FolderCrawler

[late flume]

https://github.com/HannulaTero/FolderCrawler
https://terohannula.itch.io/foldercrawler

This asset can be used to crawl out larger folder/file-structures without making the game to freeze up - mainly if structures gotta visit lot of folders inside folders. This returns structure made out of Folder and File -constructs, Folder containing arrays for Folder and File contained within.

Technically, if folder just has lot of items on same level, it could still freeze up, as I decided to not split up file_find_* calls as they ue global state; which could have ill effects if game logic elsewhere happens to use it too. Though most likely this is not the case, I may make a revision to split it up too, or add that as option. So in current state, this is good when looking folder structures, which branches a lot, not when they are wide.

Had this idea yesterday, when I was pondering getting all items within GM project, and then also wanted to split it up to several frames.

[latent valve]

So if I needed to find all files of a certain file type withing a folder or sub folder this could do that? I've been looking for a solution to do that.

[late flume]

This tries to find out all folders, subfolders and files within the given path, and returns structure representing these. Haven't put filter options there, I guess that could be a parameter.

But for now, after finding everything, you could iterate though the resulting structure, and find all those related to your criteria.

[latent valve]

I'll play around with it

[late flume]

If you wish, you could add a new array to the crawler, and in __Crawl function you can put a bit of code to push files to this array whenever they meet criteria.

Though today I most likely do some updates, such as adding "unsafe" crawl (split file_find_* loop over several frames).
I'll check whether I just make structure construction into function, so it can be changed more easily

[late flume]

Okay, made new version, which allows to do this.
In this example it will push all files which fit the criteria into the same array with custom action.
Though I am not happy how the function call looks right now.

// Stores all found images.
self.images = [ ];


// Dispatch the crawl.
folder_crawl(_path, function(_status, _result, _crawler)
{
  // This does crawl out full structure anyways.
  // -> Here it is stored.
  self.structure = _result;
}, {
  // Using optional parameters.
  // This allows further splitting to avoid stutters. 
  // -> But have to be more careful with file_find_* elsewhere.
  unsafe : true,

  // Custom action done with each item. 
  action : method(self, function(_item)
  {
    // Whether folder or file.
    if (_item.type != "file")
    {
      return;
    }
    
    // Check for the extension.
    var _ext = filename_ext(_item.name);
    if (_ext != ".png")
    {
      return;
    }
    
    // Push into array of images.
    array_push(self.images, _item);
  })
});

[latent valve]

I could hug you. This is awesome

[late flume]

Making v1.1.0 update, which would allow something like this instead:

handle = folder_crawl(path , {
  unsafe : true,
  context : self.items,
  file : function(_file, _context)
  {
    var _ext = filename_ext(_file.name);
    if (_ext == ".png")
    {
      array_push(_context, _file);
    }
  },
  callback : function(_crawler, _context)
  {
    self.timeTaken  = (get_timer() - self.timeBegin);
    self.foundCount = _crawler.DebugCount();
    self.status     = _crawler.GetStatus();
    self.structure  = _crawler.GetRoot();
    self.json       = json_stringify(self.structure, true);
  }
});

Also directory_exists seems pretty slow, and making two file_find_*seems to be faster.

edit. I have made the update.