File size: 3,440 Bytes
b8184f0 88c5e5a b8184f0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 |
import os
import random
import pandas as pd
import shutil
import uuid
# Provided file folder addresses (replace with actual paths)
common_dir = "vitsGPT/vits/"
folders = [
f"{common_dir}DUMMY5/gt_test_wav/",
f"{common_dir}ori_vits/logs/emovdb_base_pretrained16/G_100000/model_test_wav/",
f"{common_dir}emo_vits/logs/emovdb_emo_add_ave_pretrained16/G_100000/model_test_wav/",
f"{common_dir}emo_vits/logs/emovdb_emo_add_bert_cls_pretrained16/G_100000/model_test_wav/",
f"{common_dir}sem_vits/logs/emovdb_sem_mat_text_pretrained16/G_100000/model_test_wav/",
f"{common_dir}sem_vits/logs/emovdb_sem_mat_phone_pretrained16/G_100000/model_test_wav/",
f"{common_dir}sem_vits/logs/emovdb_sem_mat_bert_text_pretrained16/G_100000/model_test_wav/",
f"{common_dir}sem_vits/logs/emovdb_sem_mat_bert_phone_pretrained16/G_100000/model_test_wav/",
]
num_extracted_files = 50
# 0. Create human_mos directory if it doesn't exist
human_mos_dir = os.path.join(common_dir, "human_evaluation_emovdb/human_mos_wavs")
if not os.path.exists(human_mos_dir):
os.makedirs(human_mos_dir)
# 1. Check if every folder has the same number and names of files
reference_files = set(os.listdir(folders[0]))
consistent = all(set(os.listdir(folder)) == reference_files for folder in folders)
if not consistent:
raise ValueError("Folders do not have consistent file names or counts.")
# 2. Extract, rename, and copy files from each folder
sample_files = random.sample(list(reference_files), num_extracted_files)
# Generate a shuffled list of numbers for renaming
numbers = list(range(len(sample_files) * len(folders)))
random.shuffle(numbers)
renamed_files_paths = {}
for orig in sample_files:
for folder in folders:
# new_name = f"{len(renamed_files_paths)}.wav"
# Use UUID to generate a random file name
# new_name = f"{uuid.uuid4().hex}.wav"
# Use shuffled number for renaming
new_name = f"{numbers.pop()}.wav"
src = os.path.join(folder, orig)
dst = os.path.join(human_mos_dir, new_name)
shutil.copy(src, dst)
renamed_files_paths[src] = new_name
df_named_score = pd.DataFrame({
"original_file_path": list(renamed_files_paths.keys()),
"innominated_file_path": list(renamed_files_paths.values())
})
# 3. Create innominated_score.xlsx
df_innominated_score = pd.DataFrame({
"file_name": list(renamed_files_paths.values()),
"MOS_score": "",
"Emo/Not": "",
})
# Save to excel
df_combined = pd.concat([df_innominated_score])
df_named_score.to_excel("emovdb_named_mos_score.xlsx", index=False)
df_combined.to_excel("emovdb_innominated_mos_score.xlsx", index=False)
"Files created and copied successfully."
# 把输出的excel表格整理一下，按顺序写音频编号，方便annotator录入
# Load the original innominated_score.xlsx
df_to_sort = pd.read_excel("emovdb_innominated_mos_score.xlsx")
# Extract the number from the filename for each row
df_to_sort['sort_key'] = df_to_sort['file_name'].str.extract('(\d+)').astype(int)
# Sort the dataframe by this new column
df_sorted = df_to_sort.sort_values(by="sort_key")
# Drop the 'sort_key' column as it's no longer needed
df_sorted = df_sorted.drop(columns=['sort_key'])
# Save the sorted dataframe as scoring.xlsx
df_sorted.to_excel("emovdb_mos_scoring.xlsx", index=False)
print("File scoring.xlsx created and sorted successfully.")
|