Discussion, talk and tips
https://forum.puppylinux.com/
I don't code much but I'm looking for a way to replace everything between <p> and </p> where the word hello is with something from a text file.
since the text between the paragraph element will change every time, I got stumped
I was looking at the "sed" command initially.
This is the line:
"htmlContent":"<html><head></head><body><p>Hello,</p></p></body></html>"
Thanks. 
@spiritwild, Can you test this please?
Code: Select all
sed -i 's/^<p>=/&something_from_a_text_file /' name_of_the_file_to_modify.htmlMight need some refinement..... Like adding in a <p> in front of something_from_a_text_file
so far it's not doing it...needs some adjustment
This should work in many situations, we replace Hello with OtherWord using awk's general substitution function.
Code: Select all
awk '/<p>/{f=1} f==1{gsub(/Hello/,"OtherWord"); print $0} /<\/p>/{f=0}' filename.html > edited_file.html
f=1 and f=0 are 'on and off switches' we create to differentiate the lines where we want to apply substitutions, there could be a word Hello incrusted in a <H1></H1> tag and perhaps we do not want a substitution there.
In its basic form awk reads line by line a stream and tries to find a pattern to then execute some action or actions, you can have more than one pattern/action, so in the code snippet above we have three pattern/actions:
1st. /<p>/{f=1}, scan for the string "<p>", if you find it, then create and assign f=1
2nd. f==1{gsub(/Hello/,"OtherWord"); print $0}, if the value of f is 1 then substitute Hello for OtherWord, and then print the line
3rd. /<\/p>/{f=0}, scan for the string "</p>", if you find set variable f to 0, f=0
Further reading: https://www.gnu.org/software/gawk/manual/gawk.html
results from terminal,
Code: Select all
root# echo '"htmlContent":"<html><head></head><body><p>Hello,</p><p>Hello</p></body></html>"' | awk '/<p>/{f=1} f==1{gsub(/Hello/,"OtherWord"); print $0} /<\/p>/{f=0}'
"htmlContent":"<html><head></head><body><p>OtherWord,</p><p>OtherWord</p></body></html>"
May not be helpful, but your code isn't valid html.
There is one opening <p> tag but two closing </p> tags.
This will work in a terminal:
Code: Select all
echo 'htmlContent":"<html><head></head><body><p>Hello,</p></p></body></html>' | sed 's/Hello/Goodbye/'or make a test.html file with this in it:
Code: Select all
htmlContent":"<html><head></head><body><p>Hello,</p></p></body></html>
then a script test.sh that is next to test.html
Code: Select all
#!/bin/sh
sed -i 's/Hello/Goodbye/' test.htmlRemove tags:
Code: Select all
# echo "htmlContent":"<html><head></head><body><p>Hello,</p></p></body></html>" | sed 's%<[^>]*>%%g'
htmlContent:Hello,
Remove leading text and tags:
Code: Select all
# echo "htmlContent":"<html><head></head><body><p>Hello,</p></p></body></html>" | sed 's%.*:%%;s%<[^>]*>%%g'
Hello,
Will probably not work with other than the provided code...
@spiritwild
You have to be more precise on the searchpattern...
Quick link for basic sed commands: http://sed.sourceforge.net/sed1line.txt
This type of question is good start point for experimentation to learn how to do this.
On a php-html file I ran a very simple search for just Hello and replace it with Goodbye
Just to use some fill I am using this php-html that takes 3 camera streams from Zoneminder and displays them on a web page. I added in a fake PHP section to test out finding the pattern "Hello" and swapping it out.
Test-SED.php
Code: Select all
<!DOCTYPE html>
<head>
<meta name="viewport" content="width=device-width, initial-scale=1">
<style type='text/css'>
body {background-color: black;}
#containercams {
text-align: center;
}
#containercams > div {
display: inline-block;
vertical-align: top;
/* IE fix. */
*display: inline;
zoom: 1;
}
#containercams:after {
content: "";
width: 100%;
display: inline-block;
}
</style>
</head>
<?php
$content_string = 'htmlContent":"<html><head></head><body><p>Hello,</p></body></html>';
?>
<body>
<div id=containercams>
<div><img class="alignnone" src="http://localhost/cgi-bin/nph-zms?mode=jpeg&maxfps=&scale=75&state=1&monitor=1&auth=a732c9fb04ac5339208cdeedc9bbf94b&rand=1656735805&connkey=704099" alt="stream down" width="300" height="220"></div>
<div><img class="alignnone" src="http://localhost/cgi-bin/nph-zms?mode=jpeg&maxfps=&scale=75&state=1&monitor=2&auth=a732c9fb04ac5339208cdeedc9bbf94b&rand=1656735862&connkey=324859" alt="stream down" width="300" height="220"></div>
<div><img class="alignnone" src="http://localhost/cgi-bin/nph-zms?mode=jpeg&maxfps=&scale=100&state=1&monitor=3&auth=a732c9fb04ac5339208cdeedc9bbf94b&rand=1656735913&connkey=61446" alt="stream down...." width="300" height="220"></div>
</div>
</body>The script:
test.sh
Code: Select all
#!/bin/sh
sed -i 's|Hello|Sometimes you want to swap out a pattern|' Test-SED.phpThis should work iff:
each <p>...</p> pair is always on the same line (it's possible to work around this if you need)
<p>...</p> doesn't contain any html tag inside (any "<" actually)
In the replacement text from the something.txt, newline chars will be replaced with spaces.
Code: Select all
text=$(sed ':a;$!{N;ba};{s|[\/&]|\\&|g;s/\n/ /g}' something.txt)
sed -i "s/<p>[^<]*hello[^<]*<\/p>/<p>${text}<\/p>/ig" content.htmlEdit: only "/" was escaped in the replacement but I forgot about "\" and "&" Should be better now.
Thanks for the input, all very helpful. sed is quite a mind trip.
I was using google for smtp but they finally ditched the setting for trusted apps. Found a new smtp server and all is well now. I use it to text myself or email info from an old "pupmemo" code I modified
I was playing around with curl commands. Works but I never got it where I needed for my setup.
Thanks!!