Text-to-Sing / app.py
deleteman's picture
undoing last change
12a0eab
import gradio as gr
import re
import requests
import time
import tempfile
import os
token = os.environ['apikey']
API_HOST = "https://api.voicemod.net"
def separate_syllables(string):
#numbers = [ 4, 4, 8, 5 ]
numbers = [ 8, 9, 8, 9 ]
# Split the string into a list of characters
parts = string.split(" ")
string = "".join(parts)
print("the new string: "+ string)
chars = list(string)
# Create a list to store the syllables
syllables = []
# Iterate over the numbers
for num in numbers:
# Get the next num characters from the string
syllable = "".join(chars[:num])
print("Appending " + "".join(chars[:num]))
# Add the syllable to the list
syllables.append(syllable)
# Remove the syllable from the list of characters
chars = chars[num:]
# If there are any remaining characters, add them as a new line
if chars:
syllables.append("-")
for num in numbers:
# Get the next num characters from the remaining characters
syllable = "".join(chars[:num])
# Add the syllable to the list
syllables.append(syllable)
# Remove the syllable from the list of characters
chars = chars[num:]
# Return the list of syllables joined by spaces
if("-" in syllables):
results = [[]]
lines = 0
for i ,v in enumerate(syllables):
if v == '-':
lines=lines+1
else:
if(len(results) == 1):
results.append([])
results[lines].append(v)
return list(map(lambda words: " ".join(words), results))
else:
return " ".join(syllables)
def flattenLines(lyrics):
newValues = []
for i,val in enumerate(lyrics):
if(type(lyrics[i]) == list):
newValues.append(lyrics[i][0])
newValues.append(lyrics[i][1])
else:
newValues.append(lyrics[i])
return newValues
def getPayload(index):
print("Getting payload for index: " + str(index))
match index:
case 0:
return {
"voiceModel":"CFV_2",
"voiceModelStyle":"mezzo",
"language":"en",
"scoreFileRemotePath":"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Melody_C_200bpm_4x4_ABCD.json",
"scoreDataFileRemotePath":"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Score_Data_ABCD.json",
"dspFileRemotePath": "themes/DeckTheHalls/dsp/XML/TTS_XMAS_DeckTheHalls_DSP_C_200bpm_ABCD.xml",
"lyrics": [],
"dspEffect": "",
"backgroundSounds":False,
"artistName":"Cecilia",
"songName":"Deck The Halls",
"singerImageRemotePath":"singers/Singer-Cecilia-VideoImageBackground.png",
"backgroundImageRemotePath":"themes/DeckTheHalls/images/Song-DeckTheHalls-VideoImageBackground02.png"
}
case 1:
return {
"voiceModel":"CFV_2",
"voiceModelStyle":"mezzo",
"language":"en",
"scoreDataFileRemotePath": "themes/MoveYourBody/score/MoveYourBody_Score_Data_ABCD.json",
"scoreFileRemotePath": "themes/MoveYourBody/score/MoveYourBody_Melody_ABCD.json",
"singerImageRemotePath": "singers/Singer-Cecilia-VideoImageBackground.png",
"dspFileRemotePath": "themes/MoveYourBody/dsp/XML/MoveYourBody_DSP_ABCD.xml",
"dspEffect":"",
"lyrics": [],
"backgroundSounds":False,
"artistName":"Cecilia",
"songName":"Move Your Body",
"backgroundImageRemotePath":"themes/DeckTheHalls/images/Song-DeckTheHalls-VideoImageBackground02.png"
}
case 2:
return {
"voiceModel": "Randy_v3o",
"voiceModelStyle": "randy",
"language": "en",
"scoreFileRemotePath": "themes/DarkTrap/score/DarkTrap_Melody_ABCDEFGH.json", #"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Melody_C_200bpm_4x4_ABCD.json",
"scoreDataFileRemotePath": "themes/DarkTrap/score/DarkTrap_Score_Data_ABCDEFGH.json", #"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Score_Data_ABCD.json",
"dspFileRemotePath": "themes/DarkTrap/dsp/XML/DarkTrap_DSP_ABCDEFGH.xml", #"themes/DeckTheHalls/dsp/XML/TTS_XMAS_DeckTheHalls_DSP_C_200bpm_ABCD.xml",
"dspEffect": "",
"lyrics": [],
"backgroundSounds": False,
"artistName": "Jerry",
"songName": "Dark Trap", #"Deck The Halls",
"singerImageRemotePath": "singers/Singer-Jerry-VideoImageBackground.png",
"backgroundImageRemotePath": "themes/DeckTheHalls/images/Song-DeckTheHalls-VideoImageBackground01.png"
}
case 3:
return {
"artistName": "Jerry",
"backgroundImageRemotePath": "themes/HappyBirthday/images/Song-HappyBirthday-VideoImageBackground01.png",
"backgroundSounds": False,
"dspEffect": "",
"dspFileRemotePath": "themes/HappyBirthday/dsp/XML/HappyBirthday_DSP_ABCD.xml",
"language": "en",
"lyrics": [],
"scoreDataFileRemotePath": "themes/HappyBirthday/score/HappyBirthday_Score_Data_ABCD.json",
"scoreFileRemotePath": "themes/HappyBirthday/score/HappyBirthday_Melody_ABCD.json",
"singerImageRemotePath": "singers/Singer-Jerry-VideoImageBackground.png",
"songName": "Happy Birthday",
"voiceModel": "CFV_2",
"voiceModelStyle": "baritone"
}
case 4:
return {
"voiceModel": "Randy_v3o",
"voiceModelStyle": "randy",
"language": "en",
"scoreFileRemotePath": "themes/Levitate/score/Levitate_Melody_ABCD.json", #"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Melody_C_200bpm_4x4_ABCD.json",
"scoreDataFileRemotePath": "themes/Levitate/score/Levitate_Score_Data_ABCD.json", #"themes/DeckTheHalls/score/TTS_XMAS_DeckTheHalls_Score_Data_ABCD.json",
"dspFileRemotePath": "themes/Levitate/dsp/XML/Levitate_DSP_ABCD.xml", #"themes/DeckTheHalls/dsp/XML/TTS_XMAS_DeckTheHalls_DSP_C_200bpm_ABCD.xml",
"dspEffect": "",
"lyrics": [],
"backgroundSounds": False,
"artistName": "Jerry",
"songName": "Levitate", #"Deck The Halls",
"singerImageRemotePath": "singers/Singer-Ed-VideoImageBackground.png",
"backgroundImageRemotePath": "themes/Levitate/images/Song-Levitate-VideoImageBackground02.png"
}
def cleanUpLines(lines):
return list(filter(None, lines))
def greet(index, lyrics):
url = API_HOST + "/v2/cloud/partners/ttsing"
verification_url = API_HOST + "/v2/cloud/partners/ttsing/"
print("Calling api with "+ lyrics)
print("URL: " + url)
lines = cleanUpLines(lyrics.split("\n"))
#newLines = flattenLines(list(map(separate_syllables, lines)))
payload = getPayload(index)
payload["lyrics"] = lines
headers = {
'x-api-key': token
}
print(payload)
print("Before the call")
response = requests.request("POST", url, headers=headers, json=payload)
print("After the call...")
jsonResp = response.json()
print(response.text)
jobID = jsonResp['jobId']
verification_url += jsonResp['jobId']
while(True):
response = requests.request('GET', verification_url, headers=headers)
print(response.text)
print(jobID)
jsonResp = response.json()
if(jsonResp['status'] == 'failed'):
return ""
if('transformedAudioUrl' in jsonResp and jsonResp['transformedAudioUrl'] is not None):
return gr.make_waveform(download_file(jsonResp['transformedAudioUrl']))
#return download_file(jsonResp['urlVideoWithBackground'])
print("Still processing...")
time.sleep(1)
def download_file(url):
response = requests.get(url)
if response.status_code == 200:
with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
tmp_file.write(response.content)
tmp_file.flush()
return tmp_file.name
else:
print("Error: Unable to download file")
with gr.Blocks() as demo:
#def handleChange():
# line3.update(visible=False)
gr.Markdown("""
## Voicemod's Text-to-Sing API
To use this API follow the instructions:
1. First, select the melody you want from the dropdown
2. Then write the lyrics for it
### Disclaimer
Each melody has specific requirements for the lyrics (number of notes per verse and duration of each note).
Use the examples provided to understand how they work and write valid lyrics to get best results
Also, the length of the song is fixed, if you go over the line numbers the extra lyrics will be ignored.
## Join our Community
If you'd like to know more and meet other developers working with this technology, join our [Discord Server](https://discord.gg/vm-dev-community)!
""")
music_options = ["Deck the Halls by Cecilia",
"Move your Body by Cecilia",
"Dark Trap by Jerry",
"Happy Birthday by Jerry",
"Levitate by Ed"]
with gr.Row():
with gr.Column():
with gr.Row():
dd = gr.Dropdown(choices=music_options, type="index", label="Select the melody...")
lines = gr.Textbox(lines=10, placeholder="Write your lyrics here...", label="Lyrics")
with gr.Row():
btn = gr.Button("Run")
with gr.Column():
video = gr.Video(label="Generated output")
video.style(height=300)
gr.Markdown("""
## Example lyrics
To make your life easier, while testing the API, you can use and modify the following lyrics for each song:
""")
gr.Examples(fn=greet,
examples=[["Move your Body by Cecilia", """this is just a test
don’t worry now
the lyrics will come the less you think about them
just feel the melody"""
],
["Dark Trap by Jerry","""this is just a test
don't worry about what I say
the lyrics want to go crazy
when pushed too hard
take it easy and relax
this is moving too fast
I can help you be
a better singer"""
],
["Happy Birthday by Jerry","""happy birthday to you
happy birthday to you!
happy birthday dear Laura
happy birthday to you!"""] ,
["Deck the Halls by Cecilia", """Deck the halls with boughs of holly
Fa la la la la, la la la la
'Tis the season to be jolly
Fa la la la la, la la la la""" ],
["Levitate by Ed", """this is just a test
don’t worry about what I say
the lyrics want to go crazy
when pushed too hard"""]
],
inputs=[
dd,
lines
],
outputs=video,
cache_examples=True
)
gr.Markdown("""
## Want to use this API for your project?
If you'd like to use this API for your own project, request access through our [form here](https://voicemod.typeform.com/to/KqeNN6bO?typeform-source=huggingface) and join
our [Dev Community on Discord](https://discord.gg/vm-dev-community)!
""")
btn.click(fn=greet,
inputs=[
dd,
lines
],
outputs=video)
demo.launch()