ImageGrid/ui/ImageGrid.js

1010 lines
26 KiB
JavaScript
Raw Normal View History

/**********************************************************************
*
* Viewer Generation III
*
* Split the API into the following sections:
* - main control actions
* do main domain tasks like image and ribbon manipulation.
* - serialization and deserialization
* load and save data
* - UI
* basic align, animation and modes
*
*
* TODO group all actions into an object, referencing the viewer...
* ...this will make this reusable multiple times.
* TODO wrap the actions into an object and make all queries relative to
* a single root viewer...
* ...this will make the code reusable multiple times...
*
*
**********************************************************************/
// Data format...
var DATA = {
// the ribbon cache...
// in the simplest form this is a list of lists of GIDs
ribbons: [
$(new Array(100)).map(function(i){return i}).toArray()
],
// flat ordered list of images in current context...
// in the simplest form this is a list of GIDs.
order: $(new Array(100)).map(function(i){return i}).toArray(),
// the images object, this is indexed by image GID and contains all
// the needed data...
images: {
}
}
/**********************************************************************
* Helpers
*/
// XXX need ribbon end indicators...
// XXX might need shift left/right indicators (later)...
function flashIndicator(direction){
$({
prev: '.up-indicator',
next: '.down-indicator',
// XXX not implemented yet...
start: '.start-indicator',
end: '.end-indicator',
}[direction])
// NOTE: this needs to be visible in all cases and key press
// rhythms...
.show()
.delay(20)
.fadeOut(200)
}
function getRibbon(image){
image = image == null ? $('.current.image') : $(image)
return image.closest('.ribbon')
}
// NOTE: elem is optional and if given can be an image or a ribbon...
function getRibbonIndex(elem){
if(elem == null){
var ribbon = getRibbon()
} else {
elem = $(elem)
if(elem.hasClass('image')){
ribbon = getRibbon(elem)
} else {
ribbon = elem
}
}
return $('.ribbon').index(ribbon)
}
function getImageOrder(image){
image = image == null ? $('.current.image') : $(image)
if(image.length == 0){
return
}
return JSON.parse(image.attr('order'))
}
function getImageGID(image){
image = image == null ? $('.current.image') : $(image)
if(image.length == 0){
return
}
return JSON.parse(image.attr('gid'))
}
// ...tried to make this as brain-dead-stupidly-simple as possible...
function getRelativeVisualPosition(outer, inner){
outer = $(outer).offset()
inner = $(inner).offset()
return {
top: inner.top - outer.top,
left: inner.left - outer.left
}
}
// Returns the image size (width) as viewed on screen...
function getVisibleImageSize(){
return $('.image').outerWidth() * getElementScale($('.ribbon-set'))
}
// Return the number of images that can fit to viewer width...
function getScreenWidthInImages(){
return $('.viewer').innerWidth() / getVisibleImageSize()
}
// NOTE: this will return an empty jquery object if no image is before
// the target...
// NOTE: this might return an empty target if the ribbon is empty...
// XXX need tp make this loadable ribbon compatible -- the target may
// not be loaded...
function getImageBefore(image, ribbon, mode){
mode = mode == null ? NAV_DEFAULT : mode
image = image == null ? $('.current.image') : $(image)
if(ribbon == null){
ribbon = getRibbon(image)
}
var images = $(ribbon).find('.image').filter(mode)
// XXX need to process/format this correctly...
var order = JSON.parse(image.attr('order'))
var prev = []
images.each(function(){
if(order < JSON.parse($(this).attr('order'))){
return false
}
prev = this
})
return $(prev)
}
// same as getImageBefore, but uses gids and searches in DATA...
// XXX check for corner cases...
// XXX getGIDBefore(1, 1) does not work...
function getGIDBefore(gid, ribbon){
ribbon = DATA.ribbons[ribbon]
var order = DATA.order
var target = ribbon.indexOf(gid)
if(target >= 0){
return gid
}
target = order.indexOf(gid)
var i = ribbon.length
while(i > 0){
i = Math.floor(ribbon.length/2)
console.log('>>>', target, i, order.indexOf(ribbon[i]), order.indexOf(ribbon[i+1]))
if(target >= order.indexOf(ribbon[i]) && target < order.indexOf(ribbon[i+1])){
return ribbon[i]
// XXX I do not understand why this works correctly, think I need some sleep...
} else if(target < order.indexOf(ribbon[i])){
ribbon = ribbon.slice(0, i)
} else {
ribbon = ribbon.slice(i)
}
}
return null
}
function shiftTo(image, ribbon){
var target = getImageBefore(image, ribbon, NAV_ALL)
var cur_ribbon = getRibbon(image)
// insert before the first image if nothing is before the target...
if(target.length == 0){
image.prependTo($(ribbon))
} else {
image.insertAfter(target)
}
$('.viewer').trigger('shiftedImage', [image, cur_ribbon, ribbon])
// if removing last image out of a ribbon, remove the ribbon....
if(cur_ribbon.find('.image').length == 0){
// XXX check if the ribbon outside the loaded area is empty...
// ...do we need this check? it might be interresting to
// "collapse" disjoint, empty areas...
// ......if so, will also need to do this in DATA...
removeRibbon(cur_ribbon)
}
return image
}
function shiftImage(direction, image, force_create_ribbon){
if(image == null){
// XXX need to make this context specific...
image = $('.current.image')
} else {
image = $(image)
}
var old_ribbon = getRibbon(image)
var ribbon = old_ribbon[direction]('.ribbon')
// need to create a new ribbon...
if(ribbon.length == 0 || force_create_ribbon == true){
var index = getRibbonIndex(old_ribbon)
index = direction == 'after' ? index + 1 : index
ribbon = createRibbon(index)
shiftTo(image, ribbon)
} else {
shiftTo(image, ribbon)
}
return image
}
/**********************************************************************
* Constructors
*/
// NOTE: to avoid state sync problems this should clone an image if
// one is available...
function createImage(n){
if(n == null){
if(window._n == null){
window._n = 0
}
n = _n
_n += 1
}
var img = $('.image')
if(img.length > 0){
return img.first().clone()
.attr({
'order': JSON.stringify(n),
'gid': JSON.stringify(n),
// need to strip extra classes...
'class': 'image'
})
} else {
return $('<div order="'+n+'" gid="'+n+'" class="image"/>')
}
}
// This will create a set of new images, reusing a list of existing
// elements if given.
// XXX do we need this???
// XXX add position...
function createImages(need, have){
have = have == null ? [] : have
// we have enough elements in the cache...
if(have.length >= need){
return $(have.splice(0, need))
// need to create additional elements...
} else {
return $(have.toArray().concat(new Array(need - have.length)))
.map(function(i, elem){
if(elem != null){
return elem
}
return createImage()[0]
})
}
}
function createRibbon(index){
// make the ribbon...
var ribbon = $('<div class="ribbon"/>')
if(index == null){
return ribbon
}
var ribbons = $('.ribbon')
if(index >= ribbons.length){
ribbons.last().after(ribbon)
} else {
ribbons.eq(index).before(ribbon)
}
$('.viewer').trigger('createdRibbon', [ribbon])
return ribbon
}
// NOTE: this will pass the index where the ribbon was to the event,
// rather than an actual ribbon...
// XXX check if ribbon is empty...
function removeRibbon(ribbon){
// ribbon can be an index...
if(typeof(ribbon) == typeof(1)){
ribbon = $('.ribbon').eq(ribbon)
}
$('.viewer').trigger('removedRibbon', [getRibbonIndex(ribbon)])
return $(ribbon).remove()
}
/**********************************************************************
* Infinite ribbon machinery
*/
// NOTE: negative left or right will contract the ribbon...
// XXX check what goes on if left/right are far more than length...
function extendRibbon(left, right, ribbon){
ribbon = ribbon == null ?
getRibbon()
: $(ribbon)
left = left == null ? 0 : left
right = right == null ? 0 : right
var images = ribbon.children('.image')
var removed = []
var res = {
left: $([]),
right: $([])
}
// truncate...
// NOTE: we save the detached elements to reuse them on extending,
// if needed...
if(left < 0){
removed = $(images.splice(0, -left)).detach()
}
if(right < 0){
var l = images.length
removed = $(images.splice(l+right, l)).detach()
}
// extend...
if (left > 0){
res.left = createImages(left, removed).prependTo(ribbon)
}
if (right > 0){
res.right = createImages(right, removed).appendTo(ribbon)
}
// compensate for the truncation...
// XXX do we need to split this into a separate function?
// ...the rest of the function if pretty generic...
// XXX for some odd reason this behaves erratically when the page
// is zoomed...
if(left != 0){
var l = parseFloat(ribbon.css('left'))
l = isNaN(l) ? 0 : l
ribbon.css({
left: l + (-left * parseFloat(images.outerWidth()))
})
}
return res
}
// Roll the ribbon n positions to the left.
//
// NOTE: if n is negative the ribbon will be rolled right.
// NOTE: rollRibbon(N, R) is equivalent to extendRibbon(-N, N, R)
// NOTE: this will return a single list of relocated elements...
function rollRibbon(n, ribbon){
var res = extendRibbon(-n, n, ribbon)
return n > 0 ? res.right : res.left
}
/**********************************************************************
* Loaders
*/
// NOTE: count can be either hegative or positive, this will idicate
// load direction...
// NOTE: this will not include the 'from' GID in the resulting list...
// NOTE: this can calculate the ribbon number if an image can be only
// in one ribbon...
// NOTE: if an image can be in more than one ribbon, one MUST suply the
// correct ribbon number...
// XXX do we need more checking???
function getImageGIDs(from, count, ribbon){
if(count == 0){
return []
}
// ribbon default value...
if(ribbon == null){
$(DATA.ribbons).each(function(i, e){
if(e.indexOf(from) >= 0){
ribbon = i
return false
}
})
}
// XXX checkif this is empty...
ribbon = DATA.ribbons[ribbon]
if(count > 0){
var start = ribbon.indexOf(from) + 1
return ribbon.slice(start, start + count)
} else {
var end = ribbon.indexOf(from)
return ribbon.slice((Math.abs(count) >= end ? 0 : end + count), end)
}
}
function updateImage(image, gid, size){
image = $(image)
if(gid == null){
gid = JSON.parse(image.attr('gid'))
} else {
image.attr('gid', JSON.stringify(gid))
}
size = size == null ? getVisibleImageSize() : size
image.attr({
//order: JSON.stringify(DATA.order.indexOf(gid)),
order: JSON.stringify(gid)
// XXX update attrs
})
// XXX STUB
image.text(gid)
// XXX slect best preview by size...
// XXX
// XXX update classes...
// XXX
}
// load count images around a given image/gid into the given ribbon.
//
// NOTE: this will reload the current image elements...
// NOTE: this is similar to extendRibbon(...) but different in interface...
function loadImages(image, count, ribbon){
// XXX
}
// NOTE: this is signature-compatible with rollRibbon...
// NOTE: this will load data ONLY if it is available, otherwise this
// will have no effect...
// NOTE: this can roll past the currently loaded images (n > images.length)
function rollImages(n, ribbon){
if(n == 0){
return $([])
}
ribbon = ribbon == null ? getRibbon() : $(ribbon)
var images = ribbon.find('.image')
var from = n > 0 ? JSON.parse(ribbon.find('.image').last().attr('gid'))
: JSON.parse(ribbon.find('.image').first().attr('gid'))
var gids = getImageGIDs(from, n)
if(gids.length == 0){
return $([])
}
// truncate the results to the length of images...
if(n > images.length){
gids.reverse().splice(images.length)
gids.reverse()
} else if(Math.abs(n) > images.length){
gids.splice(images.length)
}
if(n < images.length){
images = rollRibbon(gids.length * (n > 0 ? 1 : -1), ribbon)
}
var size = getVisibleImageSize()
images.each(function(i, e){
updateImage($(e), gids[i], size)
})
return images
}
/**********************************************************************
* Modes
*/
// XXX shifting images and unmarking in this mode do not work correctly...
var toggleMarkedOnlyView = createCSSClassToggler('.viewer', 'marked-only',
function(){
var cur = $('.current.image')
// current is marked...
if(cur.hasClass('marked')){
centerImage(null, 'css')
return
}
// there is a marked image in this ribbon...
var target = getImageBefore(cur, null)
if(target.length > 0){
centerImage(focusImage(target), 'css')
return
}
// get marked image from other ribbons...
prevRibbon()
if($('.current.image').hasClass('marked')){
return
}
nextRibbon()
})
// XXX add ability to take all marked images and open them in a separate view...
// XXX should we use the createCSSClassToggler for this?
// XXX revise: does extra stuff...
function toggleImageProportions(mode){
var image = $('.image')
var h = image.outerHeight(true)
var w = image.outerWidth(true)
if(mode == '?'){
return h != w ? 'viewer' : 'square'
// square...
} else if(h != w || mode == 'square'){
var size = Math.min(w, h)
image.css({
width: size,
height: size
})
centerImage(null, 'css')
return 'square'
// viewer size...
} else {
var viewer = $('.viewer')
var W = viewer.innerWidth()
var H = viewer.innerHeight()
if(W > H){
image.css('width', W * h/H)
} else {
image.css('height', H * w/W)
}
centerImage(null, 'css')
return 'viewer'
}
}
/**********************************************************************
* Layout
*/
function focusImage(image){
image.closest('.viewer').find('.current.image').removeClass('current')
return image.addClass('current')
}
/*
// XXX need to split this into two:
// - offset calculator
// - actual move
// XXX this does not account for scale at this point...
// XXX for this to be generic, need a uniform way to get any element scale
// regardless of weather it was scaled directly or is within one or
// several scaled elements...
function alignVia(container, elem, via, valign, halign, mode){
container = $(container)
elem = $(elem)
via = $(via)
valign = valign == null ? 'center' : valign
halign = halign == null ? 'center' : halign
mode = mode == null ? 'animate' : mode
var pos = getRelativeVisualPosition(container, elem)
var dt = pos.top
var dl = pos.left
var target = {}
var t = parseFloat(via.css('top'))
t = !isNaN(t) ? t : 0
var l = parseFloat(via.css('left'))
l = !isNaN(l) ? l : 0
if(valign == 'center'){
var H = container.innerHeight()
var h = elem.outerHeight()
target.top = t - dt + (H - h)/2,
} else if(valign == 'top'){
target.top = t - dt
} else if(valign == 'bottom'){
var h = elem.outerHeight()
target.top = t - dt - h
}
if(halign == 'center'){
var W = container.innerWidth()
var w = elem.outerWidth()
target.left = l - dl + (W - w)/2
} else if(halign == 'left'){
target.left = l - dl
} else if(halign == 'right'){
var w = elem.outerWidth()
target.left = l - dl - w
}
// do the actual work...
if(mode == 'animate'){
via.stop().animate(target, 100, 'linear')
} else {
via.css(target)
}
// XXX ???
return
}
*/
// This appears to work well with scaling...
// XXX make this more configurable...
function centerImage(image, mode){
if(mode == null){
//mode = 'css'
mode = 'animate'
}
if(image == null || image.length == 0){
image = $('.current.image')
}
var viewer = $('.viewer')
// XXX should these be "inner"???
var W = viewer.innerWidth()
var H = viewer.innerHeight()
var ribbons = $('.ribbon-set')
var scale = getElementScale(ribbons)
// NOTE: these are scalable, this needs to get normalized...
var w = image.outerWidth()*scale
var h = image.outerHeight()*scale
var pos = getRelativeVisualPosition(viewer, image)
// zero out top/left if set to anything other than a specific number...
var t = parseFloat(ribbons.css('top'))
t = !isNaN(t) ? t : 0
var l = parseFloat(ribbons.css('left'))
l = !isNaN(l) ? l : 0
var res = {
'top': t - pos.top + (H - h)/2,
'left': l - pos.left + (W - w)/2
}
// do the actual work...
if(mode == 'animate'){
ribbons.stop().animate(res, 100, 'linear')
} else {
ribbons.css(res)
}
$('.viewer').trigger('centeringRibbon', [getRibbon(image), image])
return image
}
// Center a ribbon...
//
// This behaves differently for different ribbons:
// - ribbon containing the target (given) image
// center relative to the .viewer via .ribbon-set
// calls centerImage(...) directly
// both top and left are used...
// - any other ribbon
// center relative to target (given) via the ribbon left
// only left coordinate is changed...
//
// NOTE: image defaults to $('.current.image').
//
// XXX might be good to merge this and centerImage...
// ...or make a generic centering function...
//
// XXX this does not work in marked-only mode...
function centerRibbon(ribbon, image, mode){
if(mode == null){
//mode = 'css'
mode = 'animate'
}
ribbon = $(ribbon)
image = image == null ? $('.current.image') : $(image)
// if centering current ribbon, just center the image...
if(ribbon.find('.image').index(image) >= 0){
centerImage(image, mode)
// XXX should this return a ribbon or the target image???
return ribbon
}
var scale = getElementScale($('.ribbon-set'))
var target = getImageBefore(image, ribbon, null)
if(target.length > 0){
var dl = getRelativeVisualPosition(target, image).left/scale
var l = parseFloat(ribbon.css('left'))
l = !isNaN(l) ? l : 0
l = {left: l + dl - ($('.image').outerWidth()/2)}
} else {
target = ribbon.find('.image').filter(NAV_DEFAULT).first()
var dl = getRelativeVisualPosition(target, image).left/scale
var l = parseFloat(ribbon.css('left'))
l = !isNaN(l) ? l : 0
l = {left: l + dl + ($('.image').outerWidth()/2)}
}
if(mode == 'animate'){
ribbon.stop().animate(l, 100, 'linear')
} else {
ribbons.css(res)
}
$('.viewer').trigger('centeringRibbon', [ribbon, image])
// XXX should this return a ribbon or the target image???
return ribbon
}
// a shorthand...
function centerRibbons(mode){
return $('.ribbon')
.filter(':visible')
.each(function(){ centerRibbon($(this), null, mode) })
}
/**********************************************************************
* Event handlers...
*/
// NOTE: this is on purpose done relative...
function clickHandler(evt){
var img = $(evt.target).closest('.image')
centerImage(focusImage(img))
centerRibbons()
}
/**********************************************************************
* User actions
*/
// NOTE: NAV_ALL might not be practical...
var NAV_ALL = '*'
var NAV_VISIBLE = ':visible'
var NAV_MARKED = '.marked:visible'
var NAV_DEFAULT = NAV_VISIBLE
// basic navigation actions...
function nextImage(n, mode){
mode = mode == null ? NAV_DEFAULT : mode
n = n == null ? 1 : n
var target = $('.current.image').nextAll('.image' + mode)
if(target.length < n){
target = target.last()
flashIndicator('end')
} else {
target = target.eq(n-1)
}
return centerImage(focusImage(target))
}
function prevImage(n, mode){
mode = mode == null ? NAV_DEFAULT : mode
n = n == null ? 1 : n
var target = $('.current.image').prevAll('.image' + mode)
if(target.length < n){
target = target.last()
flashIndicator('start')
} else {
target = target.eq(n-1)
}
return centerImage(focusImage(target))
}
function nextScreenImages(mode){
return nextImage(Math.round(getScreenWidthInImages()), mode)
}
function prevScreenImages(mode){
return prevImage(Math.round(getScreenWidthInImages()), mode)
}
// XXX revise...
function firstImage(mode){
$('.viewer').trigger('requestedFirstImage', [getRibbon()])
mode = mode == null ? NAV_DEFAULT : mode
if($('.current.image').prevAll('.image' + mode).length == 0){
flashIndicator('start')
}
return centerImage(
focusImage(
getRibbon().find('.image').filter(mode).first()))
}
// XXX revise...
function lastImage(mode){
$('.viewer').trigger('requestedLastImage', [getRibbon()])
mode = mode == null ? NAV_DEFAULT : mode
if($('.current.image').nextAll('.image' + mode).length == 0){
flashIndicator('end')
}
return centerImage(
focusImage(
getRibbon().find('.image').filter(mode).last()))
}
// NOTE: if moving is 'next' these will chose the image after the current's order.
// NOTE: if an image with the same order is found, moving argument has no effect.
// XXX get move direction...
function prevRibbon(moving, mode){
mode = mode == null ? NAV_DEFAULT : mode
var cur = $('.current.image')
var target = getImageBefore(cur,
getRibbon(cur).prevAll('.ribbon:visible').first())
if(target.length == 0){
// XXX too complex???
target = getRibbon(cur)
.prevAll('.ribbon:visible').first()
.find('.image' + mode).first()
} else if(moving == 'next' && cur.attr('order') != target.attr('order')){
var next = target.nextAll('.image' + mode).first()
target = next.length > 0 ? next : target
}
return centerImage(focusImage(target))
}
// XXX get move direction...
function nextRibbon(moving, mode){
mode = mode == null ? NAV_DEFAULT : mode
var cur = $('.current.image')
var target = getImageBefore(cur,
getRibbon(cur).nextAll('.ribbon:visible').first())
if(target.length == 0){
// XXX too complex???
target = getRibbon(cur)
.nextAll('.ribbon:visible').first()
.find('.image' + mode).first()
} else if(moving == 'next' && cur.attr('order') != target.attr('order')){
var next = target.nextAll('.image' + mode).first()
target = next.length > 0 ? next : target
}
return centerImage(focusImage(target))
}
function fitNImages(n){
var image = $('.current.image')
var size = image.outerHeight(true)
var viewer = $('.viewer')
var W = viewer.innerWidth()
var H = viewer.innerHeight()
var scale = Math.min(W / (size * n), H / size)
// XXX if animating, the next two likes must be animated together...
setElementScale($('.ribbon-set'), scale)
centerImage(image, 'css')
}
/************************************************** Editor Actions ***/
// XXX add a shift event here...
// XXX get move direction...
function _shiftImageTo(image, direction, moving, force_create_ribbon, mode){
if(image == null){
image = $('.current.image')
}
mode = mode == null ? NAV_DEFAULT : mode
// account move for direction...
// XXX get the value from some place more logical than the argument...
var a = moving == 'prev' ? 'prevAll' : 'nextAll'
var b = moving == 'prev' ? 'nextAll' : 'prevAll'
var target = image[a]('.image' + mode).first()
target = target.length == 0 ? image[b]().first() : target
// XXX should this be in here or coupled later via an event???
flashIndicator(direction)
shiftImage(direction, image, force_create_ribbon)
// XXX does this need to be animated???
return centerImage(focusImage(target), 'css')
}
function shiftImageUp(image, moving){
return _shiftImageTo(image, 'prev', moving)
}
function shiftImageDown(image, moving){
return _shiftImageTo(image, 'next')
}
function shiftImageUpNewRibbon(image, moving){
return _shiftImageTo(image, 'prev', moving, true)
}
function shiftImageDownNewRibbon(image, moving){
return _shiftImageTo(image, 'prev', moving, false)
}
// TODO manual image ordering (shiftLeft/shiftRight functions)
// XXX
/*********************************************************** Marks ***/
// XXX if this unmarks an image in marked-only mode no visible image is
// going to be current...
var toggleImageMark = createCSSClassToggler('.current.image', 'marked')
// mode can be:
// - 'ribbon'
// - 'all'
function removeImageMarks(mode){
// remove marks from current ribbon (default)...
if(mode == 'ribbon' || mode == null){
return getRibbon()
.find('.marked')
.removeClass('marked')
// remove all marks...
} else if(mode == 'all'){
return $('.marked')
.removeClass('marked')
}
}
function markAll(mode){
// remove marks from current ribbon (default)...
if(mode == 'ribbon' || mode == null){
return getRibbon()
.find('.image:not(.marked)')
.addClass('marked')
// remove all marks...
} else if(mode == 'all'){
return $('.image:not(.marked)').addClass('marked')
}
}
// NOTE: this only does it's work in the current ribbon...
function invertImageMarks(){
return getRibbon()
.find('.image')
.toggleClass('marked')
}
// this will toggle marks in the current continuous section of marked
// or unmarked images...
function toggleImageMarkBlock(image){
if(image == null){
image = $('.current.image')
}
// we need to invert this...
var state = toggleImageMark()
var _convert = function(){
if(toggleImageMark(this, '?') == state){
return false
}
toggleImageMark(this, state)
}
image.nextAll('.image').each(_convert)
image.prevAll('.image').each(_convert)
return state
}
/**********************************************************************
* vim:set sw=4 ts=4 : */