ab2b25becddd37cc01eb547218a322 Mon Sep 17 00:00:00 2001 From: Malar Kannan Date: Thu, 28 Dec 2017 20:01:44 +0530 Subject: [PATCH] generating test for phone seg model --- speech_samplegen.py | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/speech_samplegen.py b/speech_samplegen.py index 2b6faa4..ee390dc 100644 --- a/speech_samplegen.py +++ b/speech_samplegen.py @@ -12,9 +12,9 @@ import time from tqdm import tqdm from generate_similar import similar_phoneme_phrase,similar_phrase -from speech_tools import hms_string,create_dir,format_filename +from speech_tools import hms_string,create_dir,format_filename,reservoir_sample -OUTPUT_NAME = 'story_phrases' +OUTPUT_NAME = 'test_5_words' dest_dir = os.path.abspath('.') + '/outputs/' + OUTPUT_NAME + '/' dest_file = './outputs/' + OUTPUT_NAME + '.csv' @@ -224,7 +224,7 @@ def generate_audio_for_stories(): text_list = sorted(list(set(text_list_dup))) generate_audio_for_text_list(text_list) -def generate_test_audio_for_stories(): +def generate_test_audio_for_stories(sample_count=0): story_file = './inputs/all_stories_hs.json' # story_file = './inputs/all_stories.json' stories_data = json.load(open(story_file)) @@ -234,11 +234,12 @@ def generate_test_audio_for_stories(): word_list = [i.strip('\n_') for i in open('./inputs/wordlist.txt','r').readlines()] text_set = set(text_list) new_word_list = [i for i in word_list if i not in text_set and len(i) > 4] - test_words = new_word_list[:int(len(text_list)/5+1)] + # test_words = new_word_list[:int(len(text_list)/5+1)] + test_words = reservoir_sample(new_word_list,sample_count) if sample_count > 0 else new_word_list generate_audio_for_text_list(test_words) if __name__ == '__main__': - # generate_test_audio_for_stories() + generate_test_audio_for_stories(5) # generate_audio_for_text_list(['I want to go home','education']) - generate_audio_for_stories() + # generate_audio_for_stories()