Download audio from answers.com

lourencojunior · 06-03-2008, 02:56 PM

Hallo everyone!!

I would like to download audio files from http://answers.com/ automatically.

Website usage:

Code:

  http://answers.com/choir
  http://answers.com/possible
  http://answers.com/border

I am using the following commands:

Code:

wget http://answers.com/choir -O choir.html # save a local copy 
cat choir.html | grep wav # extract line where wav file link can be found

At this point, I get an output as follows (a single line):

Code:

<h1>choir</h1>&nbsp;&nbsp;(<span style="color:blue;" class="pointer" onclick="pw = window.open('http://content.answers.com/main/content/pronkey-answers.html', 'PronunciationKey', 'height=650,width=520,resizable,scrollbars');if(pw){pw.focus();}" onmouseout="status='';return true;" onmouseover="status='Click for pronunciation key';return true;"><span class="pron">kwīr</span></span>) <span style="cursor:pointer" onmouseover="status='Click to hear pronunciation';return true;" onmouseout="status='';return true;" onclick="playIt('http://content.answers.com/main/content/ahd4/pron/C0317300.wav')"><img border="0" align="middle" src="http://content.answers.com/main/content/img/pron.gif" alt="pronunciation" /></span><br />

My doubt:
- How could I get `http://content.answers.com/main/cont...n/C0317300.wav' as result of a Regular Expression? Thus, I could use wget to download the wav file.

- Is (are) there another (other) way?

As result I would like get a pronunciation file using a command like:
[code]
pron choir
[\code]

What I've coded until now:

Code:

#! /bin/bash

wget http://answers.com/$1 -O $1.html # save a local copy 
cat $1.html | grep wav # extract line where wav file link can be found

## do something to get audio url

wget $(WAV) -O $1.wav
rm -f $1.html

Any help is valid!

Thanks in advance.
Regards,
Lourenco.

Agrouf · 06-04-2008, 08:12 AM

sed "s/.*($.*$).*/\1/g"

lourencojunior · 06-04-2008, 01:27 PM

Thank you Agrouf!

Such Regular Expression solve my problem. Now, I've got the solution.

To get a local copy of wav file:
[code]
#!/bin/bash

# TODO:
# - Verify whether $1.wav already exists. Case yes play, otherwise wget.
# - Improve first sed Regular Expression for deleting \' char.

# getting a local copy of page
wget http://answers.com/$1 -O $1.html
# extracting wav audio URL
SOUND=$(cat $1.html | grep wav | sed "s/.*($.*$).*/\1/g" | sed "s/'//g")
# saving a local copy
wget $SOUND -O $1.wav
# removing html
rm -f $1.html
# playing with a player
mplayer $1.wav
[\code]

I've also found another method (many thanks to a friend, Marcelo), which there's no need a local copy of the wav file:

Code:

#!/bin/bash

# getting a local copy of page
wget http://answers.com/$1 -O $1.html
# preparing mplayer input
SOUND=$(grep -o "playIt.*http.*wav" $1.html | sed "s/playIt('//")
# removing html
rm -f $1.html
# playing with a player
mplayer $SOUND

Best regards,
Lourenco

lourencojunior · 06-05-2008, 11:17 AM

Improved version:

Code:

#!/bin/bash

# Checking input params
if (( $# != 1 )); then
	NAME=$(basename $0)
	echo "Usage:"
	echo "  $NAME <word>"
	echo "Where:"
	echo "  <word> = a English word to get pronunciation audio file"
	echo "Return values:"
	echo "  0: Sucess"
	echo "  1: Word not found"
	echo "  2: Invalid params"
	exit 2
fi

# Where cache files will be stored
CACHE_DIR=~ljr/tmp/answers.com/
# File name
FILENAME=$CACHE_DIR$1.wav

# If there's no cache, download it
if ! test -f $FILENAME; then

	# Getting audio file URL
	SOUND=$(lynx -source http://www.answers.com/$1 | sed -n "/.*playIt.*('\(.*\.wav\).*/{s//\1/p;q;}")

	# Checking if it was found
	if test -z "$SOUND"; then
		echo "$1: word not found"
		exit 1
	fi

	# Downloading for caching
	wget $SOUND -O $FILENAME

fi

# Speak to me, Dorothy ;-)
play $FILENAME

# Signaling success
exit 0


# END OF FILE

Best regards,
Lourenco.