Spaces:
Sleeping
Sleeping
modularized code
Browse files- Dockerfile +0 -2
- app.py +87 -54
- requirements.txt +3 -2
- utlis.py +34 -1
Dockerfile
CHANGED
|
@@ -6,8 +6,6 @@ COPY requirements.txt .
|
|
| 6 |
|
| 7 |
RUN pip install --no-cache-dir -r requirements.txt
|
| 8 |
|
| 9 |
-
RUN apt-get update && apt-get install -y ffmpeg
|
| 10 |
-
|
| 11 |
ENV MPLCONFIGDIR /tmp/matplotlib
|
| 12 |
|
| 13 |
RUN mkdir /app/results_vc
|
|
|
|
| 6 |
|
| 7 |
RUN pip install --no-cache-dir -r requirements.txt
|
| 8 |
|
|
|
|
|
|
|
| 9 |
ENV MPLCONFIGDIR /tmp/matplotlib
|
| 10 |
|
| 11 |
RUN mkdir /app/results_vc
|
app.py
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
import os
|
|
|
|
| 2 |
import threading
|
| 3 |
import time
|
| 4 |
from pathlib import Path
|
| 5 |
|
| 6 |
import gradio as gr
|
| 7 |
|
| 8 |
-
from utlis import
|
| 9 |
-
table_to_ELAN_tsv, trim_media)
|
| 10 |
|
| 11 |
|
| 12 |
def delete_files(files):
|
|
@@ -18,62 +18,102 @@ def delete_files(files):
|
|
| 18 |
pass
|
| 19 |
print("files deleted")
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
def trim_video_wt(input_file, input_transcript, output_format, start_time, end_time):
|
| 22 |
# trim video with transcript
|
| 23 |
try:
|
| 24 |
-
#
|
| 25 |
output_folder = f"{os.getcwd()}/results_wt/"
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
output_file = os.path.join(output_folder, file_name)
|
| 30 |
-
print(f"Output folder: {output_folder}")
|
| 31 |
-
print(f"file_name: {file_name}")
|
| 32 |
-
print(f"Output file: {output_file}")
|
| 33 |
-
if start_time == "":
|
| 34 |
-
start_time = "00:00:00"
|
| 35 |
-
if end_time == "":
|
| 36 |
-
end_time = "00:05:00"
|
| 37 |
# Trim the video
|
| 38 |
-
|
| 39 |
-
output_file = trim_media(input_file.name, output_file, start_time, end_time)
|
| 40 |
-
|
| 41 |
-
# trim transcript
|
| 42 |
-
print("start trimming transcript")
|
| 43 |
-
table = molly_xlsx_to_table(xl_file=input_transcript.name)
|
| 44 |
-
transcript_name = f"{Path(input_transcript.name).stem.partition('.')[0]}_trimmed.tsv"
|
| 45 |
-
transcript_file = os.path.join(output_folder, transcript_name)
|
| 46 |
-
table_to_ELAN_tsv(table, transcript_file)
|
| 47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
# remove file after 10 minutes for security
|
| 49 |
print("start deleting files")
|
| 50 |
-
path_to_delete = [input_file.name, input_transcript.name, output_file,
|
| 51 |
threading.Thread(target=delete_files, args=([path_to_delete])).start()
|
| 52 |
|
| 53 |
-
return output_file,
|
| 54 |
except Exception as e:
|
| 55 |
return f"Error: {str(e)}"
|
| 56 |
|
| 57 |
def trim_video(input_file, output_format, start_time, end_time):
|
| 58 |
try:
|
| 59 |
-
# Set output file
|
| 60 |
output_folder = f"{os.getcwd()}/results_vt/"
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
file_name = f"{Path(input_file.name).stem.partition('.')[0]}_trimmed.{output_format}"
|
| 64 |
-
output_file = os.path.join(output_folder, file_name)
|
| 65 |
-
print(f"Output folder: {output_folder}")
|
| 66 |
-
print(f"file_name: {file_name}")
|
| 67 |
-
print(f"Output file: {output_file}")
|
| 68 |
-
if start_time == "":
|
| 69 |
-
start_time = "00:00:00"
|
| 70 |
-
if end_time == "":
|
| 71 |
-
end_time = "00:05:00"
|
| 72 |
# Trim the video
|
| 73 |
-
|
| 74 |
-
output_file = trim_media(input_file.name, output_file, start_time, end_time)
|
| 75 |
|
| 76 |
-
#
|
| 77 |
print("start deleting files")
|
| 78 |
path_to_delete = [input_file.name, output_file]
|
| 79 |
threading.Thread(target=delete_files, args=([path_to_delete])).start()
|
|
@@ -84,19 +124,13 @@ def trim_video(input_file, output_format, start_time, end_time):
|
|
| 84 |
|
| 85 |
def convert_video(input_file, output_format):
|
| 86 |
try:
|
| 87 |
-
# Set output file
|
| 88 |
-
output_folder = f"{os.getcwd()}/
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
output_file =
|
| 93 |
-
|
| 94 |
-
print(f"file_name: {file_name}")
|
| 95 |
-
print(f"Output file: {output_file}")
|
| 96 |
-
# Convert the video
|
| 97 |
-
print("start converting")
|
| 98 |
-
output_file = convert_video_format(input_file.name, output_file)
|
| 99 |
-
|
| 100 |
# remove file after 10 minutes for security
|
| 101 |
print("start deleting files")
|
| 102 |
path_to_delete = [input_file.name, output_file]
|
|
@@ -121,7 +155,6 @@ interface_wt = gr.Interface(fn=trim_video_wt, inputs=[input_file_wt, input_trans
|
|
| 121 |
|
| 122 |
# gr components for video trimmer
|
| 123 |
input_file = gr.File(label="Select video file")
|
| 124 |
-
# input_file.upload(show_progress=True)
|
| 125 |
output_format = gr.Dropdown(choices=["mkv", "MOV", "mp4", "wav"], label="Select output format", value="mp4", )
|
| 126 |
start_time = gr.Textbox(label="Start time (in seconds or HH:MM:SS). Leave blank to start at beginning.")
|
| 127 |
end_time = gr.Textbox(label="End time (in seconds or HH:MM:SS). Leave blank to trim a 5-minute interval since start.")
|
|
|
|
| 1 |
import os
|
| 2 |
+
import re
|
| 3 |
import threading
|
| 4 |
import time
|
| 5 |
from pathlib import Path
|
| 6 |
|
| 7 |
import gradio as gr
|
| 8 |
|
| 9 |
+
from utlis import HHMMSS_to_sec, convert_video_format, molly_xlsx_to_table, xlsx_to_table, table_to_ELAN_tsv, trim_media
|
|
|
|
| 10 |
|
| 11 |
|
| 12 |
def delete_files(files):
|
|
|
|
| 18 |
pass
|
| 19 |
print("files deleted")
|
| 20 |
|
| 21 |
+
def classify_input_format(input_string):
|
| 22 |
+
# check if the input string is empty
|
| 23 |
+
if not input_string:
|
| 24 |
+
return 0
|
| 25 |
+
|
| 26 |
+
# check if the input string is a valid time in the format HH:MM:SS
|
| 27 |
+
hhmmss_pattern = re.compile('^\d{1,2}:\d{1,2}:\d{1,2}$')
|
| 28 |
+
if hhmmss_pattern.match(input_string):
|
| 29 |
+
return HHMMSS_to_sec(time_str=input_string)
|
| 30 |
+
|
| 31 |
+
# check if the input string is a valid number in the format of seconds
|
| 32 |
+
try:
|
| 33 |
+
seconds = float(input_string)
|
| 34 |
+
return float(input_string)
|
| 35 |
+
except ValueError:
|
| 36 |
+
pass
|
| 37 |
+
|
| 38 |
+
# if the input format is neither seconds nor HH:MM:SS, return None
|
| 39 |
+
return None
|
| 40 |
+
|
| 41 |
+
def set_output_file(input_file, output_format, folder):
|
| 42 |
+
# Set output file name and extension
|
| 43 |
+
if not os.path.exists(folder):
|
| 44 |
+
os.makedirs(folder)
|
| 45 |
+
file_name = f"{Path(input_file.name).stem.partition('.')[0]}_trimmed.{output_format}"
|
| 46 |
+
output_file = os.path.join(folder, file_name)
|
| 47 |
+
print(f"Output folder: {folder}")
|
| 48 |
+
print(f"file_name: {file_name}")
|
| 49 |
+
print(f"Output file: {output_file}")
|
| 50 |
+
return output_file
|
| 51 |
+
|
| 52 |
+
def trim_video_helper(input_file, output_file, start_time, end_time):
|
| 53 |
+
if start_time == "":
|
| 54 |
+
start_time = 0
|
| 55 |
+
if end_time == "":
|
| 56 |
+
end_time = 300 + HHMMSS_to_sec(time_str=start_time)
|
| 57 |
+
# Trim the video
|
| 58 |
+
print("start trimming")
|
| 59 |
+
start_time = classify_input_format(start_time)
|
| 60 |
+
end_time = classify_input_format(end_time)
|
| 61 |
+
print("start time: ", start_time)
|
| 62 |
+
print("end time: ", end_time)
|
| 63 |
+
output_file = trim_media(input_file.name, output_file, start_time, end_time)
|
| 64 |
+
print("finish trimming")
|
| 65 |
+
return output_file
|
| 66 |
+
|
| 67 |
+
def convert_video_helper(input_file, output_file, output_format):
|
| 68 |
+
# convert video
|
| 69 |
+
print("start converting")
|
| 70 |
+
output_file = convert_video_format(input_file.name, output_file)
|
| 71 |
+
print("finish converting")
|
| 72 |
+
return output_file
|
| 73 |
+
|
| 74 |
+
def convert_transcript_helper(input_transcript, output_transcript):
|
| 75 |
+
# convert transcript
|
| 76 |
+
print("start converting transcript")
|
| 77 |
+
table = xlsx_to_table(xl_file=input_transcript)
|
| 78 |
+
print("finish converting transcript to table")
|
| 79 |
+
output_file = table_to_ELAN_tsv(table, output_transcript)
|
| 80 |
+
print("finish converting transcript")
|
| 81 |
+
return output_file
|
| 82 |
+
|
| 83 |
def trim_video_wt(input_file, input_transcript, output_format, start_time, end_time):
|
| 84 |
# trim video with transcript
|
| 85 |
try:
|
| 86 |
+
# set output file
|
| 87 |
output_folder = f"{os.getcwd()}/results_wt/"
|
| 88 |
+
output_file = set_output_file(input_file, output_format, output_folder)
|
| 89 |
+
output_transcript = set_output_file(input_transcript, "tsv", output_folder)
|
| 90 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 91 |
# Trim the video
|
| 92 |
+
output_file = trim_video_helper(input_file, output_file, start_time, end_time)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
|
| 94 |
+
# convert transcript
|
| 95 |
+
path = input_transcript.name
|
| 96 |
+
output_transcript = convert_transcript_helper(path, output_transcript)
|
| 97 |
+
# output_transcript = output_file
|
| 98 |
# remove file after 10 minutes for security
|
| 99 |
print("start deleting files")
|
| 100 |
+
path_to_delete = [input_file.name, input_transcript.name, output_file, output_transcript]
|
| 101 |
threading.Thread(target=delete_files, args=([path_to_delete])).start()
|
| 102 |
|
| 103 |
+
return output_file, output_transcript
|
| 104 |
except Exception as e:
|
| 105 |
return f"Error: {str(e)}"
|
| 106 |
|
| 107 |
def trim_video(input_file, output_format, start_time, end_time):
|
| 108 |
try:
|
| 109 |
+
# Set output file
|
| 110 |
output_folder = f"{os.getcwd()}/results_vt/"
|
| 111 |
+
output_file = set_output_file(input_file, output_format, output_folder)
|
| 112 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 113 |
# Trim the video
|
| 114 |
+
output_file = trim_video_helper(input_file, output_file, start_time, end_time)
|
|
|
|
| 115 |
|
| 116 |
+
# Remove files after 10 minutes for security
|
| 117 |
print("start deleting files")
|
| 118 |
path_to_delete = [input_file.name, output_file]
|
| 119 |
threading.Thread(target=delete_files, args=([path_to_delete])).start()
|
|
|
|
| 124 |
|
| 125 |
def convert_video(input_file, output_format):
|
| 126 |
try:
|
| 127 |
+
# Set output file
|
| 128 |
+
output_folder = f"{os.getcwd()}/results_cv/"
|
| 129 |
+
output_file = set_output_file(input_file, output_format, output_folder)
|
| 130 |
+
|
| 131 |
+
# Convert video
|
| 132 |
+
output_file = convert_video_helper(input_file, output_file, output_format)
|
| 133 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 134 |
# remove file after 10 minutes for security
|
| 135 |
print("start deleting files")
|
| 136 |
path_to_delete = [input_file.name, output_file]
|
|
|
|
| 155 |
|
| 156 |
# gr components for video trimmer
|
| 157 |
input_file = gr.File(label="Select video file")
|
|
|
|
| 158 |
output_format = gr.Dropdown(choices=["mkv", "MOV", "mp4", "wav"], label="Select output format", value="mp4", )
|
| 159 |
start_time = gr.Textbox(label="Start time (in seconds or HH:MM:SS). Leave blank to start at beginning.")
|
| 160 |
end_time = gr.Textbox(label="End time (in seconds or HH:MM:SS). Leave blank to trim a 5-minute interval since start.")
|
requirements.txt
CHANGED
|
@@ -1,3 +1,4 @@
|
|
| 1 |
-
gradio==3.
|
| 2 |
moviepy==1.0.3
|
| 3 |
-
pandas==1.1.3
|
|
|
|
|
|
| 1 |
+
gradio==3.24.1
|
| 2 |
moviepy==1.0.3
|
| 3 |
+
pandas==1.1.3
|
| 4 |
+
xlrd==1.2.0
|
utlis.py
CHANGED
|
@@ -140,6 +140,39 @@ def molly_xlsx_to_table(xl_file):
|
|
| 140 |
|
| 141 |
return table
|
| 142 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 143 |
def table_to_ELAN_tsv(table:pd.DataFrame, path:str):
|
| 144 |
# write table to tsv compatible with ELAN import
|
| 145 |
-
table.to_csv(path, index=False, float_format='%.3f',sep='\t')
|
|
|
|
|
|
| 140 |
|
| 141 |
return table
|
| 142 |
|
| 143 |
+
def xlsx_to_table(xl_file):
|
| 144 |
+
# read the first sheet of the Excel file into a DataFrame
|
| 145 |
+
print(f'...reading {xl_file}...')
|
| 146 |
+
breakpoint()
|
| 147 |
+
table = pd.read_excel(xl_file, sheet_name=0)
|
| 148 |
+
print(f'...done reading {xl_file}...')
|
| 149 |
+
# extract start and end time from the Timecode column
|
| 150 |
+
print(f'...splitting Timecode column into start and end time...')
|
| 151 |
+
timecodes = table['Timecode'].str.split(' - ', expand=True)
|
| 152 |
+
table['start_time'] = timecodes[0]
|
| 153 |
+
table['end_time'] = timecodes[1]
|
| 154 |
+
print(f'...done splitting Timecode column into start and end time...')
|
| 155 |
+
# convert start and end time to seconds using the HHMMSS_to_sec function
|
| 156 |
+
print(f'...converting start and end time to seconds...')
|
| 157 |
+
table['start_sec'] = table['start_time'].apply(HHMMSS_to_sec)
|
| 158 |
+
table['end_sec'] = table['end_time'].apply(HHMMSS_to_sec)
|
| 159 |
+
print(f'...done converting start and end time to seconds...')
|
| 160 |
+
# drop unnecessary columns
|
| 161 |
+
print(f'...dropping unnecessary columns...')
|
| 162 |
+
table.drop(['Timecode', 'Annotations', 'Error Type', 'Duration'], axis=1, inplace=True)
|
| 163 |
+
|
| 164 |
+
# rename columns
|
| 165 |
+
print(f'...renaming columns...')
|
| 166 |
+
table.rename(columns={'#': 'uttID', 'Speaker': 'speaker', 'Dialogue': 'transcript'}, inplace=True)
|
| 167 |
+
|
| 168 |
+
# reorder columns
|
| 169 |
+
print(f'...reordering columns...')
|
| 170 |
+
table = table[['uttID', 'speaker', 'transcript', 'start_sec', 'end_sec']]
|
| 171 |
+
|
| 172 |
+
return table
|
| 173 |
+
|
| 174 |
+
|
| 175 |
def table_to_ELAN_tsv(table:pd.DataFrame, path:str):
|
| 176 |
# write table to tsv compatible with ELAN import
|
| 177 |
+
table.to_csv(path, index=False, float_format='%.3f',sep='\t')
|
| 178 |
+
return path
|