-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathsadtalker_api.py
229 lines (196 loc) · 11.3 KB
/
sadtalker_api.py
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
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
from fastapi import FastAPI, File, UploadFile, Form
from fastapi.responses import StreamingResponse
from src.gradio_demo import SadTalker
from datetime import datetime
from shutil import copyfile
from interpolate import interpolate_video, optical_flow_interpolation, interpolate_frames_by_ffmpeg
app = FastAPI()
def _get_current_time():
current_time = datetime.now()
formatted_time = current_time.strftime("%Y%m%d%H%M%S")
return formatted_time
# 场景1：输入 语音+图片 ， 一次性等待返回 生成的 mp4视频
@app.post("/create_video")
async def create_video(image_file: UploadFile = File(...),
audio_file: UploadFile = File(...),
preprocess: str = Form(default='crop'),
still_mode: bool = Form(default=False),
use_enhancer: bool = Form(default=False),
fps: float = Form(default=12.),
interpolate_factor: int = Form(default=2),
add_silent_both_slides: bool = Form(default=False),
silent_seconds: float = Form(default=0.5)
):
print("DEBUG create_video 001 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
global sad_talker_crop
global sad_talker_full
full, crop, resize = None, None, None # e.g. eval("full") == full
if isinstance(eval(preprocess), str):
preprocess = eval(preprocess)
sad_talker = sad_talker_full if preprocess == 'full' else sad_talker_crop
print("DEBUG create_video 002 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# user input files
currentTime = _get_current_time()
imageFilePath = f"./image_uploaded_files/{currentTime}_{image_file.filename}"
audioFilePath = f"./audio_uploaded_files/{currentTime}_{audio_file.filename}"
# save image
with open(imageFilePath, "wb") as f:
f.write(await image_file.read())
with open(audioFilePath, "wb") as f:
f.write(await audio_file.read())
print("DEBUG create_video 003 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# other default params, TODO as params
# dynamic
pose_style: int = 0
exp_scale: float = 1.0
use_ref_video: bool = False
ref_video: str = None
input_yaw_list: list = None # [5, -1, 1]
input_pitch_list: list = None # [-3, 2, 0]
input_roll_list: list = None # [1, -1, 0]
result_dir: str = './results/'
batch_size: int = 32
fps: float = 12. # origin:25.
interpolate_dir: str = './InterpolatedVideo/'
interpolate_factor: int = 2 # origin: 1
add_silent_both_slides: bool = False
silent_seconds: float = 0.5
print("DEBUG create_video 004 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
print(f"The fps of generated video is: {fps}!")
resultVideoFilePath = sad_talker.test(imageFilePath, audioFilePath, preprocess=preprocess, still_mode=still_mode,
use_enhancer=use_enhancer, batch_size=batch_size, fps=fps,
result_dir=result_dir, exp_scale=exp_scale,
input_yaw_list=input_yaw_list, input_pitch_list=input_pitch_list,
input_roll_list=input_roll_list, add_silent_both_slides = add_silent_both_slides,
silent_seconds = silent_seconds)
print("DEBUG create_video 005 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# Interpolate Video
InterpolateVideoFilePath = interpolate_dir + str(currentTime) + '.mp4'
interpolate_video(resultVideoFilePath, InterpolateVideoFilePath, factor=interpolate_factor)
# optical_flow_interpolation(resultVideoFilePath, InterpolateVideoFilePath, factor=interpolate_factor)
# def generate_video():
# with open(resultVideoFilePath, mode="rb") as file_like:
# print("DEBUG create_video 006 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# yield from file_like
# print("DEBUG create_video 007 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
def generate_video():
with open(InterpolateVideoFilePath, mode="rb") as file_like:
print("DEBUG create_video 006 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
yield from file_like
print("DEBUG create_video 007 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
print("DEBUG create_video 008 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# 返回生成的视频
return StreamingResponse(generate_video(), media_type="video/mp4")
@app.post("/create_idle_video")
async def create_idle_video(image_file: UploadFile = File(...),
preprocess: str = Form(default='crop'),
still_mode: bool = Form(default=True),
use_enhancer: bool = Form(default=False),
fps: float = Form(default=12.),
interpolate_factor: int = Form(default=4),
):
print("DEBUG create_video 001 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
global sad_talker_crop
global sad_talker_full
full, crop, resize = None, None, None # e.g. eval("full") == full
if isinstance(eval(preprocess), str):
preprocess = eval(preprocess)
sad_talker = sad_talker_full if preprocess == 'full' else sad_talker_crop
print("DEBUG create_video 002 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# user input files
currentTime = _get_current_time()
imageFilePath = f"./image_uploaded_files/{currentTime}_{image_file.filename}"
audioFilePath = f"./audio_uploaded_files/{currentTime}_silent.wav"
defaultAudioFilePath = f"./silent.wav"
copyfile(defaultAudioFilePath, audioFilePath)
# save image
with open(imageFilePath, "wb") as f:
f.write(await image_file.read())
print("DEBUG create_video 003 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# other default params, TODO as params
# static
pose_style: int = 0
exp_scale: float = 1.0
use_ref_video: bool = False
ref_video: str = None
input_yaw_list: list = None # [5, -1, 1]
input_pitch_list: list = None # [-3, 2, 0]
input_roll_list: list = None # [1, -1, 0]
result_dir: str = './results/'
batch_size: int = 32
# fps: float = 16.
interpolate_dir: str = './InterpolatedVideo/'
print("DEBUG create_video 004 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
resultVideoFilePath = sad_talker.test(imageFilePath, audioFilePath, preprocess=preprocess, still_mode=still_mode,
use_enhancer=use_enhancer, result_dir=result_dir, exp_scale=exp_scale,
input_yaw_list=input_yaw_list, input_pitch_list=input_pitch_list,
input_roll_list=input_roll_list, batch_size=batch_size, fps=fps)
print("DEBUG create_video 005 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# Interpolate Video
InterpolateVideoFilePath = interpolate_dir + str(currentTime) + '.mp4'
interpolate_video(resultVideoFilePath, InterpolateVideoFilePath, factor=interpolate_factor)
# def generate_video():
# with open(resultVideoFilePath, mode="rb") as file_like:
# print("DEBUG create_video 006 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# yield from file_like
# print("DEBUG create_video 007 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
def generate_video():
with open(InterpolateVideoFilePath, mode="rb") as file_like:
print("DEBUG create_video 006 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
yield from file_like
print("DEBUG create_video 007 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
print("DEBUG create_video 008 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# 返回生成的视频
return StreamingResponse(generate_video(), media_type="video/mp4")
# 场景2：输入 语音+图片 ， 流式得到 多个mp4视频片段 目录,
# TODO 当然可以改造返回逐个mp4视频流，暂时未调通客户端怎么播放这些分配数据，先返回mp4片段文件名
# @app.post("/create_video_for_flow")
# async def create_video_for_flow(image_file: UploadFile = File(...), audio_file: UploadFile = File(...)):
# print("DEBUG create_video_for_flow 001 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# global sad_talker
#
# print("DEBUG create_video_for_flow 002 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# # user input files
# imageFilePath = f"./image_uploaded_files/{image_file.filename}"
# audioFilePath = f"./audio_uploaded_files/{audio_file.filename}"
#
# # save image
# with open(imageFilePath, "wb") as f:
# f.write(await image_file.read())
# with open(audioFilePath, "wb") as f:
# f.write(await audio_file.read())
#
# print("DEBUG create_video_for_flow 003 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# # other default params, TODO as params
# preprocess: str = 'crop'
# still_mode: bool = False
# use_enhancer: bool = False
# result_dir: str = './results/'
#
# print("DEBUG create_video_for_flow 004 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# resultVideoSegmentPath = sad_talker.test_flow(imageFilePath, audioFilePath, preprocess=preprocess, still_mode=still_mode,
# use_enhancer=use_enhancer, result_dir=result_dir)
#
# print("DEBUG create_video_for_flow 005")
# #
# def generate_video_fragments():
# for segFilePath in resultVideoSegmentPath:
# print("DEBUG create_video_for_flow 006 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# yield segFilePath + "\n"
# print("DEBUG create_video_for_flow 007 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
#
# print("DEBUG create_video_for_flow 008 " + datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"))
# # 返回生成的视频
# return StreamingResponse(generate_video_fragments(), media_type="text/plain")
# server demo
if __name__ == "__main__":
sad_talker_crop = SadTalker(lazy_load=False, preprocess='crop')
sad_talker_full = SadTalker(lazy_load=False, preprocess='full')
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=9009, workers=1)
# client demo: create_video
# curl -X POST -F "image_file=@/Users/weizy/PycharmProjects/pythonProject/media/art_5_256x256.png" -F "audio_file=@/Users/weizy/PycharmProjects/pythonProject/media/RD_Radio31_000.wav" http://localhost:9009/create_video -o output_test.mp4
# curl -X POST -F "audio_file=@/home/redhat/AiModels/SadTalker/examples/driven_audio/RD_Radio31_000.wav" -F "image_file=@/home/redhat/AiModels/SadTalker/examples/source_image/art_5.png" http://localhost:8008/create_video -o output_test.mp4
# client demo: create_video_for_flow
# curl -X POST -F "image_file=@/Users/weizy/PycharmProjects/pythonProject/media/art_5_256x256.png" -F "audio_file=@/Users/weizy/PycharmProjects/pythonProject/media/RD_Radio31_000.wav" http://localhost:8000/create_video_for_flow
# curl -X POST -F "audio_file=@/home/redhat/AiModels/SadTalker/examples/driven_audio/RD_Radio31_000.wav" -F "image_file=@/home/redhat/AiModels/SadTalker/examples/source_image/art_5.png" http://localhost:8000/create_video_for_flow