How can I remove certain lines in a multiline string? - php

Question

My code is receiving a string which I have no control of, which I'll call $my_string. The string is the contents of a transcript. If I echo the string, like so:
echo $my_string;
I can see the contents, which look something like this.
1
00:00:00.000 --> 00:00:04.980
[MUSIC]
2
00:00:04.980 --> 00:00:08.120
Hi, my name is holl and I am here
to write some PHP.
3
00:00:08.120 --> 00:00:10.277
You can see my screen, here.
What I'd like to do is run this through a function so it's just the actual words spoken, removing all the lines that signify time, or the order.
[MUSIC]
Hi, my name is holl and I am here
to write some php.
You can see my screen, here.
My idea is to explode the whole string by the break, and try to detect which lines which are either empty or start with a number, like so...
$lines = explode("\n", $my_string);
foreach ($lines as $line) {
if (is_numeric(line[0]) || empty(line[0]) ) {
continue;
}
$exclude[] = $line;
}
$transcript = implode("\n", $exclude);
But the result of this action is exactly the same- the output has numbers and blank lines. I clearly misunderstand something- but what is it? And is there a better way of trying to achieve my goal?
Thanks!
EDIT: Removed an echo where I wasn't actually using one in my code.

Answer 1

The problem is that you use indexing on $line:
$lines = explode("\n", $my_string);
foreach ($lines as $line) {
if (is_numeric(line[0]) || empty(line[0]) ) {//index usage?
continue;
}
$exclude[] = $line;
}
$transcript = echo implode("\n", $exclude); //remove echo
replace by:
$lines = explode("\n", $my_string);
foreach ($lines as $line) {
if (is_numeric($line) || empty($line) ) {//here
continue;
}
$exclude[] = $line;
}
$transcript = implode("\n", $exclude);
You also need regex matching to remove the 00:00:00.000 --> 00:00:04.980 fragments.
You can combine them by:
if(preg_match('/^(|\d+|\d+:\d+:\d+\.\d+\s+-->\s+\d+:\d+:\d+\.\d+)$/',$line)) { //regex
takes all possibilities into account:
$lines = explode("\n", $my_string);
foreach ($lines as $line) {
if(preg_match('/^(|\d+|\d+:\d+:\d+\.\d+\s+-->\s+\d+:\d+:\d+\.\d+)$/',$line)) {
continue;
}
$exclude[] = $line;
}
$transcript = implode("\n", $exclude);
echo $transcript;
Example (with php -a):
$ php -a
php > $my_string='1
php ' 00:00:00.000 --> 00:00:04.980
php ' [MUSIC]
php '
php ' 2
php ' 00:00:04.980 --> 00:00:08.120
php ' Hi, my name is holl and I am here
php ' to write some PHP.
php '
php ' 3
php ' 00:00:08.120 --> 00:00:10.277
php ' You can see my screen, here.';
php > $lines = explode("\n", $my_string);
php > foreach ($lines as $line) {
php { if(preg_match('/^(|\d+|\d+:\d+:\d+\.\d+\s+-->\s+\d+:\d+:\d+\.\d+)$/',$line)) {
php { continue;
php { }
php { $exclude[] = $line;
php { }
php > $transcript = implode("\n", $exclude);
php > echo $transcript;
[MUSIC]
Hi, my name is holl and I am here
to write some PHP.
You can see my screen, here.

Answer 2

Your code works almost. Just forgot $ in line[0] and " " is not empty().
$my_string = <<< EOF
1
00:00:00.000 --> 00:00:04.980
[MUSIC]
2
00:00:04.980 --> 00:00:08.120
Hi, my name is holl and I am here
to write some PHP.
3
00:00:08.120 --> 00:00:10.277
You can see my screen, here.
EOF;
$lines = explode("\n", $my_string);
foreach ($lines as $line) {
$temp = trim($line[0]);
if (is_numeric($temp) || empty($temp) ) {
continue;
}
$exclude[] = $line;
}
$transcript = implode("\n", $exclude);
echo $transcript;
Result:
[MUSIC]
Hi, my name is holl and I am here
to write some PHP.
You can see my screen, here.

Answer 3

It looks like it's a pattern. That is every first and second line contain meta data, the third is text, and the fourth is empty. If that is indeed the case, it should be trivial. You don't have to check the content at all and just grab the third line of every quartet:
$lines = explode("\n", $my_string);
$texts = array();
for ($i = 0; $i < count($lines); $i++) {
if ($i % 4 == 2) { // Index of third line is 2, of course.
$texts[] = $lines[i];
}
}
$transcript = implode($texts, "\n");
With alternative logic, because as you rightfully mentioned there can be more than one line of text, you could say that blocks/entries whatever you call them, are separated by an empty line. Each block starts with two lines of meta data, followed by one (or maybe zero) or more lines of text. With that logic you could parse it like this:
$lines = explode("\n", $my_string);
$texts = array();
$linenr = 0;
foreach ($lines as $line) {
// Keep track of the how manieth non-empty line it is.
if ($line === '')
$linenr = 0;
else
$linenr++;
// Skip the first two lines of a block.
if ($linenr > 2)
$texts[] = $line;
}
$transcript = implode($texts, "\n");
I don't know this particular format, but if I wanted to do this, I would be eager to find a pattern like this rather than parse the lines themselves. It looks like a script or subtitle file, and if you want to turn it into a transcript, it would be a shame if somebody shouted '300' and it would not be transcripted.

Answer 4

to remove theses lines try to use : preg_replace + regex
php man [1]: http://php.net/manual/en/function.preg-replace.php