[oS-en] sed question

27 Apr 2023

      -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I have a script (from somebody else) to download the image of the day from 
the NASA. I have modified the script quite a bit to download a range of 
days, even years of them.

One key function is failing on today's picture. This is the relevant 
section of the script:

+++......................
function get_page {
     echo "Downloading page to find image"
     #wget http://apod.nasa.gov/apod/ --quiet -O /tmp/apod.html
     wget https://apod.nasa.gov/apod/ap${SHORT_TODAY}.html --quiet -O /tmp/apod.html

     grep -m 1 jpg /tmp/apod.html | sed -e 's/<//' -e 's/>//' -e 's/.*=//' -e 's/"//g' -e 's/^/http:\/\/apod.nasa.gov\/apod\//' > /tmp/pic_url
}

     get_page

     # Got the link to the image
     PICURL=`/bin/cat /tmp/pic_url`
     if [ -z $PICURL ]; then
         echo "Not found picture.jpg for $TODAY. Trying .png"
         try_png
         PICURL=`/bin/cat /tmp/pic_url`
         if [ -z $PICURL ]; then
             echo "Not found picture.png for $TODAY. html file saved"
             cp /tmp/apod.html $PICTURES_DIR/$TODAY.html
             return
         fi
     fi

     NAME=`basename $PICURL`
     PICTURE_NAME=${TODAY}_${NAME}
......................++-

The script sets up variable SHORT_TODAY to 230427, so "get_page" downloads 
<https://apod.nasa.gov/apod/ap230426.html>. It is coughing up this error:

   Do 20230427
   Downloading page to find image
   /home/cer/bin/NASA Picture-Of-The-Day Wallpaper Script, mine, loop: line 98: [: too many arguments
   basename: extra operand ‘Tarantula</a>’
   Try 'basename --help' for more information.

The problem is here:

   cer@Isengard:~> grep -m 1 jpg /tmp/apod.html | sed -e 's/<//' -e 's/>//' -e 's/.*=//' -e 's/"//g' -e 's/^/http:\/\/apod.nasa.gov\/apod\//'
   http://apod.nasa.gov/apod/image/1602/Tarantula-HST-ESO-annotated1800.jpgArou... the Tarantula</a>
   cer@Isengard:~>

The source html goes like this:

   <a href="image/1602/Tarantula-HST-ESO-annotated1800.jpg">Around the Tarantula</a>
   are other star forming regions with
   young star clusters, filaments, and blown-out
   <a href="ap080327.html">bubble-shaped</a> clouds.

On the day before, it goes like this (here the script works as expected):

   2023 April 26
   <br>
    <a href="image/2304/MoonArc_zanarello_1365.jpg">
   <IMG SRC="image/2304/MoonArc_zanarello_960.jpg"
   alt="A nearly full Moon is seen through the famous Arc de Triomphi
   with trees and cars lining the foreground.
   Please see the explanation for more detailed information."
   style="max-width:100%"></a>
   </center>

where the function generates this text:

   cer@Isengard:~> echo '<a href="image/2304/MoonArc_zanarello_1365.jpg">'  | grep -m 1 jpg | sed -e 's/<//' -e 's/>//' -e 's/.*=//' -e 's/"//g' -e 's/^/http:\/\/apod.nasa.gov\/apod\//'
   http://apod.nasa.gov/apod/image/2304/MoonArc_zanarello_1365.jpg
   cer@Isengard:~>

My problem is, I don't talk sed O:-)

So, I ask here if someone knows how to change that sed command so that it
works ;-)

(the full script I modified is: https://paste.opensuse.org/7f484bb7c807)

- -- 
Cheers

 	Carlos E. R.
 	(from 15.4 x86_64 at Telcontar)
-----BEGIN PGP SIGNATURE-----

iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCZErjCRwccm9iaW4ubGlz
dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfV7RYAnR1jKw8AXonj2E837r7L
cY3g08VvAJ418VFrG/lGU3gad7PPUwFceZIfIg==
=KHba
-----END PGP SIGNATURE-----