feat: add plotly, audiorec, update requirements

README.md

@@ -1,10 +1,10 @@
 ---
-title: Test Space
+title: Music Splitter
 emoji: 🎶
 colorFrom: indigo
 colorTo: yellow
 sdk: docker
-pinned: false
+pinned: true
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app/helpers.py

@@ -0,0 +1,24 @@
+from pydub import AudioSegment
+
+import streamlit as st
+import plotly.graph_objs as go
+import plotly.express as px
+import pandas as pd
+import numpy as np
+
+@st.cache_data
+def load_audio_segment(path: str, format: str) -> AudioSegment:
+    return AudioSegment.from_file(path, format=format)
+
+#@st.cache_resource
+def plot_audio(_audio_segment: AudioSegment, title: str = None) -> go.Figure:
+    samples = _audio_segment.get_array_of_samples()
+    arr = np.array(samples)
+    df = pd.DataFrame(arr)
+    fig = px.line(df, y=0, render_mode="webgl", line_shape="linear", width=1000, height=60, title=title)
+    fig.update_layout(xaxis_fixedrange=True, yaxis_fixedrange=True, yaxis_visible=False, xaxis_visible=False, hovermode=False, margin=dict(l=0, r=0, t=0, b=0))
+            #'font': {'size': 18},
+            #'xanchor': 'center',
+            #'yanchor': 'top'})
+    st.plotly_chart(fig, use_container_width=True)
+    #return fig

app/main.py

@@ -1,18 +1,13 @@
 import logging
 import os
 from pathlib import Path
-from shutil import rmtree
 
-from pydub import AudioSegment
+import requests
 import streamlit as st
 
-# from lib.st_custom_components import st_audiorec
-
-try:
-    from app.demucs_runner import separator
-except ImportError:
-    from demucs_runner import separator
-
+from demucs_runner import separator
+from lib.st_custom_components import st_audiorec
+from helpers import load_audio_segment, plot_audio
 
 logging.basicConfig(
     format="%(asctime)s %(levelname)-8s %(message)s",
@@ -34,37 +29,28 @@ int24 = False  # output as int24 wavs, unused if 'mp3' is True.
 # You cannot set both `float32 = True` and `int24 = True` !!
 
 
-def find_files(in_path):
-    out = []
-    for file in Path(in_path).iterdir():
-        if file.suffix.lower().lstrip(".") in extensions:
-            out.append(file)
-    return out
-
-out_path = Path("tmp")
-in_path = Path("tmp")
+out_path = Path("/tmp")
+in_path = Path("/tmp")
 
-# def clean_folders():
-#     if in_path.exists():
-#         rmtree(in_path)
-#     in_path.mkdir()
-#     if out_path.exists():
-#         rmtree(out_path)
-#     out_path.mkdir()
     
 def url_is_valid(url):
-    import requests
+    if url.startswith("http") is False:
+        st.error("URL should start with http or https.")
+        return False
+    elif url.split(".")[-1] not in extensions:
+        st.error("Extension not supported.")
+        return False
     try:
         r = requests.get(url)
         r.raise_for_status()
         return True
     except Exception:
         st.error("URL is not valid.")
+        return False
 
 
 def run():
     st.markdown("<h1><center>🎶 Music Source Splitter</center></h1>", unsafe_allow_html=True)
-
     st.markdown("""
                 <style>
                 .st-af {
@@ -80,52 +66,42 @@ def run():
     filename = None
     choice =  st.radio(label=" ", options=["🔗 From URL", "⬆️ Upload File", "🎤 Record Audio"], horizontal=True)
     if choice == "🔗 From URL":
-        url = st.text_input("Paste the URL of the audio file", key="url")
+        url = st.text_input("Paste the URL of the audio file", key="url", help="Supported formats: mp3, wav, ogg, flac.")
         if url != "":
             # check if the url is valid
             if url_is_valid(url):
                 with st.spinner("Downloading audio..."):
-                    #clean_folders()
                     filename = url.split("/")[-1]
                     os.system(f"wget -O {in_path / filename} {url}")
             
     elif choice == "⬆️ Upload File":
-        uploaded_file = st.file_uploader("Choose a file")
+        uploaded_file = st.file_uploader("Choose a file", type=extensions, key="file", help="Supported formats: mp3, wav, ogg, flac.")
         if uploaded_file is not None:
-            #clean_folders()
             with open(in_path / uploaded_file.name, "wb") as f:
                 f.write(uploaded_file.getbuffer())    
             filename = uploaded_file.name        
     elif choice == "🎤 Record Audio":
-        # wav_audio_data = st_audiorec()
-        # if wav_audio_data is not None:
-        #     if wav_audio_data != b'RIFF,\x00\x00\x00WAVEfmt \x10\x00\x00\x00\x01\x00\x02\x00\x80>\x00\x00\x00\xfa\x00\x00\x04\x00\x10\x00data\x00\x00\x00\x00':
-        #         clean_folders()
-        #         filename = "recording.wav"
-        #         with open(in_path / filename, "wb") as f:
-        #             f.write(wav_audio_data)
-        pass
+        wav_audio_data = st_audiorec()
+        if wav_audio_data is not None:
+            if wav_audio_data != b'RIFF,\x00\x00\x00WAVEfmt \x10\x00\x00\x00\x01\x00\x02\x00\x80>\x00\x00\x00\xfa\x00\x00\x04\x00\x10\x00data\x00\x00\x00\x00':
+                filename = "recording.wav"
+                with open(in_path / filename, "wb") as f:
+                    f.write(wav_audio_data)                
                 
     if filename is not None:
-        st.markdown("<hr>", unsafe_allow_html=True)
-        cols = st.columns(2)
-        with cols[0]:
-            st.markdown("<h3>Original Audio</h3>", unsafe_allow_html=True)
-        with cols[1]:
-            audio_file = open(in_path / filename, "rb")
-            audio_bytes = audio_file.read()
-            _ = st.audio(audio_bytes)
+        song = load_audio_segment(in_path / filename, filename.split(".")[-1])
         
-        song = AudioSegment.from_file(in_path / filename, filename.split(".")[-1])
         n_secs = round(len(song) / 1000)
-        start_time = st.slider("Choose the start time", min_value=0, max_value=n_secs, value=0, help=f"Maximum duration is {max_duration} seconds.")
+        audio_file = open(in_path / filename, "rb")
+        audio_bytes = audio_file.read()
+        start_time = st.slider("Choose the start time", min_value=0, max_value=n_secs, step=1, value=0, help=f"Maximum duration is {max_duration} seconds.")
+        _ = st.audio(audio_bytes, start_time=start_time)
         end_time = min(start_time + max_duration, n_secs)
+        song = song[start_time*1000:end_time*1000]
         tot_time = end_time - start_time
         st.info(f"Audio source will be processed from {start_time} to {end_time} seconds.", icon="⏱")
-         
         execute = st.button("Split Music 🎶", type="primary")
         if execute:
-            song = song[start_time*1000:end_time*1000]
             song.export(in_path / filename, format=filename.split(".")[-1])
             with st.spinner(f"Splitting source audio, it will take almost {round(tot_time*3.6)} seconds..."):
                 separator(
@@ -144,22 +120,25 @@ def run():
                     jobs=os.cpu_count(),
                     verbose=True,
                 )
-                pass
+
             last_dir = ".".join(filename.split(".")[:-1])
             for file in ["vocals.mp3", "drums.mp3", "bass.mp3", "other.mp3"]:
                 file = out_path / Path(model) / last_dir / file
                 st.markdown("<hr>", unsafe_allow_html=True)
+                label = file.name.split(".")[0].replace("_", " ").capitalize()
+                # add emoji to label
+                label = {
+                    "Drums": "🥁",
+                    "Bass": "🎸",
+                    "Other": "🎹",
+                    "Vocals": "🎤",
+                }.get(label) + " " + label
+                st.markdown("<center><h3>" + label + "</h3></center>", unsafe_allow_html=True)
+ 
                 cols = st.columns(2)
                 with cols[0]:
-                    label = file.name.split(".")[0].replace("_", " ").capitalize()
-                    # add emoji to label
-                    label = {
-                        "Drums": "🥁",
-                        "Bass": "🎸",
-                        "Other": "🎹",
-                        "Vocals": "🎤",
-                    }.get(label) + " " + label
-                    st.markdown("<h3>" + label + "</h3>", unsafe_allow_html=True)
+                   auseg = load_audio_segment(file, "mp3")
+                   plot_audio(auseg)
                 with cols[1]:
                     audio_file = open(file, "rb")
                     audio_bytes = audio_file.read()

requirements.txt

@@ -1,3 +1,171 @@
-streamlit==1.18.1
+#
+# This file is autogenerated by pip-compile with Python 3.10
+# by the following command:
+#
+#    pip-compile --output-file=requirements.txt --resolver=backtracking requirements.in
+#
+altair==4.2.2
+    # via streamlit
+antlr4-python3-runtime==4.9.3
+    # via omegaconf
+attrs==22.2.0
+    # via jsonschema
+blinker==1.5
+    # via streamlit
+cachetools==5.3.0
+    # via streamlit
+certifi==2022.12.7
+    # via requests
+charset-normalizer==3.0.1
+    # via requests
+click==8.1.3
+    # via streamlit
+cloudpickle==2.2.1
+    # via submitit
+cython==0.29.33
+    # via diffq
+decorator==5.1.1
+    # via validators
 demucs==4.0.0
+    # via -r requirements.in
+diffq==0.2.3
+    # via demucs
+dora-search==0.1.11
+    # via demucs
+einops==0.6.0
+    # via demucs
+entrypoints==0.4
+    # via altair
+gitdb==4.0.10
+    # via gitpython
+gitpython==3.1.31
+    # via streamlit
+idna==3.4
+    # via requests
+importlib-metadata==6.0.0
+    # via streamlit
+jinja2==3.1.2
+    # via
+    #   altair
+    #   pydeck
+jsonschema==4.17.3
+    # via altair
+julius==0.2.7
+    # via demucs
+lameenc==1.4.2
+    # via demucs
+markdown-it-py==2.2.0
+    # via rich
+markupsafe==2.1.2
+    # via jinja2
+mdurl==0.1.2
+    # via markdown-it-py
+numpy==1.24.2
+    # via
+    #   altair
+    #   diffq
+    #   openunmix
+    #   pandas
+    #   pyarrow
+    #   pydeck
+    #   streamlit
+omegaconf==2.3.0
+    # via dora-search
+openunmix==1.2.1
+    # via demucs
+packaging==23.0
+    # via streamlit
+pandas==1.5.3
+    # via
+    #   -r requirements.in
+    #   altair
+    #   streamlit
+pillow==9.4.0
+    # via streamlit
+plotly==5.13.0
+    # via -r requirements.in
+protobuf==3.20.3
+    # via streamlit
+pyarrow==11.0.0
+    # via streamlit
+pydeck==0.8.0
+    # via streamlit
 pydub==0.25.1
+    # via -r requirements.in
+pygments==2.14.0
+    # via rich
+pympler==1.0.1
+    # via streamlit
+pyrsistent==0.19.3
+    # via jsonschema
+python-dateutil==2.8.2
+    # via
+    #   pandas
+    #   streamlit
+pytz==2022.7.1
+    # via pandas
+pytz-deprecation-shim==0.1.0.post0
+    # via tzlocal
+pyyaml==6.0
+    # via
+    #   demucs
+    #   omegaconf
+requests==2.28.2
+    # via streamlit
+retrying==1.3.4
+    # via dora-search
+rich==13.3.1
+    # via streamlit
+semver==2.13.0
+    # via streamlit
+six==1.16.0
+    # via
+    #   python-dateutil
+    #   retrying
+smmap==5.0.0
+    # via gitdb
+streamlit==1.18.1
+    # via -r requirements.in
+submitit==1.4.5
+    # via dora-search
+tenacity==8.2.1
+    # via plotly
+toml==0.10.2
+    # via streamlit
+toolz==0.12.0
+    # via altair
+torch==1.13.1
+    # via
+    #   demucs
+    #   diffq
+    #   dora-search
+    #   julius
+    #   openunmix
+    #   torchaudio
+torchaudio==0.13.1
+    # via
+    #   demucs
+    #   openunmix
+tornado==6.2
+    # via streamlit
+tqdm==4.64.1
+    # via
+    #   demucs
+    #   openunmix
+treetable==0.2.5
+    # via dora-search
+typing-extensions==4.5.0
+    # via
+    #   streamlit
+    #   submitit
+    #   torch
+tzdata==2022.7
+    # via pytz-deprecation-shim
+tzlocal==4.2
+    # via streamlit
+urllib3==1.26.14
+    # via requests
+validators==0.20.0
+    # via streamlit
+zipp==3.14.0
+    # via importlib-metadata