How can I do this

[hmehta]

I have text of this format throughout the file

<audio src=&quot;../prompts/say_name_you_want_to.wav&quot;>Please say the skill code </audio>

What I need to do is create a file with mapping between .wav file and the text above so the mapping in this case would be :

say_name_you_want_to.wav Please say the skill code

Essentially wherever I say .wav file, I need to put the correspoding text whihc would be between
<audio.....> and /<audio>

How can I do this for all files under one dir?

 

[Grant]

Hi hmehta,

There are numerous problems with doing this in awk. The main issues are that there is often no control over whether this type of code is 100% consistent in use of upper- and lower-cases, the <audio...>/</audio> tags may be split over multiple lines, and special characters may cause confusion if they happen to occur inside text areas and strings. So, be warned that this a dubious proposition at best. Which is why I would rather do it in a language like Perl, where there are numerous HTML and XML parsing tools to draw on.

Having said that, I still would like to take a crack at it.

The following is a sketch of how I would try it if I were constrained to do this in awk. (I would much prefer to do it in Perl). Please bear in mind that I wrote this at home, without access to awk -- not even an awk manual -- so nothing is tested, and even the syntax should be double- and triple-checked. Try to extract the idea, and re-write it your own way.


/<audio src=.*wav.>/{
line=$0;

# In case <audio>...</audio> crosses many lines,
# append the lines together.
while (match(/<\/audio>/, $0) == 0)
{
getline;
line=line $0;
}

# Get the wave.
match(/audio src=.[^>]*/, line);
wave=substr(line,RSTART+12,RLENGTH-12);

# Simplify: reduce the 'line' to what is essential.
line=substr(line, RSTART+RLENGTH+1);

# Get the text.
match(/.*<\/audio>/, line);
text=substr(line,RSTART+1,RLENGTH-8);

printf(&quot;wav: %s text: %s\n&quot;,wave, text);
}


I hope this gives you the starting point to achieve your goal, and that you are able to straighten out whatever mistakes I have made in syntax, etc.

Good luck,
Grant.

 

[Grant]

Hi hmehta,

I am at work and had a free moment, so I quickly checked the code I suggested above in my previous post. As promised, there were some mistakes, so I made some changes and came up with the following:

% cat wavetext.awk
#!/usr/bin/awk -f
/<audio src=.*wav.>/ {
line=$0;

while ( match($0, /<\/audio>/) == 0)
{
getline;
line=line $0;
}

match(line, /audio src=.[^>]*/);
wave=substr(line,RSTART+11,RLENGTH-12);

line=substr(line, RSTART+RLENGTH+1);

match(line, /.*<\/audio>/);
text=substr(line,RSTART,RLENGTH-8);

printf(&quot;wav: %s text: %s\n&quot;,wave, text);
}


I created a simple file, to test:
% cat mywave

<audio src=&quot;../prompts/say_name_you_want_to.wav&quot;>Please say the skill code </audio>
<audio src=&quot;../mydir/something_else.wav&quot;>Jitterbug</audio>


I ran it and the result is:
wav: ../prompts/say_name_you_want_to.wav text: Please say the skill code
wav: ../mydir/something_else.wav text: Jitterbug


It passed an easy test. That's nice, but it would definitely fail tougher tests. For instance, if there were multiple <audio src=...> tags on the same line. It might have trouble if there were other tags on the same line, and so on. As long as you have a very simple file with no surprises, this might work for you, but I can't vouch for it if used on a file that presents any complications.

Good luck,
Grant.