Batch replace serialized CDN subdomains

Question

We are moving an existing website to use CDN distribution of img resources. The project manager has mandated that on each page we have to distribute img src's to one of six different CDN subdomains -- the idea being that since download requests to different domains execute simultaneously instead of serially, we will get speed benefits.

In other words, I need to turn this...

<p>Argle bargle.</p>
<img src="http://old-domain.com/_img/image-a.jpg">
<img src="http://old-domain.com/_img/image-b.jpg">
<img src="http://old-domain.com/_img/image-c.jpg">
<img src="http://old-domain.com/_img/image-d.jpg">
<img src="http://old-domain.com/_img/image-e.jpg">
<img src="http://old-domain.com/_img/image-f.jpg">
<img src="http://old-domain.com/_img/image-g.jpg">
<img src="http://old-domain.com/_img/image-h.jpg">
<p>Whiz bang.</p>

Into this...

<p>Argle bargle.</p>
<img src="http://cdn1.cdn.com/_img/image-a.jpg">
<img src="http://cdn2.cdn.com/_img/image-b.jpg">
<img src="http://cdn3.cdn.com/_img/image-c.jpg">
<img src="http://cdn4.cdn.com/_img/image-d.jpg">
<img src="http://cdn5.cdn.com/_img/image-e.jpg">
<img src="http://cdn6.cdn.com/_img/image-f.jpg">
<img src="http://cdn1.cdn.com/_img/image-g.jpg">
<img src="http://cdn2.cdn.com/_img/image-h.jpg">
<p>Whiz bang.</p>

I have hundreds of files to update and they are all much more complex than the above sample. If the CDN was just one domain, I would batch replace all the files in an instant using TextWrangler. But I need to somehow serialize (or even randomize?) the replacement strings.

I use FileMaker Pro as a production front-end (to automate navigation construction, meta-tagging, etc.), and so I tried to craft a Calculation on my HTML Output field that would serialize every src, but I think it needs a for-each loop, which you can't do in a Calculation field (I have FM Pro, not FM Pro Advanced, so I can't use a Custom Function).

Anybody ever do anything similar in AppleScript? Or maybe leverage a Terminal text-processor? Any recommendations would be appreciated.

Answer 1

Are you actually writing HTML from FileMaker? If that's the case, and you want to manipulate the HTML using Perform AppleScript, then you can do something like the following.

Best Practice: Always use "Native AppleScript" instead of "Calculated AppleScript". To get data in and out of AppleScript, create a pair of global fields for reading and writing between FileMaker and AppleScript. This helps you isolate AppleScript errors when talking to other applications so that they don't bleed into later script steps. (I prefer to create a Global table and create a table occurrence called @Global. In that Global table I create two fields: AppleScriptInput and AppleScriptOutput. Note that you must create a record in this table or you will get mysterious "object not found" errors in AppleScript.)

Set Field [@Global::AppleScriptInput; $myHtml]
Perform AppleScript [“
    -- Read HTML from a global field in FileMaker
    set html to get data (table "@Global")'s (field "AppleScriptInput")

    -- Split html on old domain fragment
    set my text item delimiters to "<img src=\"http://old-domain.com/_img/"
    set html_fragments to text items of html

    -- Prefix resource with CDN domain fragment
    repeat with i from 2 to length of html_fragments
      set cdn_number to (i - 2) mod 6 + 1
      set cdn_fragment to "<img src=\"http://cdn" & cdn_number & ".cdn.com/_img/"
      set item i of html_fragments to cdn_fragment & item i of html_fragments
    end repeat

    -- Re-join html fragments
    set my text item delimiters to ""
    set html to html_fragments as text

    -- Write html back to a global field in FileMaker
    set data (table "@Global")'s (cell "AppleScriptInput") to html
”]
Set Variable [$revisedHtml; @Global::AppleScriptOutput]

The above script steps will read the input you provided from the FileMaker variable $html and save the output you provided to $revisedHtml.

You can do the same thing in pure FileMaker script of course:

Set Variable [$html; "<p>Argle bargle.</p>
    <img src=\"http://old-domain.com/_img/image-a.jpg\">
    <img src=\"http://old-domain.com/_img/image-b.jpg\">
    <img src=\"http://old-domain.com/_img/image-c.jpg\">
    <img src=\"http://old-domain.com/_img/image-d.jpg\">
    <img src=\"http://old-domain.com/_img/image-e.jpg\">
    <img src=\"http://old-domain.com/_img/image-f.jpg\">
    <img src=\"http://old-domain.com/_img/image-g.jpg\">
    <img src=\"http://old-domain.com/_img/image-h.jpg\">
    <p>Whiz bang.</p>"
    ]
Set Variable [$oldDomain; "http://old-domain.com/_img/"]
Set Variable [$itemCount; PatternCount ( $html ; $oldDomain )]
Loop
    Exit Loop If [Let ( $i = $i + 1 ; If ( $i > $itemCount ; Let ( $i = "" ; True ) ) )]
    Set Variable [$html; 
        Let ( [
          domainPosition = Position ( $html ; "http://old-domain.com/_img/" ; 1 ; 1 );
          domainLength = Length ( "http://old-domain.com/_img/" );
          cdnNumber = Mod ( $i - 1 ; 6 ) + 1
        ] ;
          Replace ( $html ; domainPosition ; domainLength ; "http://cdn" & cdnNumber & ".cdn.com/_img/" )
        )
    ]
End Loop
Show Custom Dialog ["Result"; $html]

Answer 2

I suggest you to use a simple bash script like that:

#!/bin/bash

OLD="old-domain.com\/_img"
NEW="cdn1.cdn.com\/_img"
DPATH="/home/of/your/files/*.html"

for f in $DPATH
do
   sed "s/$OLD/$NEW/g" "$f" > temp; mv temp "$f";
done

It replaces old-domain.com/_img with cdn1.cdn.com/_img on all your .html files on /home/of/your/files/.