I have a .srt file with text like this:

19
00:01:05,100 --> 00:01:08,820
countries such as Spain. Another factor to

20
00:01:08,820 --> 00:01:11,850
consider is the southern tip of Spain's coast

21
00:01:11,850 --> 00:01:15,060
being so close to northern Africa could have

I've found this code which works pretty well at cleaning the information but this code leaves in the initial numbers (these can be from one digit to four digits)

the result:

19countries such as Spain. Another factor to 20consider is the southern tip of Spain's coast 21being so close to northern Africa could have

Any idea how to remove the digits?

This is my code:

 <script>
            document.querySelector('#files').addEventListener('change', (e) => {
                
                let files = e.target.files,
                    i = 0,
                    reader = new FileReader;
            
                
                reader.onload = (e) => {
                    //console.log(files[i].name, e.target.result);
                    var fileName = files[i].name;
                    var text = e.target.result;

                    text = text.replace(/WEBVTT[\r\n]/,"");
                    text = text.replace(/NOTE duration:.*[\r\n]/,"");
                    text = text.replace(/NOTE language:.*[\r\n]/,"");
                    text = text.replace(/NOTE Confidence:. \d/g,"");
                    text = text.replace(/NOTE recognizability. \d/g,"");
                    text = text.replace(/[\r\n]. -. -. -. -. /g,"");
                    text = text.replace(/[\r\n].  --> . [\r\n]/g,"");
                    text = text.replace(/.[\r\n]. --> . [\r\n]/g,"");
                    text = text.replace(/[\n](.)/g," $1");
                    text = text.replace(/[\r\n] /g,"");
                    text = text.replace(/^ /,"");
                
                    var heading = document.createElement('h3');
                    document.body.appendChild(heading);
                    heading.innerHTML = "Transcript for '"   files[i].name   "'";
                
                    var copyButton = document.createElement('button');
                    document.body.appendChild(copyButton);
                    copyButton.onclick = function() {copyToClip(text,fileName); };
                    copyButton.innerHTML = "Copy transcript";
                    copyButton.className = "copyButton";
                
                    var div = document.createElement('div');
                    document.body.appendChild(div);
                    div.className = "cleanVTTText";
                    div.innerHTML = text;
            
                    //console.log(files[i].name, text);
                    console.log(files[i].name);
                    
                    
                    if (i   < files.length - 1) {
                        reader.readAsText(files[i]);
                    } else {
                        console.log('done');
                        
                    }
                };
                
                reader.readAsText(files[i]);
            
            }, false);
            
            function copyToClip(str,fileName) {
                function listener(e) {
                e.clipboardData.setData("text/html", str);
                e.clipboardData.setData("text/plain", str);
                e.preventDefault();
                }
                document.addEventListener("copy", listener);
                document.execCommand("copy");
                document.removeEventListener("copy", listener);
                alert("Copied transcript to clipboard:\n'" fileName "'");
            };     
            </script>

CodePudding user response：

Instead of using replace I propose a different solution, using split to split the string based on the newlines and then you get an array where you can construct the string however you want.

let text = `19
00:01:05,100 --> 00:01:08,820
countries such as Spain. Another factor to

20
00:01:08,820 --> 00:01:11,850
consider is the southern tip of Spain's coast

21
00:01:11,850 --> 00:01:15,060
being so close to northern Africa could have`
let newtext = text.split('\n').filter(el => el !== '' && !el.includes('-->') && el.match(/[^A-Za-z0-9\-_]/) );
console.log(newtext);
console.log('Just for example:');
console.log(`${newtext[0]} ${newtext[1]} ${newtext[2]}`);

Reference:

• String.prototype.replace()
• Array.prototype.filter()

CodePudding user response：

You could delete all text that starts an optional empty line, then a line with only a single integer, and then finally a single line with time stamps separated by "-->" (including any terminating newline characters).

The optional WEBVTT line is not part of the original crt standard, but can be removed in the same pass:

function cleanCrt(text) {
    return text.replace(/^(WEBVTT\b.*)?[\r\n]*\d (?:\r\n?|\n)[\d:,.]* --> [\d:,.]*[\r\n]*/gm, "");
}

const text = `WEBVTT
1
00:00:00,940 --> 00:00:04,630
Donkeys were first domesticated around 6000 years

2
00:00:04,630 --> 00:00:08,620
ago in northern Africa and Egypt, primarily for

3
00:00:08,620 --> 00:00:12,820
their milk and their meat. And around 2000 years

4
00:00:12,820 --> 00:00:15,970
ago, donkeys were used as draft animals, carrying 

5
00:00:15,970 --> 00:00:19,720
silk from the Pacific Ocean to the Mediterranean

6
00:00:19,720 --> 00:00:23,350
along the silk route. This was in return for trade [...]

19
00:01:05,100 --> 00:01:08,820
countries such as Spain. Another factor to

20
00:01:08,820 --> 00:01:11,850
consider is the southern tip of Spain's coast

21
00:01:11,850 --> 00:01:15,060
being so close to northern Africa could have`;

console.log(cleanCrt(text));