michael/dotfiles
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Support a time range in the transcribe video scripts

Browse Source
This commit is contained in:
Michael Campagnaro 2023-08-07 17:22:07 -04:00
parent 90c312e954
commit ba9a776867
15 changed files with 201 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
.aliases
View File

@ -354,36 +354,51 @@ function my_transcribe_video() {
file="$1"
output="$2"
include_small=$3
start_time="$4"
end_time="$5"
if [[ $file == "" ]]; then
error "Usage: <input video> <optional output name>"
error "Usage: <input video> <optional output name> <optional include small model (1 or 0)> <optional start time> <optional end time>"
return
fi
if [[ $output == "" ]]; then
output="${1%.*}" # just use the input name without the extension.
fi
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
# Tiny is fast and semi-accurate, so whatever.
# Base is pretty good overall. It has good punctuation inserting and
# catches most words. Small and medium can do better word detection at
# times, but suffer from bad punctuation. Medium is particularly bad and
# not adding commas and periods.
if [[ $include_small -eq 1 ]]; then
transcribe-video "$file" "$output" tiny base small
transcribe-video "$file" "$output" $start_time $end_time tiny base small
else
transcribe-video "$file" "$output" tiny base
transcribe-video "$file" "$output" $start_time $end_time tiny base
fi
}
function my_transcribe_video_all_models() {
file="$1"
output="$2"
start_time="$3"
end_time="$4"
if [[ $file == "" ]]; then
error "Usage: <input video> <optional output name>"
error "Usage: <input video> <optional output name> <optional start time> <optional end time>"
return
fi
if [[ $output == "" ]]; then
output="${1%.*}" # just use the input name without the extension.
fi
transcribe-video "$file" "$output" tiny base small medium
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
transcribe-video "$file" "$output" $start_time $end_time tiny base small medium
}

2
dotfiles/bin/compress-video-with-crf
View File

@ -26,7 +26,7 @@ fi
use_gpu=0
if [[ "$#" < 3 || "$#" > 6 ]]; then
printf "${BOLD}${RED}Usage: $0 <crf value> <filename> <output name> <optional: use-gpu (1|0), defaults to $use_gpu> <optional: start time HH:MM:SS> <optional: end time HH:MM:SS>\n\nIf you want to encode a range of CRF values then use -1 as the crf value.${NORMAL}\n"
printf "${BOLD}${RED}Usage: compress-video-with-crf <crf value> <filename> <output name> <optional: use-gpu (1|0), defaults to $use_gpu> <optional: start time HH:MM:SS> <optional: end time HH:MM:SS>\n\nIf you want to encode a range of CRF values then use -1 as the crf value.${NORMAL}\n"
exit 1
fi

2
dotfiles/bin/convert-video-avi-to-mp4
View File

@ -29,7 +29,7 @@ else
fi
if [[ $1 == "" || $2 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <filename> <bitrate, e.g. \"4000k\"> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: convert-video-avi-to-mp4 <filename> <bitrate, e.g. \"4000k\"> <optional output name>${NORMAL}\n"
exit 1
fi

2
dotfiles/bin/convert-video-flv-to-mp4
View File

@ -27,7 +27,7 @@ else
fi
if [[ $1 == "" || $2 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <compress 1|0> <filename> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: convert-video-flv-to-mp4 <compress 1|0> <filename> <optional output name>${NORMAL}\n"
exit 1
fi

2
dotfiles/bin/convert-video-mkv-to-mp4
View File

@ -27,7 +27,7 @@ else
fi
if [[ $1 == "" || $2 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <compress 1|0> <filename> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: convert-video-mkv-to-mp4 <compress 1|0> <filename> <optional output name>${NORMAL}\n"
exit 1
fi

22
dotfiles/bin/extract-audio-from-video
View File

@ -30,9 +30,11 @@ fi
filename=$(basename -- "$1")
format="$2"
start_time="$3"
end_time="$4"
if [[ $1 == "" || $2 == "" ]]; then
printf "${BOLD}${RED}Usage: extract-audio-from-video <filename> <format (mp3, m4a, aac, etc)>${NORMAL}\n"
printf "${BOLD}${RED}Usage: extract-audio-from-video <filename> <format (mp3, m4a, aac, etc)> <optional: start time HH:MM:SS, use empty string or 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
@ -40,9 +42,23 @@ extension="${filename##*.}"
filename="${filename%.*}"
output_name="$filename.$format"
printf "\n${YELLOW}${BOLD}Extracting audio from $filename.$extension | output: $output_name${NORMAL}\n"
timing_args=""
if [[ $start_time != "" ]]; then
timing_args="-ss $start_time "
fi
if [[ $end_time != "" ]]; then
if [[ $start_time == "0" && $end_time == "0" ]]; then
# We treat a start and end with 0 values as no op.
timing_args=""
elif [[ $end_time != "0" ]]; then
# Handle having a start time but end time is set to 0, can just ignore it and it'll use the remainder of the video.
timing_args+="-to $end_time"
fi
fi
ffmpeg -y -stats -loglevel level+error -i "$filename.$extension" -vn -acodec copy "$output_name"
printf "\n${YELLOW}${BOLD}Extracting audio from $filename.$extension | output: $output_name | start: $start_time | end: $end_time${NORMAL}\n"
ffmpeg -y -stats -loglevel level+error $timing_args -i "$filename.$extension" -vn -acodec copy "$output_name"
printf "\n${GREEN}${BOLD}Done extracting audio from $filename.$extension | output name '$output_name'${NORMAL}\n\n"

10
dotfiles/bin/extract-audio-from-video-and-transcode
View File

@ -37,7 +37,7 @@ output_base=$(basename -- "$output_name")
output_extension="${output_base##*.}"
if [[ $filename == "" || $output_extension == "" || $output_extension == $output_base ]]; then
printf "${BOLD}${RED}Usage: extract-audio-from-video-and-transcode <filename> <output name w/ extension> <optional: sample rate. Defaults to 44100> <optional: channel count. Defaults to 2> <optional: start time HH:MM:SS> <optional: end time HH:MM:SS>${NORMAL}\n"
printf "${BOLD}${RED}Usage: extract-audio-from-video-and-transcode <filename> <output name w/ extension> <optional: sample rate. Defaults to 44100> <optional: channel count. Defaults to 2> <optional: start time HH:MM:SS, use empty string or 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
@ -54,7 +54,13 @@ if [[ $start_time != "" ]]; then
timing_args="-ss $start_time "
fi
if [[ $end_time != "" ]]; then
timing_args+="-to $end_time"
if [[ $start_time == "0" && $end_time == "0" ]]; then
# We treat a start and end with 0 values as no op.
timing_args=""
elif [[ $end_time != "0" ]]; then
# Handle having a start time but end time is set to 0, can just ignore it and it'll use the remainder of the video.
timing_args+="-to $end_time"
fi
fi
printf "\n${YELLOW}${BOLD}Extracting audio from '$filename' | output: '$output_name' | sample rate: $sample_rate | channels: $channel_count | start: $start_time | end: $end_time${NORMAL}\n"

2
dotfiles/bin/fix-audio-in-one-channel
View File

@ -27,7 +27,7 @@ else
fi
if [[ $1 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <filename> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: fix-audio-in-one-channel <filename> <optional output name>${NORMAL}\n"
exit 1
fi

2
dotfiles/bin/normalize-video-volume
View File

@ -31,7 +31,7 @@ else
fi
if [[ $1 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <filename> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: normalize-video-volume <filename> <optional output name>${NORMAL}\n"
exit 1
fi

2
dotfiles/bin/remove-audio-from-video
View File

@ -24,7 +24,7 @@ else
fi
if [[ $1 == "" ]]; then
printf "${BOLD}${RED}Usage: $0 <filename> <optional output name>${NORMAL}\n"
printf "${BOLD}${RED}Usage: remove-audio-from-video <filename> <optional output name>${NORMAL}\n"
exit 1
fi

10
dotfiles/bin/transcribe-video
View File

@ -25,11 +25,13 @@ fi
input_video="$1"
output_name_without_ext="$2"
shift 2
start_time="$3"
end_time="$4"
shift 4
models=("$@")
if [[ $input_video == "" || $output_name_without_ext == "" || ${#models[@]} -eq 0 ]]; then
printf "${BOLD}${RED}Usage: transcribe-video <input.mp4> <output name without extension> <list of model names to use>${NORMAL}\n"
if [[ $input_video == "" || $output_name_without_ext == "" || $start_time == "" || $end_time == "" || ${#models[@]} -eq 0 ]]; then
printf "${BOLD}${RED}Usage: transcribe-video <input.mp4> <output name without extension> <start time HH:MM:SS, use 0 for start> <end time HH:MM:SS, use 0 for no value> <list of model names to use>${NORMAL}\n"
exit 1
fi
@ -45,7 +47,7 @@ fi
channel_count=1
sample_rate=16000
extract-audio-from-video-and-transcode "$input_video" "$wav_name" $sample_rate $channel_count
extract-audio-from-video-and-transcode "$input_video" "$wav_name" $sample_rate $channel_count $start_time $end_time
if [[ $? == 1 ]]; then exit 1; fi
for model in "$@"; do

35
dotfiles/bin/transcribe-video-base
View File

@ -1,2 +1,37 @@
#!/usr/bin/env bash
if which tput >/dev/null 2>&1; then
ncolors=$(tput colors)
fi
if [ -t 1 ] && [ -n "$ncolors" ] && [ "$ncolors" -ge 8 ]; then
RED="$(tput setaf 1)"
GREEN="$(tput setaf 2)"
YELLOW="$(tput setaf 3)"
BLUE="$(tput setaf 4)"
MAGENTA="$(tput setaf 5)"
CYAN="$(tput setaf 6)"
BOLD="$(tput bold)"
NORMAL="$(tput sgr0)"
else
RED=""
GREEN=""
YELLOW=""
BLUE=""
MAGENTA=""
CYAN=""
BOLD=""
NORMAL=""
fi
if [[ $2 == "" ]]; then
printf "${BOLD}${RED}Usage: transcribe-video-base <input.mp4> <output name without extension> <optional: start time HH:MM:SS, use 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
start_time="$3"
end_time="$4"
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
transcribe-video "$1" "$2" "base"

35
dotfiles/bin/transcribe-video-medium
View File

@ -1,2 +1,37 @@
#!/usr/bin/env bash
if which tput >/dev/null 2>&1; then
ncolors=$(tput colors)
fi
if [ -t 1 ] && [ -n "$ncolors" ] && [ "$ncolors" -ge 8 ]; then
RED="$(tput setaf 1)"
GREEN="$(tput setaf 2)"
YELLOW="$(tput setaf 3)"
BLUE="$(tput setaf 4)"
MAGENTA="$(tput setaf 5)"
CYAN="$(tput setaf 6)"
BOLD="$(tput bold)"
NORMAL="$(tput sgr0)"
else
RED=""
GREEN=""
YELLOW=""
BLUE=""
MAGENTA=""
CYAN=""
BOLD=""
NORMAL=""
fi
if [[ $2 == "" ]]; then
printf "${BOLD}${RED}Usage: transcribe-video-medium <input.mp4> <output name without extension> <optional: start time HH:MM:SS, use 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
start_time="$3"
end_time="$4"
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
transcribe-video "$1" "$2" "medium"

35
dotfiles/bin/transcribe-video-small
View File

@ -1,2 +1,37 @@
#!/usr/bin/env bash
if which tput >/dev/null 2>&1; then
ncolors=$(tput colors)
fi
if [ -t 1 ] && [ -n "$ncolors" ] && [ "$ncolors" -ge 8 ]; then
RED="$(tput setaf 1)"
GREEN="$(tput setaf 2)"
YELLOW="$(tput setaf 3)"
BLUE="$(tput setaf 4)"
MAGENTA="$(tput setaf 5)"
CYAN="$(tput setaf 6)"
BOLD="$(tput bold)"
NORMAL="$(tput sgr0)"
else
RED=""
GREEN=""
YELLOW=""
BLUE=""
MAGENTA=""
CYAN=""
BOLD=""
NORMAL=""
fi
if [[ $2 == "" ]]; then
printf "${BOLD}${RED}Usage: transcribe-video-small <input.mp4> <output name without extension> <optional: start time HH:MM:SS, use 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
start_time="$3"
end_time="$4"
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
transcribe-video "$1" "$2" "small"

37
dotfiles/bin/transcribe-video-tiny
View File

@ -1,2 +1,37 @@
#!/usr/bin/env bash
transcribe-video "$1" "$2" "tiny"
if which tput >/dev/null 2>&1; then
ncolors=$(tput colors)
fi
if [ -t 1 ] && [ -n "$ncolors" ] && [ "$ncolors" -ge 8 ]; then
RED="$(tput setaf 1)"
GREEN="$(tput setaf 2)"
YELLOW="$(tput setaf 3)"
BLUE="$(tput setaf 4)"
MAGENTA="$(tput setaf 5)"
CYAN="$(tput setaf 6)"
BOLD="$(tput bold)"
NORMAL="$(tput sgr0)"
else
RED=""
GREEN=""
YELLOW=""
BLUE=""
MAGENTA=""
CYAN=""
BOLD=""
NORMAL=""
fi
if [[ $2 == "" ]]; then
printf "${BOLD}${RED}Usage: transcribe-video-tiny <input.mp4> <output name without extension> <optional: start time HH:MM:SS, use 0 for start> <optional: end time HH:MM:SS, use empty string or 0 for no value>${NORMAL}\n"
exit 1
fi
start_time="$3"
end_time="$4"
if [[ $start_time == "" ]]; then start_time="0"; fi
if [[ $end_time == "" ]]; then end_time="0"; fi
transcribe-video "$1" "$2" $start_time $end_time "tiny"