jsgui3-server 0.0.121 → 0.0.123

package/resources/old/website-image-resource.js

@@ -1,1186 +1,1186 @@
-/*
- if (typeof define !== 'function') {
- var define = require('amdefine')(module);
- }
-
- define(['module', 'path', 'fs', 'url', '../../web/jsgui-html', 'os', 'http', 'url', './resource',
- '../../web/jsgui-je-suis-xml', 'cookies', '../../fs/jsgui-node-fs2-core',
- //"jpeg-js",
- '../../image/node/jsgui-node-jpeg',
- '../../image/node/jsgui-node-png',
- 'webworker-threads'],
-
- function(module, path, fs, url, jsgui, os, http, libUrl,
- Resource, JeSuisXML, Cookies, fs2,
- jsgui_jpeg,
- jsgui_png,
- Worker) {
- */
-
-
-// 24/07/2019 - Looks like this needs more work.
-//  Want to provide simple and powerful API.
-
-// Specifically being able to deal with icons and a directory of icons.
-
-
-
-
-
-
-
-var path = require('path'),
-    fs = require('fs'),
-    url = require('url'),
-    jsgui = require('jsgui3-html'),
-    os = require('os'),
-    http = require('http'),
-    libUrl = require('url'),
-    Resource = jsgui.Resource;
-
-fs2 = require('../fs2');
-
-// Then could this connect to a cms db?
-//  Or a wider source of images?
-// Mapping the ws image requests to acquiring the image data.
-//  Won't necessarily be on disk. Could call to another resource, such as network / file / files on network.
-//  Could map over / use another resource.
-
-/*
-
-jsgui_jpeg = require('../../image/node/jsgui-node-jpeg'),
-jsgui_png = require('../../image/node/jsgui-node-png');
-
-
-*/
-//Worker = require('webworker-threads');
-
-var stringify = jsgui.stringify,
-    each = jsgui.eac,
-    arrayify = jsgui.arrayify,
-    tof = jsgui.tof;
-var call_multi = jsgui.call_multi;
-var filter_map_by_regex = jsgui.filter_map_by_regex;
-var Class = jsgui.Class,
-    Data_Object = jsgui.Data_Object,
-    Enhanced_Data_Object = jsgui.Enhanced_Data_Object;
-var fp = jsgui.fp,
-    is_defined = jsgui.is_defined;
-var get_item_sig = jsgui.get_item_sig;
-var Collection = jsgui.Collection;
-
-// This resource may have quite a lot of functionality put in to deal with
-//  seriving images in an optimized way.
-// Want this so it also serves the developer in that images can be put in place
-//  and modified easily in an unoptimized way. Then they get put in place in an
-//  optimized way on request or automatically. Want to incorporate sprite sheet generation
-//  here. This should handle requests for individual images, but it's also possible sprites
-//  could be requested. If that's the case, maybe it should pass it onto a sprite resource.
-// It would be really good to have practically instant sprite generation, using C / C++
-//  and maybe OpenCL.
-// It could be requests to a sprite resource that returns a predefined selection of
-//  assets.
-// The image resource may also be called upon to return rescaled versions.
-//  That's something that OpenCL would also be very fast at.
-// The image / spritesheet resources may be connected to other database resources that provide
-//  caching and indexing of images.
-//   These would most likely be abstractions over data resources.
-
-// Want to make an image interface to a flexible db interface.
-//  Serving and caching optimized images would be cool.
-
-// Rather than relying on storing image versions in a DB (which would be cool, esp for distributed)
-//  we can store image versions + spritesheets on the local disk.
-// Cache them in RAM as well.
-//  Storing them in GPU RAM would be very cool, especially if from there we can get versions with
-//   different operations done, or the output of operations.
-
-
-// May need to change around a fair few references to make it workable.
-// May need some more complicated logic to change it to the path for service.
-
-
-
-// This is going to be extended so that it works with the Web_DB_Resource system.
-
-// Will have further handling on images, hopefully just deal with the DB side of things as file storage.
-//  Will keep image processing algorithms (and probably calls) out of the Web_DB_Postgres module, but use that interface
-//  to interact with images.
-
-// This module should also be comfortable dealing with images from the file system.
-
-// Also want to be able to quickly get thumbnails of images.
-
-// Will refer to images by a 'key', and also specify a size to get / display the image at
-//  The point will be to make it easy to program the appearance of the image, not having to keep track of the URL, having the image
-//   within a convenient admin system.
-
-// Serving a thumbnail, based on the image key, makes the most sense.
-
-// Want to be able to get the URL for an image from the key.
-//  Will also be able to get thumbnail URLs.
-
-// I think this Resource will also have more specialised functionality, possibly handlers too, that will be used by other modules.
-
-// Add an image to the database, along with it's metadata, and thumbnail versions.
-//  Thumbnail versions under key _400x300
-//   I think x will be an OK separator here.
-//   That size will refer to their maximum dimensions.
-
-
-
-
-// need to see what type of image it is.
-var mime_types = {
-    'jpg': 'image/jpeg',
-    'jpeg': 'image/jpeg',
-    'gif': 'image/gif',
-    'png': 'image/png',
-    'svg': 'image/svg+xml',
-    'otf': 'application/font-sfnt',
-    'ttf': 'application/font-sfnt'
-}
-
-var serve_image_file_from_buffer = function (buffer, filename, response) {
-    console.log('filename', filename);
-    var extname = path.extname(filename);
-    console.log('extname ' + extname);
-
-    var extension = extname.substr(1);
-    console.log('extension ' + extension);
-
-    response.writeHead(200, {
-        'Content-Type': mime_types[extension]
-    });
-    response.end(buffer, 'binary');
-}
-
-
-var serve_image_file_from_disk = function (filePath, response) {
-    var extname = path.extname(filePath);
-    console.log('extname ' + extname);
-    var extension = extname.substr(1);
-    console.log('extension ' + extension);
-
-    // then return the right MIME type for that extension.
-
-    // No, don't load the image file as a string.
-
-    // fs loadfile
-    //  async, then serve it with the correct mime type, write to the response buffer.
-
-    // can this be streamed to the response buffer?
-
-    fs.readFile(filePath, function (err, data) {
-        if (err) {
-            throw err;
-        } else {
-            response.writeHead(200, {
-                'Content-Type': mime_types[extension]
-            });
-            response.end(data, 'binary');
-        }
-    });
-
-
-    //fs2.load_file_as_string(filePath, function (err, data) {
-    //	if (err) {
-    //		throw err;
-    //	} else {
-    //		//var servableJs = updateReferencesForServing(data);
-    //		response.writeHead(200, {'Content-Type': mime_types[extension]});
-    //		response.end(data);
-    //	}
-    //});
-}
-
-
-class Site_Images extends Resource {
-
-
-    constructor(spec) {
-        super(spec);
-
-        //this.meta.set('custom_paths', new Data_Object({}));
-
-        //this.custom_paths = new Data_Object({});
-        this.custom_paths = {};
-
-
-        // Those are custom file paths.
-
-        // could have a collection of directories, indexed by name, that get served.
-
-        // Index the collection by string value?
-        //this.meta.set('served_directories', new Collection({'index_by': 'name'}));
-        this.served_directories = new Collection({
-            'index_by': 'name'
-        });
-
-    }
-    'start'(callback) {
-        callback(null, true);
-    }
-    'get_image_url'(image_key) {
-
-    }
-    'get_icon_url'(icon_key) {
-        // May have different size icons
-        //  For the moment will keep this simple.
-
-        return '/img/icons/' + icon_key + '.png';
-
-
-    }
-
-
-    /*
-    'get_metadata': fp(function(a, sig) {
-        console.log('get_metadata sig', sig);
-
-        var value, type_name, callback;
-        if (sig == '[B,s,f]') {
-            value = a[0]; type_name = a[1]; callback = a[2];
-
-            if (type_name === 'jpeg') {
-
-                // do this in a web worker?
-                //console.log('pre run web worker')
-
-
-                 //var worker = new Worker.Worker(function() {
-                 //console.log('running web worker')
-                 ////postMessage("I'm working before postMessage('ali').");
-                 ////this.onmessage = function(event) {
-                 //// But does the worker have access to jpeg_js?
-                 //console.log('pre decode');
-                 //var decoded = jpeg_js.decode(value);
-                 //console.log('post decode');
-                 //var width = decoded.width;
-                 //var height = decoded.height;
-                 //var res = {
-                 //'width': width,
-                 //'height': height,
-                 //'type_name': type_name
-                 //}
-                 //callback(null, res);
-                 //postMessage(res);
-                 //worker.close();
-                 //});
-                 //worker.onmessage = function(event) {
-                 //console.log("Worker said : " + event.data);
-                 //};
-
-
-
-                // It would be nice to improve the speed of the JPEG_JS module.
-
-                var jpeg = new jsgui_jpeg.JPEG({});
-
-                // load_size_from_buffer
-
-
-
-
-                jpeg.load_size_from_buffer(value, function(err, size) {
-                    if (err) {
-                        // But if there was an error loading the JPEG, it needs to raise an error.
-                        //  I'm now going to use
-
-
-                        throw err;
-
-                    } else {
-
-
-                        //console.log('jpeg loaded');
-
-                        //console.log('size', size);
-
-                        //throw 'stop';
-
-                        callback(null, {
-                            'size': size
-                        });
-
-
-
-                    }
-                })
-
-
-
-
-
-
-
-            }
-        }
-    }),
-    */
-
-    /*
-
-    '__get_resized_versions': fp(function(a, sig) {
-
-        // No, this needs to get resized versions of itself!
-        //  Also, it may make sense to work on this in Pixel_Buffer for the moment.
-
-
-
-
-        var that = this, callback, arr_image_item, buffer_image, type_name;
-
-        console.log('get_resized_versions sig', sig);
-
-        // if it's an array, we see what type of array it is.
-        //  it could be the buffer, then the type name
-
-        // May be given an array of sizes to get.
-        //  Would do the processing on Pixel Buffers.
-
-        // Possibly asyncronous function to get this as a Pixel Buffer?
-        //  Could we make a Pixel Buffer class operate directly on this class's Buffer? Should we?
-
-        console.trace('');
-        throw 'stop';
-
-        if (sig == '[a,f]') {
-            var sig0 = get_item_sig(a[0], 1);
-
-            // Will be getting an array of sizes.
-
-            console.log('sig0', sig0);
-
-            if (sig0 == '[B,o]') {
-                arr_image_item = a[0];
-                buffer_image = arr_image_item[0];
-                metadata = arr_image_item[1];
-
-                type_name = metadata.type_name;
-
-
-
-                // Maybe it would be better to use a better JPEG loading library.
-                //  node-webkit could be overkill.
-
-                // Could even translate the existing JavaScript code to C++.
-                //  It may be more effective at loading JPEGs.
-
-                //var Worker = require('webworker-threads').Worker;
-                // var w = new Worker('worker.js'); // Standard API
-
-                // You may also pass in a function:
-
-                // Load the JPEG from a buffer, then generate resized versions from the RGBA pixel buffer.
-
-                var jpeg = new jsgui_jpeg.JPEG({});
-
-
-                jpeg.load_from_buffer(buffer_image, function(err, cb_jpeg_loaded) {
-                    if (err) {
-                        throw err;
-                    } else {
-                        console.log('jpeg loaded');
-
-                        // We then want to get an RGBA Pixel Buffer from it.
-
-
-
-                        throw 'stop';
-                        //postMessage('done');
-                        worker.close();
-                    }
-                })
-
-
-
-
-
-
-
-                //var pb = jsgui_jpeg.load_pixel_buffer_from_disk()
-
-
-                // need to do the actual resizing too.
-                //  seems best to call on code in other module.
-                //  This resource system is nice, bur we really want to be using it to expose other functionality.
-
-                // Could maybe use a jpeg library.
-
-
-
-
-
-
-
-
-
-            }
-
-            callback = a[1];
-
-        }
-
-
-        //throw 'stop';
-
-
-    }),
-
-*/
-
-    /*
-
-    'get_document': function(key, callback) {
-        // get the doc from the db.
-        var that = this;
-        var pool = that.meta.get('pool');
-        var web_db = pool.get_resource('Web DB');
-
-
-        // get document with type and metadata
-
-        // get_document_full
-        //  would return an object with a few properties, including 'document'.
-
-        // {document, type, metadata}
-        //  or value, type, metadata
-
-
-
-
-        web_db.get_document_full(key, function(err, res_web_db_get_document) {
-            if (err) {
-                throw err;
-            } else {
-
-                // The web db could return more data.
-                //  That data could include the document type, and the metadata
-
-
-                // get_document_with_metadata
-                //  would also include the mime type
-
-
-
-                //console.log('res_web_db_get_document', res_web_db_get_document);
-
-                // callback with the document value, no metadata.
-
-                callback(null, res_web_db_get_document);
-
-            }
-        });
-    },
-
-    'set_document': fp(function(a, sig) {
-        // This needs to get metadata about the image document it is setting.
-        //  Will also make the document in various thumbnail sizes.
-
-        console.log('Image Resource set_document sig', sig);
-
-        // Also want to make and save resized versions of it.
-        //  Want to have a system of setting up different versions and automatic filters.
-        //  When one image changes, it could / should then sequentially make all the various derived images.
-        //   Don't want to implement that yet.
-        //   Just want a system where it makes resized versions of the images that are set as documents.
-
-        // I think we do want some things in the database about the id of the original image
-        //  Possibly also transformation info?
-        //   Maybe that should be in the metadata system?
-        //   That would mean a less complicated database.
-        //    I think queries for metadata should probably be really fast anyway.
-        //    Get all images with an original image of ...
-        //    Want to easily and quickly be able to get the resized versions of images.
-
-
-
-
-        // Also needs to set the non-resized as well.
-
-        var key, value, type_name, metadata, callback, that = this;
-
-        var pool = that.meta.get('pool');
-        //console.log('pool', pool);
-
-        var web_db = pool.get_resource('Web DB');
-
-        //console.log('web_db', web_db);
-
-        //throw 'stop';
-
-
-        if (sig == '[s,B,s,f]') {
-            // we work out the metadata and then use the web db resource's set_document.
-
-            key = a[0]; value = a[1]; type_name = a[2]; callback = a[3];
-
-            //console.log('callback', callback);
-            //throw 'stop';
-
-            // Best to get the image metadata.
-            // get_metadata
-            // value, type_name
-
-            // and the format is in the metadata.
-            //  maybe have type_name as fairly simple format info.
-            //  .format could be more complicated.
-            //  type_name should be fine.
-
-            if (type_name === 'jpeg') {
-                var jpeg = new jsgui_jpeg.JPEG({});
-
-
-                // I think load 24 bpp buffer by default when loading JPEGs.
-
-                jpeg.load_from_buffer(value, function(err, cb_jpeg_loaded) {
-                    if (err) {
-                        throw err;
-                    } else {
-                        console.log('jpeg loaded');
-
-                        //pb = jpeg.get_rgb_enhanced_pixel_buffer();
-
-                        var pb = jpeg.get_rgba_enhanced_pixel_buffer();
-
-                        // then save a variety of smaller jpeg versions
-
-
-
-
-                        // Want to save a few resized / thumbnail versions.
-                        //  Want the various sizes to be configurable - part of the DB?
-
-
-
-
-                        // get an rgba one.
-
-
-                        // I think by default will keep the ARs.
-                        // Same bits per pixel as well?
-
-                        // would be good to save that enhanced pixel buffer as a JPEG now.
-
-                        //  It would work out from the file name that it's a JPEG.
-
-
-                        // pb.save('test.jpeg');
-
-
-                        // Can save the various resized versions to the database.
-
-                        // Could call a function to save the various resized objects.
-
-                        console.log('jpeg.size', jpeg.size);
-
-                        var md = {
-                            'width': jpeg.size[0],
-                            'height': jpeg.size[1]
-                        }
-
-
-
-                        // In some situations may want to load, encode before saving, maybe strip out dodgy metadata.
-                        web_db.set_document(key, value, type_name, md, function(err, uploaded_image_document_id) {
-                            if (err) {
-                                throw err;
-                            } else {
-                                console.log('cb save_square_sized_version set_document uploaded_image_document_id: ' + uploaded_image_document_id);
-
-                                // When describing transformed documents, we can reference them by key rather than integer id.
-                                //  The different versions will still have different names, most likely given suffixes.
-
-                                // Then transform and save various versions of it.
-
-
-                                //throw 'stop';
-                                //callback2(null, res_saved);
-
-
-                                // need to look at the id of the item that was saved.
-                                //  could possibly use its key to save the resized versions.
-                                //  The resized versions will differ by name as well, but the transformations will be saved within the
-                                //  database as well. That way it will be possible to logically view the original images, being served the
-                                //  appropriate one for the device and required image size.
-
-                                var save_square_sized_version = function(square_size, callback2) {
-                                    console.log('save_square_sized_version');
-                                    //throw 'stop';
-
-                                    var resized_128x128 = pb.get_resized([square_size, square_size]);
-                                    var jpeg_128x128 = new jsgui_jpeg.JPEG({});
-                                    jpeg_128x128.load_from_rgb_pixel_buffer(resized_128x128);
-                                    var buffer_jpeg_128x128 = jpeg_128x128.save_to_buffer();
-                                    var md = {
-                                        'width': resized_128x128.size[0],
-                                        'height': resized_128x128.size[1]
-                                    }
-                                    // Need to change the key, so insert the _128x128 before the extension.
-
-                                    //var s_key = key.split('.');
-                                    // need to replace the last full stop?
-                                    var i = key.lastIndexOf('.');
-                                    var key2 = key;
-                                    if (i > -1) {
-                                        key2 = key.substr(0, i) + '_' + square_size + 'x' + square_size + '.' + key.substr(i + 1);
-                                    }
-
-                                    console.log('key2', key2);
-
-                                    web_db.set_document(key2, buffer_jpeg_128x128, type_name, md, function(err, res_saved) {
-                                        if (err) {
-                                            throw err;
-                                        } else {
-                                            console.log('cb save_square_sized_version set_document size: ' + square_size);
-                                            //throw 'stop';
-                                            callback2(null, res_saved);
-
-
-
-
-
-
-
-                                        }
-                                    });
-
-                                    // Needs to set a document with a transformation reference.
-
-                                    // Needs original document id.
-
-                                    // set transformed document
-                                    //  orig id uploaded_image_document_id
-                                    //  transformation info
-                                    //   (include the transformation type)
-                                    //   possibly include the transformation parameters.
-                                    //   we know it's resize from the type, but the transformation params can include the size it
-                                    //    was transformed to. Could possibly use a normalized transformation parameters system.
-
-                                    // .set_transformed_document(key2, transformed_buffer, type_name, metadata, transformation_type_name, transformation_params,
-                                    //  Need to create the document, while also creating the transformation record.
-
-                                    // Or a document itself could have a transformation source?
-                                    //  Separate records would be more normalized.
-                                    //  I think just a bit more sensible.
-
-                                    // A bit like the metadata record system.
-                                    //  Each transformation can have parameters which are similar in implementation to the metadata system.
-
-
-                                    // Document transformations will be generally about reencoding a document.
-
-                                    // Document_Transformations
-                                    //  id
-                                    //  source_document_id
-                                    //  target_document_id
-                                    //  transformation_verb_id
-
-                                    // Document_Transformation_Parameters
-                                    //  could have different types, like integer parameters, string parameters
-                                    //   this system would help a lot with searching for documents of particular sizes.
-                                    //   Would be a bit time consuming to make, but would be useful for the end system where transformed
-                                    //   documents can better be searched for and retrieved very quickly.
-
-                                    // dtrans_params
-                                    //  id, transformation id, key
-
-                                    // Document Transformation Parameter (Record)s
-                                    //  id, transformation id, key
-                                    // Document Transformation Integer Parameter (Record)s
-                                    //  These will be some of the most useful searchable and sortable parameters
-
-                                    // Possible transformation: encode so that it's 50KB or less. Parameter could be an integer size.
-
-                                    // Document Transformation Integer Parameters
-                                    //  id, transformation_param_id, value
-
-                                    // Make the document transformation record + param records after the target document
-                                    //  has been put into the database.
-
-                                    // Then save the resized one, as before,
-
-                                    // Then save the transformation that connects the two saved images.
-
-
-
-
-
-                                };
-
-                                // Looks like it needs to save the original version and get the id first.
-
-
-
-                                save_square_sized_version(128, function(err, transformed_doc_id) {
-                                    console.log('saved 128');
-                                    //throw 'stop';
-
-                                    if (err) {
-                                        throw err;
-                                    } else {
-
-                                        console.log('transformed_doc_id', transformed_doc_id);
-
-                                        web_db.ensure_transformation('resize', {'size': 128}, uploaded_image_document_id, transformed_doc_id, function(err, res_ensure_transformation) {
-                                            if (err) {
-                                                throw err;
-                                            } else {
-                                                console.log('res_ensure_transformation', res_ensure_transformation);
-
-                                                callback2(null, res_saved);
-
-
-                                            }
-                                        })
-
-
-
-
-                                        // need to save the resize transformation that connects them.
-
-
-
-                                        //throw 'stop';
-
-
-
-
-                                    }
-
-                                });
-
-                            }
-                        });
-
-
-
-                        // Should save the original file first, and then save the transformations based on it.
-
-
-
-
-
-
-                        //  Could give an array of different sizes and filenames.
-
-                        // Would be good to get an array of the different sizes to save as.
-                        //  Could also be useful to have fields for a description of the size
-                        //   eg iPhone 5 portrait full screen
-                        //    also iPhone 5 portrait portrait safari iOS 6 full window
-                        //     could potentially be different in iOS 7.
-
-                        // Also different icon sizes.
-                        //  For the moment, would be good to store it in a variety of different sizes.
-                        //  A few thumbnail image sizes would be useful.
-
-                        // Don't have this configurable in the DB at the moment...
-                        //  However the DB needs to know which images are related.
-                        //   Could have image_versions, with size restrictions and filters applied.
-
-                        // Want to be able to handle icons in the DB.
-
-                        // Will be storing a binary document value with modified name and size metadata
-                        //  For thumbnail images, may want 256x256 max.
-                        //  Or 512x512, possibly then scaled down in the browser.
-
-                        // Then we may want 128x128 max or 64x64, scaled for viewing in a list or grid in the browser.
-
-                        // [64, 64], [128, 128], [256, 256], [512, 512], [800, 600], [1024, 768]
-                        //  [64, 64], [128, 128], [256, 256], [512, 512], [1024, 1024]
-                        //   second option seems better.
-
-
-
-
-
-
-                        // jpeg.save_resized()
-
-                        // Want to make and save a max 64x64 version.
-                        //  It needs to be saved in the database as a different size.
-
-                        // put this in a function, so we can set a bunch of them quickly
-
-                        // And the save square sized version will deal with the db document transformation system.
-                        //  Needs to mark the result images as being resized transformations based on an original.
-
-
-
-
-
-
-
-
-                        // Set a few resized versions....
-                        // 128x128
-                        // 258x256
-                        // 512x512
-
-
-
-
-
-
-
-                        // then save that JPEG to the database.
-
-                        // want to get the encoded JPEG buffer.
-
-
-
-                        // then create and save a JPEG with those max dimensions.
-
-
-
-                        // We then want to get an RGBA Pixel Buffer from it.
-
-
-
-                        //throw 'stop';
-                        //postMessage('done');
-                        //worker.close();
-                    }
-                })
-            }
-
-
-        }
-
-    }),
-
-    */
-
-    // can have both a directory path and a URL path.
-    //  don't want to necessarily serve them under /img/
-
-
-
-
-    // want to be able to serve a path better.
-    //  still has bug here.
-    'serve_directory'(path) {
-        // Serves that directory, as any files given in that directory can be served from /js
-
-        //var served_directories = this.meta.get('served_directories');
-        //console.log('served_directories ' + stringify(served_directories));
-        //served_directories.push(path);
-        this.served_directories.push({
-            'name': path
-        });
-        //console.log('served_directories ' + stringify(served_directories));
-        //console.log('path ' + path);
-
-        // We may be serving directories AS something.
-
-        //throw 'stop';
-
-    }
-
-    // basically get requests, but can handle more than just get.
-    'process'(req, res) {
-        //console.log('Site_Images processing');
-        var remoteAddress = req.connection.remoteAddress;
-
-        var custom_paths = this.custom_paths;
-
-        var rurl = req.url;
-        //var pool = this.pool;
-        // should have a bunch of resources from the pool.
-
-        //var pool_resources = pool.resources();
-        //console.log('pool_resources ' + stringify(pool_resources));
-
-        var url_parts = url.parse(req.url, true);
-        //console.log('url_parts ' + stringify(url_parts));
-        var splitPath = url_parts.path.substr(1).split('/');
-        //console.log('resource site css splitPath ' + stringify(splitPath));
-
-        if (rurl.substr(0, 1) == '/') rurl = rurl.substr(1);
-        rurl = rurl.replace(/\./g, '☺');
-        //console.log('rurl ' + rurl);
-
-        var custom_response_entry = custom_paths[rurl];
-
-
-        var that = this;
-        //console.log('custom_response_entry ' + stringify(custom_response_entry));
-
-        if (custom_response_entry) {
-            var tcr = tof(custom_response_entry);
-            //console.log('tcr ' + tcr);
-
-            if (tcr == 'data_value') {
-                val = custom_response_entry.value();
-                //console.log('val ' + val);
-
-                var tval = tof(val);
-
-                if (tval == 'string') {
-                    // then it should be a local file path, serve it.
-                    serve_image_file_from_disk(val, res);
-                }
-            }
-            //throw 'stop';
-        } else {
-            //console.log('splitPath', splitPath);
-            if (splitPath.length > 0) {
-
-                // Can check for /js folder.
-                //  There will be some fixed resources for the site.
-                //   They will be served by Resource objects.
-                //  There may be some overlap of resources, with there being some very fixed purpose
-                //   specific resources that could duplicate some features of the more general ones.
-                //   Eventually, some of the code from the more specific resources will be
-                //   replacable with code from the more general ones.
-
-                // Site_JavaScript resource
-                //  Will serve JavaScript files needed for the site.
-                //   Could become more advanced at some points, serving particular builds.
-
-
-                // img or images
-
-                if (splitPath[0] == 'img' || splitPath[0] == 'images') {
-                    // Possibly could be something different?
-                    //  May be best to look at the path within the image resource.
-                    //var sjs = pool.get_resource('Site JavaScript');
-                    //console.log('sjs ' + sjs);
-
-                    //throw 'stop';
-
-                    // determine the name of the file to serve, serve that file
-                    //  Could use some more general kind of file server.
-
-                    // We may want to serve from disk.
-                    //  I think this could work differently to an interface to the disk resource.
-                    //  The image resource, like some others, could directly access the disk.
-                    //   They may have their own bit of caching.
-
-
-                    //console.log('splitPath', splitPath);
-
-                    if (splitPath.length > 1) {
-                        // At this point, can look for the file on disk within the app directory.
-                        //console.log('rurl', rurl);
-                        //console.log('req.url', req.url);
-                        // replace /images/ with /img/
-                        var project_disk_path = req.url.replace('images/', 'img/');
-                        if (project_disk_path.substr(0, 1) == '/') project_disk_path = project_disk_path.substr(1);
-
-                        // try to load it from disk.
-                        //  try to load it as a buffer
-                        fs2.load_file_as_buffer(project_disk_path, function (err, buffer) {
-                            if (err) {
-                                console.log('error loading ' + project_disk_path + ': ' + err);
-
-
-                                // Then try serving the file using the other methods, but will need to take care not to expect some resources to be there,
-                                //  such as the web_data resource.
-                                // Want top have a very versitile web_data resource that uses a DB, but also to have the means to interact with the files on disk
-                                //  in the project directory.
-
-                                res.writeHead(404, {
-                                    "Content-Type": "text/plain"
-                                });
-                                res.write("404 Not Found\n");
-                                res.end();
-
-                                //throw 'stop';
-
-
-                                const old_fn = () => {
-                                    if (splitPath.length == 2) {
-
-
-                                        // Could check options to see if to look in the database, or the file system.
-                                        //  Perhaps could have a website document storage abstraction that will store in the file system or the
-                                        //   database.
-                                        var fileName = splitPath[1];
-
-                                        // See if we can get the document with the key of the filename of the image.
-                                        //  Perhaps it will be stored as a site image.
-
-                                        that.get_document(fileName, function (err, res_get_document) {
-                                            if (err) {
-                                                throw err;
-                                            } else {
-                                                //console.log('res_get_document', res_get_document);
-                                                console.log('cb get document');
-
-                                                // Now we have it,
-
-                                                var doc_keys = Object.keys(res_get_document);
-                                                console.log('doc_keys', doc_keys);
-
-                                                // Need to know more information, such as the type_name
-                                                //  that extra info should probably get returned from the database.
-                                            }
-                                        });
-
-                                        /*
-                                         //console.log('url_parts.path ' + url_parts.path);
-                                         var filePath = url_parts.path.substr(1);
-                                         //console.log('module.uri ' + module.uri);
-                                         var val2 =  path.dirname(module.uri);
-                                         console.log('val2 ' + val2);
-                                         //throw '9) stop';
-                                         //var diskPath = val2 + '/../../images/' + fileName;
-                                         var diskPath = '../../ws/img/' + fileName;
-                                         // Also making use of custom paths...
-                                         // First check if such an image is in a specifically served directory.
-                                         var served_directories = this.meta.get('served_directories');
-                                         //console.log('served_directories ' + stringify(served_directories));
-                                         // see if the file exists in any of the served directories
-                                         // search for the file in the served directories.
-                                         //  will be an asyncronous search, and use an array of function calls with call_multi.
-                                         var fns = [];
-                                         var found_path;
-                                         each(served_directories, function(served_directory) {
-                                         //console.log('served_directory', served_directory);
-                                         var dir_val = served_directory.value();
-                                         //console.log('dir_val', dir_val);
-                                         var dir_name = dir_val.name;
-                                         //console.log('dir_name', dir_name);
-                                         var search_path = dir_name + '/' + fileName;
-                                         //console.log('search_path', search_path);
-                                         fns.push(function(callback) {
-                                         // check that directory
-                                         fs.exists(search_path, function(exists) {
-                                         //console.log('exists', exists);
-                                         if (!found_path && exists) {
-                                         found_path = search_path;
-                                         }
-                                         callback(null, exists);
-                                         })
-                                         })
-                                         });
-                                         call_multi(fns, function(err, res2) {
-                                         if (err) {
-                                         throw err;
-                                         } else {
-                                         //console.log('found_path', found_path);
-                                         if (found_path) {
-                                         diskPath = found_path;
-                                         serve_image_file_from_disk(diskPath, res);
-                                         }
-                                         }
-                                         });
-                                         //throw 'stop';
-                                         */
-
-                                        /*
-                                         fs2.load_file_as_string(diskPath, function (err, data) {
-                                         if (err) {
-                                         throw err;
-                                         } else {
-                                         //var servableJs = updateReferencesForServing(data);
-                                         res.writeHead(200, {'Content-Type': 'text/css'});
-                                         res.end(data);
-                                         }
-                                         });
-                                         */
-                                    } else {
-                                        if (splitPath.length > 2) {
-                                            // need to put the rest of it together...
-                                            var fileName = splitPath.slice(1, splitPath.length).join('/');
-                                            console.log('fileName', fileName);
-                                            var filePath = url_parts.path.substr(1);
-                                            //console.log('module.uri ' + module.uri);
-                                            var val2 = path.dirname(module.uri);
-                                            console.log('val2 ' + val2);
-                                            //throw '9) stop';
-                                            //var diskPath = val2 + '/../../images/' + fileName;
-                                            var diskPath = '../../ws/img/' + fileName;
-                                            // Do the search in the various served directories for the file.
-
-                                            var served_directories = this.served_directories;
-                                            //console.log('served_directories ' + stringify(served_directories));
-                                            // see if the file exists in any of the served directories
-                                            // search for the file in the served directories.
-                                            //  will be an asyncronous search, and use an array of function calls with call_multi.
-
-                                            var fns = [];
-                                            var found_path;
-                                            each(served_directories, function (served_directory) {
-                                                //console.log('served_directory', served_directory);
-                                                var dir_val = served_directory.value();
-                                                //console.log('dir_val', dir_val);
-                                                var dir_name = dir_val.name;
-                                                //console.log('dir_name', dir_name);
-
-                                                var search_path = dir_name + '/' + fileName;
-                                                //console.log('search_path', search_path);
-                                                fns.push(function (callback) {
-                                                    // check that directory
-
-                                                    fs.exists(search_path, function (exists) {
-                                                        //console.log('exists', exists);
-
-                                                        if (!found_path && exists) {
-                                                            found_path = search_path;
-                                                        }
-                                                        callback(null, exists);
-                                                    })
-                                                })
-                                            });
-
-                                            call_multi(fns, function (err, res2) {
-                                                if (err) {
-                                                    throw err;
-                                                } else {
-                                                    console.log('found_path', found_path);
-                                                    if (found_path) {
-                                                        diskPath = found_path;
-                                                        //serve_image_file_from_disk(diskPath, res);
-                                                    }
-                                                    serve_image_file_from_disk(diskPath, res);
-                                                }
-                                            });
-                                            // /js/core/jsgui-lang-enh
-                                            //console.log('!*!*!*! url_parts.path ' + url_parts.path);
-                                            /*
-                                             if (splitPath[1] == 'core') {
-                                             var fileName = splitPath[2];
-                                             var val2 =  path.dirname(module.uri);
-                                             //console.log('val2 ' + val2);
-                                             var diskPath = val2 + '/../core/' + fileName;
-                                             fs2.load_file_as_string(diskPath, function (err, data) {
-                                             if (err) {
-                                             throw err;
-                                             } else {
-                                             //var servableJs = updateReferencesForServing(data);
-                                             res.writeHead(200, {'Content-Type': 'text/css'});
-                                             res.end(data);
-                                             }
-                                             });
-                                             }
-                                             */
-                                        }
-                                    }
-                                }
-                            } else {
-                                console.log('loaded file, need to serve it');
-                                // need to set the mime type correctly.
-                                //  We can get this from the file name for the moment.
-                                // serve_image_file_from_buffer
-                                //serve_image_file_from_disk(diskPath, res);
-                                console.log('buffer', buffer);
-                                console.log('project_disk_path', project_disk_path);
-                                serve_image_file_from_buffer(buffer, project_disk_path, res);
-                            }
-                        })
-                    }
-                }
-            }
-        }
-    }
-}
-
-
-//return Site_Images;
-
-
-//});
+/*
+ if (typeof define !== 'function') {
+ var define = require('amdefine')(module);
+ }
+
+ define(['module', 'path', 'fs', 'url', '../../web/jsgui-html', 'os', 'http', 'url', './resource',
+ '../../web/jsgui-je-suis-xml', 'cookies', '../../fs/jsgui-node-fs2-core',
+ //"jpeg-js",
+ '../../image/node/jsgui-node-jpeg',
+ '../../image/node/jsgui-node-png',
+ 'webworker-threads'],
+
+ function(module, path, fs, url, jsgui, os, http, libUrl,
+ Resource, JeSuisXML, Cookies, fs2,
+ jsgui_jpeg,
+ jsgui_png,
+ Worker) {
+ */
+
+
+// 24/07/2019 - Looks like this needs more work.
+//  Want to provide simple and powerful API.
+
+// Specifically being able to deal with icons and a directory of icons.
+
+
+
+
+
+
+
+var path = require('path'),
+    fs = require('fs'),
+    url = require('url'),
+    jsgui = require('jsgui3-html'),
+    os = require('os'),
+    http = require('http'),
+    libUrl = require('url'),
+    Resource = jsgui.Resource;
+
+fs2 = require('../fs2');
+
+// Then could this connect to a cms db?
+//  Or a wider source of images?
+// Mapping the ws image requests to acquiring the image data.
+//  Won't necessarily be on disk. Could call to another resource, such as network / file / files on network.
+//  Could map over / use another resource.
+
+/*
+
+jsgui_jpeg = require('../../image/node/jsgui-node-jpeg'),
+jsgui_png = require('../../image/node/jsgui-node-png');
+
+
+*/
+//Worker = require('webworker-threads');
+
+var stringify = jsgui.stringify,
+    each = jsgui.eac,
+    arrayify = jsgui.arrayify,
+    tof = jsgui.tof;
+var call_multi = jsgui.call_multi;
+var filter_map_by_regex = jsgui.filter_map_by_regex;
+var Class = jsgui.Class,
+    Data_Object = jsgui.Data_Object,
+    Enhanced_Data_Object = jsgui.Enhanced_Data_Object;
+var fp = jsgui.fp,
+    is_defined = jsgui.is_defined;
+var get_item_sig = jsgui.get_item_sig;
+var Collection = jsgui.Collection;
+
+// This resource may have quite a lot of functionality put in to deal with
+//  seriving images in an optimized way.
+// Want this so it also serves the developer in that images can be put in place
+//  and modified easily in an unoptimized way. Then they get put in place in an
+//  optimized way on request or automatically. Want to incorporate sprite sheet generation
+//  here. This should handle requests for individual images, but it's also possible sprites
+//  could be requested. If that's the case, maybe it should pass it onto a sprite resource.
+// It would be really good to have practically instant sprite generation, using C / C++
+//  and maybe OpenCL.
+// It could be requests to a sprite resource that returns a predefined selection of
+//  assets.
+// The image resource may also be called upon to return rescaled versions.
+//  That's something that OpenCL would also be very fast at.
+// The image / spritesheet resources may be connected to other database resources that provide
+//  caching and indexing of images.
+//   These would most likely be abstractions over data resources.
+
+// Want to make an image interface to a flexible db interface.
+//  Serving and caching optimized images would be cool.
+
+// Rather than relying on storing image versions in a DB (which would be cool, esp for distributed)
+//  we can store image versions + spritesheets on the local disk.
+// Cache them in RAM as well.
+//  Storing them in GPU RAM would be very cool, especially if from there we can get versions with
+//   different operations done, or the output of operations.
+
+
+// May need to change around a fair few references to make it workable.
+// May need some more complicated logic to change it to the path for service.
+
+
+
+// This is going to be extended so that it works with the Web_DB_Resource system.
+
+// Will have further handling on images, hopefully just deal with the DB side of things as file storage.
+//  Will keep image processing algorithms (and probably calls) out of the Web_DB_Postgres module, but use that interface
+//  to interact with images.
+
+// This module should also be comfortable dealing with images from the file system.
+
+// Also want to be able to quickly get thumbnails of images.
+
+// Will refer to images by a 'key', and also specify a size to get / display the image at
+//  The point will be to make it easy to program the appearance of the image, not having to keep track of the URL, having the image
+//   within a convenient admin system.
+
+// Serving a thumbnail, based on the image key, makes the most sense.
+
+// Want to be able to get the URL for an image from the key.
+//  Will also be able to get thumbnail URLs.
+
+// I think this Resource will also have more specialised functionality, possibly handlers too, that will be used by other modules.
+
+// Add an image to the database, along with it's metadata, and thumbnail versions.
+//  Thumbnail versions under key _400x300
+//   I think x will be an OK separator here.
+//   That size will refer to their maximum dimensions.
+
+
+
+
+// need to see what type of image it is.
+var mime_types = {
+    'jpg': 'image/jpeg',
+    'jpeg': 'image/jpeg',
+    'gif': 'image/gif',
+    'png': 'image/png',
+    'svg': 'image/svg+xml',
+    'otf': 'application/font-sfnt',
+    'ttf': 'application/font-sfnt'
+}
+
+var serve_image_file_from_buffer = function (buffer, filename, response) {
+    console.log('filename', filename);
+    var extname = path.extname(filename);
+    console.log('extname ' + extname);
+
+    var extension = extname.substr(1);
+    console.log('extension ' + extension);
+
+    response.writeHead(200, {
+        'Content-Type': mime_types[extension]
+    });
+    response.end(buffer, 'binary');
+}
+
+
+var serve_image_file_from_disk = function (filePath, response) {
+    var extname = path.extname(filePath);
+    console.log('extname ' + extname);
+    var extension = extname.substr(1);
+    console.log('extension ' + extension);
+
+    // then return the right MIME type for that extension.
+
+    // No, don't load the image file as a string.
+
+    // fs loadfile
+    //  async, then serve it with the correct mime type, write to the response buffer.
+
+    // can this be streamed to the response buffer?
+
+    fs.readFile(filePath, function (err, data) {
+        if (err) {
+            throw err;
+        } else {
+            response.writeHead(200, {
+                'Content-Type': mime_types[extension]
+            });
+            response.end(data, 'binary');
+        }
+    });
+
+
+    //fs2.load_file_as_string(filePath, function (err, data) {
+    //	if (err) {
+    //		throw err;
+    //	} else {
+    //		//var servableJs = updateReferencesForServing(data);
+    //		response.writeHead(200, {'Content-Type': mime_types[extension]});
+    //		response.end(data);
+    //	}
+    //});
+}
+
+
+class Site_Images extends Resource {
+
+
+    constructor(spec) {
+        super(spec);
+
+        //this.meta.set('custom_paths', new Data_Object({}));
+
+        //this.custom_paths = new Data_Object({});
+        this.custom_paths = {};
+
+
+        // Those are custom file paths.
+
+        // could have a collection of directories, indexed by name, that get served.
+
+        // Index the collection by string value?
+        //this.meta.set('served_directories', new Collection({'index_by': 'name'}));
+        this.served_directories = new Collection({
+            'index_by': 'name'
+        });
+
+    }
+    'start'(callback) {
+        callback(null, true);
+    }
+    'get_image_url'(image_key) {
+
+    }
+    'get_icon_url'(icon_key) {
+        // May have different size icons
+        //  For the moment will keep this simple.
+
+        return '/img/icons/' + icon_key + '.png';
+
+
+    }
+
+
+    /*
+    'get_metadata': fp(function(a, sig) {
+        console.log('get_metadata sig', sig);
+
+        var value, type_name, callback;
+        if (sig == '[B,s,f]') {
+            value = a[0]; type_name = a[1]; callback = a[2];
+
+            if (type_name === 'jpeg') {
+
+                // do this in a web worker?
+                //console.log('pre run web worker')
+
+
+                 //var worker = new Worker.Worker(function() {
+                 //console.log('running web worker')
+                 ////postMessage("I'm working before postMessage('ali').");
+                 ////this.onmessage = function(event) {
+                 //// But does the worker have access to jpeg_js?
+                 //console.log('pre decode');
+                 //var decoded = jpeg_js.decode(value);
+                 //console.log('post decode');
+                 //var width = decoded.width;
+                 //var height = decoded.height;
+                 //var res = {
+                 //'width': width,
+                 //'height': height,
+                 //'type_name': type_name
+                 //}
+                 //callback(null, res);
+                 //postMessage(res);
+                 //worker.close();
+                 //});
+                 //worker.onmessage = function(event) {
+                 //console.log("Worker said : " + event.data);
+                 //};
+
+
+
+                // It would be nice to improve the speed of the JPEG_JS module.
+
+                var jpeg = new jsgui_jpeg.JPEG({});
+
+                // load_size_from_buffer
+
+
+
+
+                jpeg.load_size_from_buffer(value, function(err, size) {
+                    if (err) {
+                        // But if there was an error loading the JPEG, it needs to raise an error.
+                        //  I'm now going to use
+
+
+                        throw err;
+
+                    } else {
+
+
+                        //console.log('jpeg loaded');
+
+                        //console.log('size', size);
+
+                        //throw 'stop';
+
+                        callback(null, {
+                            'size': size
+                        });
+
+
+
+                    }
+                })
+
+
+
+
+
+
+
+            }
+        }
+    }),
+    */
+
+    /*
+
+    '__get_resized_versions': fp(function(a, sig) {
+
+        // No, this needs to get resized versions of itself!
+        //  Also, it may make sense to work on this in Pixel_Buffer for the moment.
+
+
+
+
+        var that = this, callback, arr_image_item, buffer_image, type_name;
+
+        console.log('get_resized_versions sig', sig);
+
+        // if it's an array, we see what type of array it is.
+        //  it could be the buffer, then the type name
+
+        // May be given an array of sizes to get.
+        //  Would do the processing on Pixel Buffers.
+
+        // Possibly asyncronous function to get this as a Pixel Buffer?
+        //  Could we make a Pixel Buffer class operate directly on this class's Buffer? Should we?
+
+        console.trace('');
+        throw 'stop';
+
+        if (sig == '[a,f]') {
+            var sig0 = get_item_sig(a[0], 1);
+
+            // Will be getting an array of sizes.
+
+            console.log('sig0', sig0);
+
+            if (sig0 == '[B,o]') {
+                arr_image_item = a[0];
+                buffer_image = arr_image_item[0];
+                metadata = arr_image_item[1];
+
+                type_name = metadata.type_name;
+
+
+
+                // Maybe it would be better to use a better JPEG loading library.
+                //  node-webkit could be overkill.
+
+                // Could even translate the existing JavaScript code to C++.
+                //  It may be more effective at loading JPEGs.
+
+                //var Worker = require('webworker-threads').Worker;
+                // var w = new Worker('worker.js'); // Standard API
+
+                // You may also pass in a function:
+
+                // Load the JPEG from a buffer, then generate resized versions from the RGBA pixel buffer.
+
+                var jpeg = new jsgui_jpeg.JPEG({});
+
+
+                jpeg.load_from_buffer(buffer_image, function(err, cb_jpeg_loaded) {
+                    if (err) {
+                        throw err;
+                    } else {
+                        console.log('jpeg loaded');
+
+                        // We then want to get an RGBA Pixel Buffer from it.
+
+
+
+                        throw 'stop';
+                        //postMessage('done');
+                        worker.close();
+                    }
+                })
+
+
+
+
+
+
+
+                //var pb = jsgui_jpeg.load_pixel_buffer_from_disk()
+
+
+                // need to do the actual resizing too.
+                //  seems best to call on code in other module.
+                //  This resource system is nice, bur we really want to be using it to expose other functionality.
+
+                // Could maybe use a jpeg library.
+
+
+
+
+
+
+
+
+
+            }
+
+            callback = a[1];
+
+        }
+
+
+        //throw 'stop';
+
+
+    }),
+
+*/
+
+    /*
+
+    'get_document': function(key, callback) {
+        // get the doc from the db.
+        var that = this;
+        var pool = that.meta.get('pool');
+        var web_db = pool.get_resource('Web DB');
+
+
+        // get document with type and metadata
+
+        // get_document_full
+        //  would return an object with a few properties, including 'document'.
+
+        // {document, type, metadata}
+        //  or value, type, metadata
+
+
+
+
+        web_db.get_document_full(key, function(err, res_web_db_get_document) {
+            if (err) {
+                throw err;
+            } else {
+
+                // The web db could return more data.
+                //  That data could include the document type, and the metadata
+
+
+                // get_document_with_metadata
+                //  would also include the mime type
+
+
+
+                //console.log('res_web_db_get_document', res_web_db_get_document);
+
+                // callback with the document value, no metadata.
+
+                callback(null, res_web_db_get_document);
+
+            }
+        });
+    },
+
+    'set_document': fp(function(a, sig) {
+        // This needs to get metadata about the image document it is setting.
+        //  Will also make the document in various thumbnail sizes.
+
+        console.log('Image Resource set_document sig', sig);
+
+        // Also want to make and save resized versions of it.
+        //  Want to have a system of setting up different versions and automatic filters.
+        //  When one image changes, it could / should then sequentially make all the various derived images.
+        //   Don't want to implement that yet.
+        //   Just want a system where it makes resized versions of the images that are set as documents.
+
+        // I think we do want some things in the database about the id of the original image
+        //  Possibly also transformation info?
+        //   Maybe that should be in the metadata system?
+        //   That would mean a less complicated database.
+        //    I think queries for metadata should probably be really fast anyway.
+        //    Get all images with an original image of ...
+        //    Want to easily and quickly be able to get the resized versions of images.
+
+
+
+
+        // Also needs to set the non-resized as well.
+
+        var key, value, type_name, metadata, callback, that = this;
+
+        var pool = that.meta.get('pool');
+        //console.log('pool', pool);
+
+        var web_db = pool.get_resource('Web DB');
+
+        //console.log('web_db', web_db);
+
+        //throw 'stop';
+
+
+        if (sig == '[s,B,s,f]') {
+            // we work out the metadata and then use the web db resource's set_document.
+
+            key = a[0]; value = a[1]; type_name = a[2]; callback = a[3];
+
+            //console.log('callback', callback);
+            //throw 'stop';
+
+            // Best to get the image metadata.
+            // get_metadata
+            // value, type_name
+
+            // and the format is in the metadata.
+            //  maybe have type_name as fairly simple format info.
+            //  .format could be more complicated.
+            //  type_name should be fine.
+
+            if (type_name === 'jpeg') {
+                var jpeg = new jsgui_jpeg.JPEG({});
+
+
+                // I think load 24 bpp buffer by default when loading JPEGs.
+
+                jpeg.load_from_buffer(value, function(err, cb_jpeg_loaded) {
+                    if (err) {
+                        throw err;
+                    } else {
+                        console.log('jpeg loaded');
+
+                        //pb = jpeg.get_rgb_enhanced_pixel_buffer();
+
+                        var pb = jpeg.get_rgba_enhanced_pixel_buffer();
+
+                        // then save a variety of smaller jpeg versions
+
+
+
+
+                        // Want to save a few resized / thumbnail versions.
+                        //  Want the various sizes to be configurable - part of the DB?
+
+
+
+
+                        // get an rgba one.
+
+
+                        // I think by default will keep the ARs.
+                        // Same bits per pixel as well?
+
+                        // would be good to save that enhanced pixel buffer as a JPEG now.
+
+                        //  It would work out from the file name that it's a JPEG.
+
+
+                        // pb.save('test.jpeg');
+
+
+                        // Can save the various resized versions to the database.
+
+                        // Could call a function to save the various resized objects.
+
+                        console.log('jpeg.size', jpeg.size);
+
+                        var md = {
+                            'width': jpeg.size[0],
+                            'height': jpeg.size[1]
+                        }
+
+
+
+                        // In some situations may want to load, encode before saving, maybe strip out dodgy metadata.
+                        web_db.set_document(key, value, type_name, md, function(err, uploaded_image_document_id) {
+                            if (err) {
+                                throw err;
+                            } else {
+                                console.log('cb save_square_sized_version set_document uploaded_image_document_id: ' + uploaded_image_document_id);
+
+                                // When describing transformed documents, we can reference them by key rather than integer id.
+                                //  The different versions will still have different names, most likely given suffixes.
+
+                                // Then transform and save various versions of it.
+
+
+                                //throw 'stop';
+                                //callback2(null, res_saved);
+
+
+                                // need to look at the id of the item that was saved.
+                                //  could possibly use its key to save the resized versions.
+                                //  The resized versions will differ by name as well, but the transformations will be saved within the
+                                //  database as well. That way it will be possible to logically view the original images, being served the
+                                //  appropriate one for the device and required image size.
+
+                                var save_square_sized_version = function(square_size, callback2) {
+                                    console.log('save_square_sized_version');
+                                    //throw 'stop';
+
+                                    var resized_128x128 = pb.get_resized([square_size, square_size]);
+                                    var jpeg_128x128 = new jsgui_jpeg.JPEG({});
+                                    jpeg_128x128.load_from_rgb_pixel_buffer(resized_128x128);
+                                    var buffer_jpeg_128x128 = jpeg_128x128.save_to_buffer();
+                                    var md = {
+                                        'width': resized_128x128.size[0],
+                                        'height': resized_128x128.size[1]
+                                    }
+                                    // Need to change the key, so insert the _128x128 before the extension.
+
+                                    //var s_key = key.split('.');
+                                    // need to replace the last full stop?
+                                    var i = key.lastIndexOf('.');
+                                    var key2 = key;
+                                    if (i > -1) {
+                                        key2 = key.substr(0, i) + '_' + square_size + 'x' + square_size + '.' + key.substr(i + 1);
+                                    }
+
+                                    console.log('key2', key2);
+
+                                    web_db.set_document(key2, buffer_jpeg_128x128, type_name, md, function(err, res_saved) {
+                                        if (err) {
+                                            throw err;
+                                        } else {
+                                            console.log('cb save_square_sized_version set_document size: ' + square_size);
+                                            //throw 'stop';
+                                            callback2(null, res_saved);
+
+
+
+
+
+
+
+                                        }
+                                    });
+
+                                    // Needs to set a document with a transformation reference.
+
+                                    // Needs original document id.
+
+                                    // set transformed document
+                                    //  orig id uploaded_image_document_id
+                                    //  transformation info
+                                    //   (include the transformation type)
+                                    //   possibly include the transformation parameters.
+                                    //   we know it's resize from the type, but the transformation params can include the size it
+                                    //    was transformed to. Could possibly use a normalized transformation parameters system.
+
+                                    // .set_transformed_document(key2, transformed_buffer, type_name, metadata, transformation_type_name, transformation_params,
+                                    //  Need to create the document, while also creating the transformation record.
+
+                                    // Or a document itself could have a transformation source?
+                                    //  Separate records would be more normalized.
+                                    //  I think just a bit more sensible.
+
+                                    // A bit like the metadata record system.
+                                    //  Each transformation can have parameters which are similar in implementation to the metadata system.
+
+
+                                    // Document transformations will be generally about reencoding a document.
+
+                                    // Document_Transformations
+                                    //  id
+                                    //  source_document_id
+                                    //  target_document_id
+                                    //  transformation_verb_id
+
+                                    // Document_Transformation_Parameters
+                                    //  could have different types, like integer parameters, string parameters
+                                    //   this system would help a lot with searching for documents of particular sizes.
+                                    //   Would be a bit time consuming to make, but would be useful for the end system where transformed
+                                    //   documents can better be searched for and retrieved very quickly.
+
+                                    // dtrans_params
+                                    //  id, transformation id, key
+
+                                    // Document Transformation Parameter (Record)s
+                                    //  id, transformation id, key
+                                    // Document Transformation Integer Parameter (Record)s
+                                    //  These will be some of the most useful searchable and sortable parameters
+
+                                    // Possible transformation: encode so that it's 50KB or less. Parameter could be an integer size.
+
+                                    // Document Transformation Integer Parameters
+                                    //  id, transformation_param_id, value
+
+                                    // Make the document transformation record + param records after the target document
+                                    //  has been put into the database.
+
+                                    // Then save the resized one, as before,
+
+                                    // Then save the transformation that connects the two saved images.
+
+
+
+
+
+                                };
+
+                                // Looks like it needs to save the original version and get the id first.
+
+
+
+                                save_square_sized_version(128, function(err, transformed_doc_id) {
+                                    console.log('saved 128');
+                                    //throw 'stop';
+
+                                    if (err) {
+                                        throw err;
+                                    } else {
+
+                                        console.log('transformed_doc_id', transformed_doc_id);
+
+                                        web_db.ensure_transformation('resize', {'size': 128}, uploaded_image_document_id, transformed_doc_id, function(err, res_ensure_transformation) {
+                                            if (err) {
+                                                throw err;
+                                            } else {
+                                                console.log('res_ensure_transformation', res_ensure_transformation);
+
+                                                callback2(null, res_saved);
+
+
+                                            }
+                                        })
+
+
+
+
+                                        // need to save the resize transformation that connects them.
+
+
+
+                                        //throw 'stop';
+
+
+
+
+                                    }
+
+                                });
+
+                            }
+                        });
+
+
+
+                        // Should save the original file first, and then save the transformations based on it.
+
+
+
+
+
+
+                        //  Could give an array of different sizes and filenames.
+
+                        // Would be good to get an array of the different sizes to save as.
+                        //  Could also be useful to have fields for a description of the size
+                        //   eg iPhone 5 portrait full screen
+                        //    also iPhone 5 portrait portrait safari iOS 6 full window
+                        //     could potentially be different in iOS 7.
+
+                        // Also different icon sizes.
+                        //  For the moment, would be good to store it in a variety of different sizes.
+                        //  A few thumbnail image sizes would be useful.
+
+                        // Don't have this configurable in the DB at the moment...
+                        //  However the DB needs to know which images are related.
+                        //   Could have image_versions, with size restrictions and filters applied.
+
+                        // Want to be able to handle icons in the DB.
+
+                        // Will be storing a binary document value with modified name and size metadata
+                        //  For thumbnail images, may want 256x256 max.
+                        //  Or 512x512, possibly then scaled down in the browser.
+
+                        // Then we may want 128x128 max or 64x64, scaled for viewing in a list or grid in the browser.
+
+                        // [64, 64], [128, 128], [256, 256], [512, 512], [800, 600], [1024, 768]
+                        //  [64, 64], [128, 128], [256, 256], [512, 512], [1024, 1024]
+                        //   second option seems better.
+
+
+
+
+
+
+                        // jpeg.save_resized()
+
+                        // Want to make and save a max 64x64 version.
+                        //  It needs to be saved in the database as a different size.
+
+                        // put this in a function, so we can set a bunch of them quickly
+
+                        // And the save square sized version will deal with the db document transformation system.
+                        //  Needs to mark the result images as being resized transformations based on an original.
+
+
+
+
+
+
+
+
+                        // Set a few resized versions....
+                        // 128x128
+                        // 258x256
+                        // 512x512
+
+
+
+
+
+
+
+                        // then save that JPEG to the database.
+
+                        // want to get the encoded JPEG buffer.
+
+
+
+                        // then create and save a JPEG with those max dimensions.
+
+
+
+                        // We then want to get an RGBA Pixel Buffer from it.
+
+
+
+                        //throw 'stop';
+                        //postMessage('done');
+                        //worker.close();
+                    }
+                })
+            }
+
+
+        }
+
+    }),
+
+    */
+
+    // can have both a directory path and a URL path.
+    //  don't want to necessarily serve them under /img/
+
+
+
+
+    // want to be able to serve a path better.
+    //  still has bug here.
+    'serve_directory'(path) {
+        // Serves that directory, as any files given in that directory can be served from /js
+
+        //var served_directories = this.meta.get('served_directories');
+        //console.log('served_directories ' + stringify(served_directories));
+        //served_directories.push(path);
+        this.served_directories.push({
+            'name': path
+        });
+        //console.log('served_directories ' + stringify(served_directories));
+        //console.log('path ' + path);
+
+        // We may be serving directories AS something.
+
+        //throw 'stop';
+
+    }
+
+    // basically get requests, but can handle more than just get.
+    'process'(req, res) {
+        //console.log('Site_Images processing');
+        var remoteAddress = req.connection.remoteAddress;
+
+        var custom_paths = this.custom_paths;
+
+        var rurl = req.url;
+        //var pool = this.pool;
+        // should have a bunch of resources from the pool.
+
+        //var pool_resources = pool.resources();
+        //console.log('pool_resources ' + stringify(pool_resources));
+
+        var url_parts = url.parse(req.url, true);
+        //console.log('url_parts ' + stringify(url_parts));
+        var splitPath = url_parts.path.substr(1).split('/');
+        //console.log('resource site css splitPath ' + stringify(splitPath));
+
+        if (rurl.substr(0, 1) == '/') rurl = rurl.substr(1);
+        rurl = rurl.replace(/\./g, '☺');
+        //console.log('rurl ' + rurl);
+
+        var custom_response_entry = custom_paths[rurl];
+
+
+        var that = this;
+        //console.log('custom_response_entry ' + stringify(custom_response_entry));
+
+        if (custom_response_entry) {
+            var tcr = tof(custom_response_entry);
+            //console.log('tcr ' + tcr);
+
+            if (tcr == 'data_value') {
+                val = custom_response_entry.value();
+                //console.log('val ' + val);
+
+                var tval = tof(val);
+
+                if (tval == 'string') {
+                    // then it should be a local file path, serve it.
+                    serve_image_file_from_disk(val, res);
+                }
+            }
+            //throw 'stop';
+        } else {
+            //console.log('splitPath', splitPath);
+            if (splitPath.length > 0) {
+
+                // Can check for /js folder.
+                //  There will be some fixed resources for the site.
+                //   They will be served by Resource objects.
+                //  There may be some overlap of resources, with there being some very fixed purpose
+                //   specific resources that could duplicate some features of the more general ones.
+                //   Eventually, some of the code from the more specific resources will be
+                //   replacable with code from the more general ones.
+
+                // Site_JavaScript resource
+                //  Will serve JavaScript files needed for the site.
+                //   Could become more advanced at some points, serving particular builds.
+
+
+                // img or images
+
+                if (splitPath[0] == 'img' || splitPath[0] == 'images') {
+                    // Possibly could be something different?
+                    //  May be best to look at the path within the image resource.
+                    //var sjs = pool.get_resource('Site JavaScript');
+                    //console.log('sjs ' + sjs);
+
+                    //throw 'stop';
+
+                    // determine the name of the file to serve, serve that file
+                    //  Could use some more general kind of file server.
+
+                    // We may want to serve from disk.
+                    //  I think this could work differently to an interface to the disk resource.
+                    //  The image resource, like some others, could directly access the disk.
+                    //   They may have their own bit of caching.
+
+
+                    //console.log('splitPath', splitPath);
+
+                    if (splitPath.length > 1) {
+                        // At this point, can look for the file on disk within the app directory.
+                        //console.log('rurl', rurl);
+                        //console.log('req.url', req.url);
+                        // replace /images/ with /img/
+                        var project_disk_path = req.url.replace('images/', 'img/');
+                        if (project_disk_path.substr(0, 1) == '/') project_disk_path = project_disk_path.substr(1);
+
+                        // try to load it from disk.
+                        //  try to load it as a buffer
+                        fs2.load_file_as_buffer(project_disk_path, function (err, buffer) {
+                            if (err) {
+                                console.log('error loading ' + project_disk_path + ': ' + err);
+
+
+                                // Then try serving the file using the other methods, but will need to take care not to expect some resources to be there,
+                                //  such as the web_data resource.
+                                // Want top have a very versitile web_data resource that uses a DB, but also to have the means to interact with the files on disk
+                                //  in the project directory.
+
+                                res.writeHead(404, {
+                                    "Content-Type": "text/plain"
+                                });
+                                res.write("404 Not Found\n");
+                                res.end();
+
+                                //throw 'stop';
+
+
+                                const old_fn = () => {
+                                    if (splitPath.length == 2) {
+
+
+                                        // Could check options to see if to look in the database, or the file system.
+                                        //  Perhaps could have a website document storage abstraction that will store in the file system or the
+                                        //   database.
+                                        var fileName = splitPath[1];
+
+                                        // See if we can get the document with the key of the filename of the image.
+                                        //  Perhaps it will be stored as a site image.
+
+                                        that.get_document(fileName, function (err, res_get_document) {
+                                            if (err) {
+                                                throw err;
+                                            } else {
+                                                //console.log('res_get_document', res_get_document);
+                                                console.log('cb get document');
+
+                                                // Now we have it,
+
+                                                var doc_keys = Object.keys(res_get_document);
+                                                console.log('doc_keys', doc_keys);
+
+                                                // Need to know more information, such as the type_name
+                                                //  that extra info should probably get returned from the database.
+                                            }
+                                        });
+
+                                        /*
+                                         //console.log('url_parts.path ' + url_parts.path);
+                                         var filePath = url_parts.path.substr(1);
+                                         //console.log('module.uri ' + module.uri);
+                                         var val2 =  path.dirname(module.uri);
+                                         console.log('val2 ' + val2);
+                                         //throw '9) stop';
+                                         //var diskPath = val2 + '/../../images/' + fileName;
+                                         var diskPath = '../../ws/img/' + fileName;
+                                         // Also making use of custom paths...
+                                         // First check if such an image is in a specifically served directory.
+                                         var served_directories = this.meta.get('served_directories');
+                                         //console.log('served_directories ' + stringify(served_directories));
+                                         // see if the file exists in any of the served directories
+                                         // search for the file in the served directories.
+                                         //  will be an asyncronous search, and use an array of function calls with call_multi.
+                                         var fns = [];
+                                         var found_path;
+                                         each(served_directories, function(served_directory) {
+                                         //console.log('served_directory', served_directory);
+                                         var dir_val = served_directory.value();
+                                         //console.log('dir_val', dir_val);
+                                         var dir_name = dir_val.name;
+                                         //console.log('dir_name', dir_name);
+                                         var search_path = dir_name + '/' + fileName;
+                                         //console.log('search_path', search_path);
+                                         fns.push(function(callback) {
+                                         // check that directory
+                                         fs.exists(search_path, function(exists) {
+                                         //console.log('exists', exists);
+                                         if (!found_path && exists) {
+                                         found_path = search_path;
+                                         }
+                                         callback(null, exists);
+                                         })
+                                         })
+                                         });
+                                         call_multi(fns, function(err, res2) {
+                                         if (err) {
+                                         throw err;
+                                         } else {
+                                         //console.log('found_path', found_path);
+                                         if (found_path) {
+                                         diskPath = found_path;
+                                         serve_image_file_from_disk(diskPath, res);
+                                         }
+                                         }
+                                         });
+                                         //throw 'stop';
+                                         */
+
+                                        /*
+                                         fs2.load_file_as_string(diskPath, function (err, data) {
+                                         if (err) {
+                                         throw err;
+                                         } else {
+                                         //var servableJs = updateReferencesForServing(data);
+                                         res.writeHead(200, {'Content-Type': 'text/css'});
+                                         res.end(data);
+                                         }
+                                         });
+                                         */
+                                    } else {
+                                        if (splitPath.length > 2) {
+                                            // need to put the rest of it together...
+                                            var fileName = splitPath.slice(1, splitPath.length).join('/');
+                                            console.log('fileName', fileName);
+                                            var filePath = url_parts.path.substr(1);
+                                            //console.log('module.uri ' + module.uri);
+                                            var val2 = path.dirname(module.uri);
+                                            console.log('val2 ' + val2);
+                                            //throw '9) stop';
+                                            //var diskPath = val2 + '/../../images/' + fileName;
+                                            var diskPath = '../../ws/img/' + fileName;
+                                            // Do the search in the various served directories for the file.
+
+                                            var served_directories = this.served_directories;
+                                            //console.log('served_directories ' + stringify(served_directories));
+                                            // see if the file exists in any of the served directories
+                                            // search for the file in the served directories.
+                                            //  will be an asyncronous search, and use an array of function calls with call_multi.
+
+                                            var fns = [];
+                                            var found_path;
+                                            each(served_directories, function (served_directory) {
+                                                //console.log('served_directory', served_directory);
+                                                var dir_val = served_directory.value();
+                                                //console.log('dir_val', dir_val);
+                                                var dir_name = dir_val.name;
+                                                //console.log('dir_name', dir_name);
+
+                                                var search_path = dir_name + '/' + fileName;
+                                                //console.log('search_path', search_path);
+                                                fns.push(function (callback) {
+                                                    // check that directory
+
+                                                    fs.exists(search_path, function (exists) {
+                                                        //console.log('exists', exists);
+
+                                                        if (!found_path && exists) {
+                                                            found_path = search_path;
+                                                        }
+                                                        callback(null, exists);
+                                                    })
+                                                })
+                                            });
+
+                                            call_multi(fns, function (err, res2) {
+                                                if (err) {
+                                                    throw err;
+                                                } else {
+                                                    console.log('found_path', found_path);
+                                                    if (found_path) {
+                                                        diskPath = found_path;
+                                                        //serve_image_file_from_disk(diskPath, res);
+                                                    }
+                                                    serve_image_file_from_disk(diskPath, res);
+                                                }
+                                            });
+                                            // /js/core/jsgui-lang-enh
+                                            //console.log('!*!*!*! url_parts.path ' + url_parts.path);
+                                            /*
+                                             if (splitPath[1] == 'core') {
+                                             var fileName = splitPath[2];
+                                             var val2 =  path.dirname(module.uri);
+                                             //console.log('val2 ' + val2);
+                                             var diskPath = val2 + '/../core/' + fileName;
+                                             fs2.load_file_as_string(diskPath, function (err, data) {
+                                             if (err) {
+                                             throw err;
+                                             } else {
+                                             //var servableJs = updateReferencesForServing(data);
+                                             res.writeHead(200, {'Content-Type': 'text/css'});
+                                             res.end(data);
+                                             }
+                                             });
+                                             }
+                                             */
+                                        }
+                                    }
+                                }
+                            } else {
+                                console.log('loaded file, need to serve it');
+                                // need to set the mime type correctly.
+                                //  We can get this from the file name for the moment.
+                                // serve_image_file_from_buffer
+                                //serve_image_file_from_disk(diskPath, res);
+                                console.log('buffer', buffer);
+                                console.log('project_disk_path', project_disk_path);
+                                serve_image_file_from_buffer(buffer, project_disk_path, res);
+                            }
+                        })
+                    }
+                }
+            }
+        }
+    }
+}
+
+
+//return Site_Images;
+
+
+//});
 module.exports = Site_Images;