pitts should work
Browse files- App/TTS/Schemas.py +5 -0
- App/TTS/TTSRoutes.py +48 -0
- App/TTS/utils/Pi.py +189 -0
- App/TTS/utils/Picsart.py +0 -33
App/TTS/Schemas.py
CHANGED
@@ -30,6 +30,11 @@ class DescriptTranscript(BaseModel):
|
|
30 |
file_extenstion: str = ".wav"
|
31 |
|
32 |
|
|
|
|
|
|
|
|
|
|
|
33 |
class DescriptRequest(BaseModel):
|
34 |
text: str
|
35 |
speaker: Optional[str] = Field(default="Lawrance")
|
|
|
30 |
file_extenstion: str = ".wav"
|
31 |
|
32 |
|
33 |
+
class PiTTSRequest(BaseModel):
|
34 |
+
text: str
|
35 |
+
voice: Optional[str]
|
36 |
+
|
37 |
+
|
38 |
class DescriptRequest(BaseModel):
|
39 |
text: str
|
40 |
speaker: Optional[str] = Field(default="Lawrance")
|
App/TTS/TTSRoutes.py
CHANGED
@@ -9,13 +9,19 @@ from .Schemas import (
|
|
9 |
DescriptStatusRequest,
|
10 |
DescriptSfxRequest,
|
11 |
DescriptTranscript,
|
|
|
12 |
)
|
13 |
from .utils.Podcastle import PodcastleAPI
|
14 |
from .utils.HeyGen import HeygenAPI
|
|
|
15 |
from .utils.Descript import DescriptTTS
|
16 |
import os
|
17 |
import asyncio
|
18 |
|
|
|
|
|
|
|
|
|
19 |
tts_router = APIRouter(tags=["TTS"])
|
20 |
data = {"username": os.environ.get("USERNAME"), "password": os.environ.get("PASSWORD")}
|
21 |
tts = PodcastleAPI(**data)
|
@@ -27,6 +33,7 @@ data = {
|
|
27 |
|
28 |
descript_tts = DescriptTTS()
|
29 |
heyGentts = HeygenAPI(**data)
|
|
|
30 |
|
31 |
|
32 |
@tts_router.post("/generate_tts")
|
@@ -86,3 +93,44 @@ async def auto_refresh():
|
|
86 |
@tts_router.post("/status")
|
87 |
async def search_id(req: StatusRequest):
|
88 |
return await tts.check_status(req)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
DescriptStatusRequest,
|
10 |
DescriptSfxRequest,
|
11 |
DescriptTranscript,
|
12 |
+
PiTTSRequest,
|
13 |
)
|
14 |
from .utils.Podcastle import PodcastleAPI
|
15 |
from .utils.HeyGen import HeygenAPI
|
16 |
+
from .utils.Pi import PiAIClient
|
17 |
from .utils.Descript import DescriptTTS
|
18 |
import os
|
19 |
import asyncio
|
20 |
|
21 |
+
from fastapi import FastAPI, Request, HTTPException
|
22 |
+
from fastapi.responses import StreamingResponse, FileResponse
|
23 |
+
import os
|
24 |
+
|
25 |
tts_router = APIRouter(tags=["TTS"])
|
26 |
data = {"username": os.environ.get("USERNAME"), "password": os.environ.get("PASSWORD")}
|
27 |
tts = PodcastleAPI(**data)
|
|
|
33 |
|
34 |
descript_tts = DescriptTTS()
|
35 |
heyGentts = HeygenAPI(**data)
|
36 |
+
pi = PiAIClient()
|
37 |
|
38 |
|
39 |
@tts_router.post("/generate_tts")
|
|
|
93 |
@tts_router.post("/status")
|
94 |
async def search_id(req: StatusRequest):
|
95 |
return await tts.check_status(req)
|
96 |
+
|
97 |
+
|
98 |
+
@tts_router.post("/pi_tts")
|
99 |
+
async def pi_tts(req: PiTTSRequest):
|
100 |
+
return await pi.say(text=req.text, voice=req.voice)
|
101 |
+
|
102 |
+
|
103 |
+
@tts_router.get("/audio/{audio_name}")
|
104 |
+
async def serve_audio(request: Request, audio_name: str):
|
105 |
+
audio_directory = "/tmp/Audio"
|
106 |
+
audio_path = os.path.join(audio_directory, audio_name)
|
107 |
+
if not os.path.isfile(audio_path):
|
108 |
+
raise HTTPException(status_code=404, detail="Audio not found")
|
109 |
+
|
110 |
+
range_header = request.headers.get("Range", None)
|
111 |
+
audio_size = os.path.getsize(audio_path)
|
112 |
+
|
113 |
+
if range_header:
|
114 |
+
start, end = range_header.strip().split("=")[1].split("-")
|
115 |
+
start = int(start)
|
116 |
+
end = audio_size if end == "" else int(end)
|
117 |
+
|
118 |
+
headers = {
|
119 |
+
"Content-Range": f"bytes {start}-{end}/{audio_size}",
|
120 |
+
"Accept-Ranges": "bytes",
|
121 |
+
# Optionally, you might want to force download by uncommenting the next line:
|
122 |
+
# "Content-Disposition": f"attachment; filename={audio_name}",
|
123 |
+
}
|
124 |
+
|
125 |
+
content = read_file_range(audio_path, start, end)
|
126 |
+
return StreamingResponse(content, media_type="audio/mpeg", headers=headers)
|
127 |
+
|
128 |
+
return FileResponse(audio_path, media_type="audio/mpeg")
|
129 |
+
|
130 |
+
|
131 |
+
def read_file_range(path, start, end):
|
132 |
+
"""Helper function to read specific range of bytes from a file."""
|
133 |
+
with open(path, "rb") as file:
|
134 |
+
file.seek(start)
|
135 |
+
# Be sure to handle the case where `end` is not the last byte
|
136 |
+
return file.read(end - start + 1)
|
App/TTS/utils/Pi.py
ADDED
@@ -0,0 +1,189 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import aiohttp
|
2 |
+
import asyncio
|
3 |
+
import enum
|
4 |
+
import requests
|
5 |
+
import os
|
6 |
+
from functools import cache
|
7 |
+
import tempfile
|
8 |
+
import uuid
|
9 |
+
|
10 |
+
|
11 |
+
class VoiceType(enum.Enum):
|
12 |
+
voice1 = "voice1"
|
13 |
+
voice2 = "voice2"
|
14 |
+
voice3 = "voice3"
|
15 |
+
voice4 = "voice4"
|
16 |
+
voice5 = "voice5"
|
17 |
+
voice5_update = "voice5-update"
|
18 |
+
voice6 = "voice6"
|
19 |
+
voice7 = "voice7"
|
20 |
+
voice8 = "voice8"
|
21 |
+
voice9 = "voice9"
|
22 |
+
voice10 = "voice10"
|
23 |
+
voice11 = "voice11"
|
24 |
+
voice12 = "voice12"
|
25 |
+
qdpi = "qdpi"
|
26 |
+
|
27 |
+
|
28 |
+
class PiAIClient:
|
29 |
+
def __init__(self):
|
30 |
+
self.dir = "/tmp/Audio"
|
31 |
+
self.base_url = "https://pi.ai/api/chat"
|
32 |
+
self.referer = "https://pi.ai/talk"
|
33 |
+
self.origin = "https://pi.ai"
|
34 |
+
self.user_agent = (
|
35 |
+
"Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/113.0"
|
36 |
+
)
|
37 |
+
self.cookie = None
|
38 |
+
self.headers = {
|
39 |
+
"User-Agent": self.user_agent,
|
40 |
+
"Accept": "text/event-stream",
|
41 |
+
"Referer": self.referer,
|
42 |
+
"X-Api-Version": "3",
|
43 |
+
"Content-Type": "application/json",
|
44 |
+
"Origin": self.origin,
|
45 |
+
"Connection": "keep-alive",
|
46 |
+
"Sec-Fetch-Dest": "empty",
|
47 |
+
"Sec-Fetch-Mode": "no-cors",
|
48 |
+
"Sec-Fetch-Site": "same-origin",
|
49 |
+
"DNT": "1",
|
50 |
+
"Sec-GPC": "1",
|
51 |
+
"TE": "trailers",
|
52 |
+
"Pragma": "no-cache",
|
53 |
+
"Cache-Control": "no-cache",
|
54 |
+
}
|
55 |
+
|
56 |
+
async def get_cookie(self) -> str:
|
57 |
+
headers = self.headers.copy()
|
58 |
+
async with aiohttp.ClientSession() as session:
|
59 |
+
async with session.post(
|
60 |
+
f"{self.base_url}/start", headers=headers, json={}
|
61 |
+
) as response:
|
62 |
+
self.cookie = response.headers["Set-Cookie"]
|
63 |
+
return self.cookie
|
64 |
+
|
65 |
+
async def make_request(
|
66 |
+
self, endpoint: str, headers: dict, json: dict = None, method: str = "POST"
|
67 |
+
):
|
68 |
+
async with aiohttp.ClientSession() as session:
|
69 |
+
if method == "POST":
|
70 |
+
async with session.post(
|
71 |
+
endpoint, headers=headers, json=json
|
72 |
+
) as response:
|
73 |
+
return await response.text()
|
74 |
+
elif method == "GET":
|
75 |
+
async with session.get(endpoint, headers=headers) as response:
|
76 |
+
return response
|
77 |
+
|
78 |
+
async def get_response(self, input_text) -> tuple[list[str], list[str]]:
|
79 |
+
if self.cookie is None:
|
80 |
+
self.cookie = await self.get_cookie()
|
81 |
+
|
82 |
+
headers = self.headers.copy()
|
83 |
+
headers["Cookie"] = self.cookie
|
84 |
+
|
85 |
+
data = {"text": input_text}
|
86 |
+
response_text = await self.make_request(self.base_url, headers, json=data)
|
87 |
+
|
88 |
+
response_lines = response_text.split("\n")
|
89 |
+
response_texts = []
|
90 |
+
response_sids = []
|
91 |
+
print(response_lines)
|
92 |
+
for line in response_lines:
|
93 |
+
if line.startswith('data: {"text":"'):
|
94 |
+
start = len('data: {"text":')
|
95 |
+
end = line.rindex("}")
|
96 |
+
text_dict = line[start + 1 : end - 1].strip()
|
97 |
+
response_texts.append(text_dict)
|
98 |
+
elif line.startswith('data: {"sid":'):
|
99 |
+
start = len('data: {"sid":')
|
100 |
+
end = line.rindex("}")
|
101 |
+
sid_dict = line[start : end - 1].strip()
|
102 |
+
sid_dict = sid_dict.split(",")[0][1:-1]
|
103 |
+
response_sids.append(sid_dict)
|
104 |
+
|
105 |
+
return response_texts, response_sids
|
106 |
+
|
107 |
+
async def speak_response(
|
108 |
+
self, message_sid: str, voice: VoiceType = VoiceType.voice4
|
109 |
+
) -> None:
|
110 |
+
if self.cookie is None:
|
111 |
+
self.cookie = await self.get_cookie()
|
112 |
+
|
113 |
+
headers = self.headers.copy()
|
114 |
+
headers.update(
|
115 |
+
{
|
116 |
+
"Host": "pi.ai",
|
117 |
+
"Accept": "audio/webm,audio/ogg,audio/wav,audio/*;q=0.9,application/ogg;q=0.7,video/*;q=0.6,*/*;q=0.5",
|
118 |
+
"Accept-Language": "en-US,en;q=0.9",
|
119 |
+
"Range": "bytes=0-",
|
120 |
+
"Sec-Fetch-Dest": "audio",
|
121 |
+
"Sec-Fetch-Mode": "no-cors",
|
122 |
+
"Sec-Fetch-Site": "same-origin",
|
123 |
+
"Sec-CH-UA": '"Chromium";v="124", "Google Chrome";v="124", "Not-A.Brand";v="99"',
|
124 |
+
"Sec-CH-UA-Mobile": "?0",
|
125 |
+
"Sec-CH-UA-Platform": '"Windows"',
|
126 |
+
}
|
127 |
+
)
|
128 |
+
|
129 |
+
headers = {
|
130 |
+
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/113.0",
|
131 |
+
"Accept": "audio/webm,audio/ogg,audio/wav,audio/*;q=0.9,application/ogg;q=0.7,video/*;q=0.6,*/*;q=0.5",
|
132 |
+
"Accept-Language": "en-US,en;q=0.5",
|
133 |
+
"Range": "bytes=0-",
|
134 |
+
"Connection": "keep-alive",
|
135 |
+
"Referer": "https://pi.ai/talk",
|
136 |
+
# "Cookie": cookie,
|
137 |
+
"Sec-Fetch-Dest": "audio",
|
138 |
+
"Sec-Fetch-Mode": "no-cors",
|
139 |
+
"Sec-Fetch-Site": "same-origin",
|
140 |
+
"DNT": "1",
|
141 |
+
"Sec-GPC": "1",
|
142 |
+
"Accept-Encoding": "identity",
|
143 |
+
"TE": "trailers",
|
144 |
+
}
|
145 |
+
headers["Cookie"] = self.cookie
|
146 |
+
print(headers)
|
147 |
+
endpoint = f"{self.base_url}/voice?mode=eager&voice={voice.value}&messageSid={message_sid}"
|
148 |
+
async with aiohttp.ClientSession() as session:
|
149 |
+
async with session.get(endpoint, headers=headers) as response:
|
150 |
+
print(response.status)
|
151 |
+
file_name = str(uuid.uuid4()) + ".mp3"
|
152 |
+
file_path = os.path.join(self.dir, file_name)
|
153 |
+
os.makedirs(file_path, exist_ok=True)
|
154 |
+
if response.status == 200:
|
155 |
+
with open(file_path, "wb") as file:
|
156 |
+
async for chunk in response.content.iter_chunked(128):
|
157 |
+
file.write(chunk)
|
158 |
+
return {
|
159 |
+
"url": f"https://yakova-embedding.hf.space/audio/{file_name}"
|
160 |
+
}
|
161 |
+
# Run command vlc to play the audio file
|
162 |
+
# os.system("vlc speak.wav --intf dummy --play-and-exit")
|
163 |
+
else:
|
164 |
+
temp = await response.text()
|
165 |
+
print(temp)
|
166 |
+
return "Error: Unable to retrieve audio."
|
167 |
+
|
168 |
+
async def say(self, text, voice=VoiceType.qdpi):
|
169 |
+
_, response_sids = await self.get_response(text)
|
170 |
+
|
171 |
+
if response_sids:
|
172 |
+
return await self.speak_response(response_sids[0], voice=voice)
|
173 |
+
|
174 |
+
|
175 |
+
# async def main():
|
176 |
+
# client = PiAIClient()
|
177 |
+
# response_texts, response_sids = await client.get_response(
|
178 |
+
# "Write a ryme to introduce yourself."
|
179 |
+
# )
|
180 |
+
# print(response_texts, response_sids)
|
181 |
+
# import time
|
182 |
+
|
183 |
+
# if response_sids:
|
184 |
+
# return await client.speak_response(response_sids[1])
|
185 |
+
|
186 |
+
|
187 |
+
# # Run the main function
|
188 |
+
# if __name__ == "__main__":
|
189 |
+
# asyncio.run(main())
|
App/TTS/utils/Picsart.py
DELETED
@@ -1,33 +0,0 @@
|
|
1 |
-
import asyncio
|
2 |
-
from playwright.async_api import async_playwright
|
3 |
-
|
4 |
-
async def extract_bootstrap_data():
|
5 |
-
async with async_playwright() as p:
|
6 |
-
browser = await p.chromium.launch()
|
7 |
-
page = await browser.new_page()
|
8 |
-
|
9 |
-
# Navigate to the Pixabay Sound Effects page
|
10 |
-
await page.goto('https://pixabay.com/sound-effects/search/door%20creaking/')
|
11 |
-
|
12 |
-
# Wait for the content to load (you can adjust the timeout as needed)
|
13 |
-
# await page.wait_for_selector('.js-media-list-wrapper')
|
14 |
-
# await page.wait_for_selector('.js-media-item')
|
15 |
-
|
16 |
-
# Get the content of the 5th script tag
|
17 |
-
# script_content = await page.evaluate('''() => {
|
18 |
-
# const scripty=document.querySelectorAll('script')[0];
|
19 |
-
# return scripty.content
|
20 |
-
# }''')
|
21 |
-
# print(script_content)
|
22 |
-
# await page.evaluate(f'''{script_content}(''')
|
23 |
-
|
24 |
-
page_content = await page.content()
|
25 |
-
# Print the content of the 5th script tag
|
26 |
-
# print(page_content)
|
27 |
-
|
28 |
-
# Close the browser
|
29 |
-
await browser.close()
|
30 |
-
|
31 |
-
# Run the extraction function
|
32 |
-
if __name__ == '__main__':
|
33 |
-
asyncio.run(extract_bootstrap_data())
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|