app.py

import time
import streamlit as st
# from transformers import pipeline
import os
import torch
import datetime
import numpy as np
import soundfile
from wavmark.utils import file_reader


# pipeline = pipeline(task="image-classification", model="julien-c/hotdog-not-hotdog")

# st.title("Hot Dog? Or Not?")

# file_name = st.file_uploader("Upload a hot dog candidate image")

# if file_name is not None:
#     col1, col2 = st.columns(2)

#     image = Image.open(file_name)
#     col1.image(image, use_column_width=True)
#     predictions = pipeline(image)

#     col2.header("Probabilities")
#     for p in predictions:
#         col2.subheader(f"{ p['label'] }: { round(p['score'] * 100, 1)}%")

def create_default_value():
    if "def_value" not in st.session_state:
        def_val_npy = np.random.choice([0, 1], size=32 - len_start_bit)
        def_val_str = "".join([str(i) for i in def_val_npy])
        st.session_state.def_value = def_val_str

# Main web app
def main():
    create_default_value()

    # st.title("MDS07")
    # st.write("https://github.com/wavmark/wavmark")
    markdown_text = """
    # MDS07
    [AudioSeal](https://github.com/jcha0155/AudioSealEnhanced) is the next-generation watermarking tool driven by AI. 
    You can upload an audio file and encode a custom 16-bit watermark or perform decoding from a watermarked audio.
    
    This page is for demonstration usage and only process **the first minute** of the audio. 
    If you have longer files for processing, we recommend using [our python toolkit](https://github.com/jcha0155/AudioSealEnhanced).
    """

    # 使用st.markdown渲染Markdown文本
    st.markdown(markdown_text)

    audio_file = st.file_uploader("Upload Audio", type=["wav", "mp3"], accept_multiple_files=False)

    if audio_file:
        # 保存文件到本地：
        tmp_input_audio_file = os.path.join("/tmp/", audio_file.name)
        with open(tmp_input_audio_file, "wb") as f:
            f.write(audio_file.getbuffer())

        # 展示文件到页面上
        # st.audio(tmp_input_audio_file, format="audio/wav")

        action = st.selectbox("Select Action", ["Add Watermark", "Decode Watermark"])

        if action == "Add Watermark":
            watermark_text = st.text_input("The watermark (0, 1 list of length-16):", value=st.session_state.def_value)
            add_watermark_button = st.button("Add Watermark", key="add_watermark_btn")
            if add_watermark_button:  # 点击按钮后执行的
                if audio_file and watermark_text:
                    with st.spinner("Adding Watermark..."):
                        # watermarked_audio, encode_time_cost = add_watermark(tmp_input_audio_file, watermark_text)
#                         st.write("Watermarked Audio:")
#                         print("watermarked_audio:", watermarked_audio)
#                         st.audio(watermarked_audio, format="audio/wav")
#                         st.write("Time Cost: %d seconds" % encode_time_cost)

#                         # st.button("Add Watermark", disabled=False)
#         elif action == "Decode Watermark":
#             if st.button("Decode"):
#                 with st.spinner("Decoding..."):
#                     decode_watermark(tmp_input_audio_file)


if __name__ == "__main__":
    default_sr = 16000
    max_second_encode = 60
    max_second_decode = 30
    len_start_bit = 16
    device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
    # model = wavmark.load_model().to(device)
    main()

    # audio_path = "/Users/my/Library/Mobile Documents/com~apple~CloudDocs/CODE/PycharmProjects/4_语音水印/419_huggingface水印/WavMark/example.wav"

    # decoded_watermark, decode_cost = decode_watermark(audio_path)
    # print(decoded_watermark)