2019-03-07 01:55:38 +03:00
|
|
|
package brotli
|
|
|
|
|
|
|
|
/* NOLINT(build/header_guard) */
|
|
|
|
/* Copyright 2013 Google Inc. All Rights Reserved.
|
|
|
|
|
|
|
|
Distributed under MIT license.
|
|
|
|
See file LICENSE for detail or copy at https://opensource.org/licenses/MIT
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Computes the bit cost reduction by combining out[idx1] and out[idx2] and if
|
|
|
|
it is below a threshold, stores the pair (idx1, idx2) in the *pairs queue. */
|
2019-03-16 03:24:40 +03:00
|
|
|
func compareAndPushToQueueLiteral(out []histogramLiteral, cluster_size []uint32, idx1 uint32, idx2 uint32, max_num_pairs uint, pairs []histogramPair, num_pairs *uint) {
|
2019-03-07 01:55:38 +03:00
|
|
|
var is_good_pair bool = false
|
2019-03-15 22:05:31 +03:00
|
|
|
var p histogramPair
|
2019-03-07 01:55:38 +03:00
|
|
|
p.idx2 = 0
|
|
|
|
p.idx1 = p.idx2
|
|
|
|
p.cost_combo = 0
|
|
|
|
p.cost_diff = p.cost_combo
|
|
|
|
if idx1 == idx2 {
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
if idx2 < idx1 {
|
|
|
|
var t uint32 = idx2
|
|
|
|
idx2 = idx1
|
|
|
|
idx1 = t
|
|
|
|
}
|
|
|
|
|
|
|
|
p.idx1 = idx1
|
|
|
|
p.idx2 = idx2
|
2019-03-15 22:05:31 +03:00
|
|
|
p.cost_diff = 0.5 * clusterCostDiff(uint(cluster_size[idx1]), uint(cluster_size[idx2]))
|
2019-03-07 01:55:38 +03:00
|
|
|
p.cost_diff -= out[idx1].bit_cost_
|
|
|
|
p.cost_diff -= out[idx2].bit_cost_
|
|
|
|
|
|
|
|
if out[idx1].total_count_ == 0 {
|
|
|
|
p.cost_combo = out[idx2].bit_cost_
|
|
|
|
is_good_pair = true
|
|
|
|
} else if out[idx2].total_count_ == 0 {
|
|
|
|
p.cost_combo = out[idx1].bit_cost_
|
|
|
|
is_good_pair = true
|
|
|
|
} else {
|
|
|
|
var threshold float64
|
|
|
|
if *num_pairs == 0 {
|
|
|
|
threshold = 1e99
|
|
|
|
} else {
|
|
|
|
threshold = brotli_max_double(0.0, pairs[0].cost_diff)
|
|
|
|
}
|
2019-03-16 03:24:40 +03:00
|
|
|
var combo histogramLiteral = out[idx1]
|
2019-03-07 01:55:38 +03:00
|
|
|
var cost_combo float64
|
2019-03-16 03:24:40 +03:00
|
|
|
histogramAddHistogramLiteral(&combo, &out[idx2])
|
2019-03-10 04:11:32 +03:00
|
|
|
cost_combo = populationCostLiteral(&combo)
|
2019-03-07 01:55:38 +03:00
|
|
|
if cost_combo < threshold-p.cost_diff {
|
|
|
|
p.cost_combo = cost_combo
|
|
|
|
is_good_pair = true
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if is_good_pair {
|
|
|
|
p.cost_diff += p.cost_combo
|
2019-03-15 22:05:31 +03:00
|
|
|
if *num_pairs > 0 && histogramPairIsLess(&pairs[0], &p) {
|
2019-03-07 01:55:38 +03:00
|
|
|
/* Replace the top of the queue if needed. */
|
|
|
|
if *num_pairs < max_num_pairs {
|
|
|
|
pairs[*num_pairs] = pairs[0]
|
|
|
|
(*num_pairs)++
|
|
|
|
}
|
|
|
|
|
|
|
|
pairs[0] = p
|
|
|
|
} else if *num_pairs < max_num_pairs {
|
|
|
|
pairs[*num_pairs] = p
|
|
|
|
(*num_pairs)++
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-03-16 03:24:40 +03:00
|
|
|
func histogramCombineLiteral(out []histogramLiteral, cluster_size []uint32, symbols []uint32, clusters []uint32, pairs []histogramPair, num_clusters uint, symbols_size uint, max_clusters uint, max_num_pairs uint) uint {
|
2019-03-07 01:55:38 +03:00
|
|
|
var cost_diff_threshold float64 = 0.0
|
|
|
|
var min_cluster_size uint = 1
|
|
|
|
var num_pairs uint = 0
|
|
|
|
{
|
|
|
|
/* We maintain a vector of histogram pairs, with the property that the pair
|
|
|
|
with the maximum bit cost reduction is the first. */
|
|
|
|
var idx1 uint
|
|
|
|
for idx1 = 0; idx1 < num_clusters; idx1++ {
|
|
|
|
var idx2 uint
|
|
|
|
for idx2 = idx1 + 1; idx2 < num_clusters; idx2++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
compareAndPushToQueueLiteral(out, cluster_size, clusters[idx1], clusters[idx2], max_num_pairs, pairs[0:], &num_pairs)
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for num_clusters > min_cluster_size {
|
|
|
|
var best_idx1 uint32
|
|
|
|
var best_idx2 uint32
|
|
|
|
var i uint
|
|
|
|
if pairs[0].cost_diff >= cost_diff_threshold {
|
|
|
|
cost_diff_threshold = 1e99
|
|
|
|
min_cluster_size = max_clusters
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Take the best pair from the top of heap. */
|
|
|
|
best_idx1 = pairs[0].idx1
|
|
|
|
|
|
|
|
best_idx2 = pairs[0].idx2
|
2019-03-16 03:24:40 +03:00
|
|
|
histogramAddHistogramLiteral(&out[best_idx1], &out[best_idx2])
|
2019-03-07 01:55:38 +03:00
|
|
|
out[best_idx1].bit_cost_ = pairs[0].cost_combo
|
|
|
|
cluster_size[best_idx1] += cluster_size[best_idx2]
|
|
|
|
for i = 0; i < symbols_size; i++ {
|
|
|
|
if symbols[i] == best_idx2 {
|
|
|
|
symbols[i] = best_idx1
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for i = 0; i < num_clusters; i++ {
|
|
|
|
if clusters[i] == best_idx2 {
|
|
|
|
copy(clusters[i:], clusters[i+1:][:num_clusters-i-1])
|
|
|
|
break
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
num_clusters--
|
|
|
|
{
|
|
|
|
/* Remove pairs intersecting the just combined best pair. */
|
|
|
|
var copy_to_idx uint = 0
|
|
|
|
for i = 0; i < num_pairs; i++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
var p *histogramPair = &pairs[i]
|
2019-03-07 01:55:38 +03:00
|
|
|
if p.idx1 == best_idx1 || p.idx2 == best_idx1 || p.idx1 == best_idx2 || p.idx2 == best_idx2 {
|
|
|
|
/* Remove invalid pair from the queue. */
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
2019-03-15 22:05:31 +03:00
|
|
|
if histogramPairIsLess(&pairs[0], p) {
|
2019-03-07 01:55:38 +03:00
|
|
|
/* Replace the top of the queue if needed. */
|
2019-03-15 22:05:31 +03:00
|
|
|
var front histogramPair = pairs[0]
|
2019-03-07 01:55:38 +03:00
|
|
|
pairs[0] = *p
|
|
|
|
pairs[copy_to_idx] = front
|
|
|
|
} else {
|
|
|
|
pairs[copy_to_idx] = *p
|
|
|
|
}
|
|
|
|
|
|
|
|
copy_to_idx++
|
|
|
|
}
|
|
|
|
|
|
|
|
num_pairs = copy_to_idx
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Push new pairs formed with the combined histogram to the heap. */
|
|
|
|
for i = 0; i < num_clusters; i++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
compareAndPushToQueueLiteral(out, cluster_size, best_idx1, clusters[i], max_num_pairs, pairs[0:], &num_pairs)
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return num_clusters
|
|
|
|
}
|
|
|
|
|
|
|
|
/* What is the bit cost of moving histogram from cur_symbol to candidate. */
|
2019-03-16 03:24:40 +03:00
|
|
|
func histogramBitCostDistanceLiteral(histogram *histogramLiteral, candidate *histogramLiteral) float64 {
|
2019-03-07 01:55:38 +03:00
|
|
|
if histogram.total_count_ == 0 {
|
|
|
|
return 0.0
|
|
|
|
} else {
|
2019-03-16 03:24:40 +03:00
|
|
|
var tmp histogramLiteral = *histogram
|
|
|
|
histogramAddHistogramLiteral(&tmp, candidate)
|
2019-03-10 04:11:32 +03:00
|
|
|
return populationCostLiteral(&tmp) - candidate.bit_cost_
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Find the best 'out' histogram for each of the 'in' histograms.
|
|
|
|
When called, clusters[0..num_clusters) contains the unique values from
|
|
|
|
symbols[0..in_size), but this property is not preserved in this function.
|
|
|
|
Note: we assume that out[]->bit_cost_ is already up-to-date. */
|
2019-03-16 03:24:40 +03:00
|
|
|
func histogramRemapLiteral(in []histogramLiteral, in_size uint, clusters []uint32, num_clusters uint, out []histogramLiteral, symbols []uint32) {
|
2019-03-07 01:55:38 +03:00
|
|
|
var i uint
|
|
|
|
for i = 0; i < in_size; i++ {
|
|
|
|
var best_out uint32
|
|
|
|
if i == 0 {
|
|
|
|
best_out = symbols[0]
|
|
|
|
} else {
|
|
|
|
best_out = symbols[i-1]
|
|
|
|
}
|
2019-03-15 22:05:31 +03:00
|
|
|
var best_bits float64 = histogramBitCostDistanceLiteral(&in[i], &out[best_out])
|
2019-03-07 01:55:38 +03:00
|
|
|
var j uint
|
|
|
|
for j = 0; j < num_clusters; j++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
var cur_bits float64 = histogramBitCostDistanceLiteral(&in[i], &out[clusters[j]])
|
2019-03-07 01:55:38 +03:00
|
|
|
if cur_bits < best_bits {
|
|
|
|
best_bits = cur_bits
|
|
|
|
best_out = clusters[j]
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
symbols[i] = best_out
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Recompute each out based on raw and symbols. */
|
|
|
|
for i = 0; i < num_clusters; i++ {
|
2019-03-16 03:24:40 +03:00
|
|
|
histogramClearLiteral(&out[clusters[i]])
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
for i = 0; i < in_size; i++ {
|
2019-03-16 03:24:40 +03:00
|
|
|
histogramAddHistogramLiteral(&out[symbols[i]], &in[i])
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Reorders elements of the out[0..length) array and changes values in
|
|
|
|
symbols[0..length) array in the following way:
|
|
|
|
* when called, symbols[] contains indexes into out[], and has N unique
|
|
|
|
values (possibly N < length)
|
|
|
|
* on return, symbols'[i] = f(symbols[i]) and
|
|
|
|
out'[symbols'[i]] = out[symbols[i]], for each 0 <= i < length,
|
|
|
|
where f is a bijection between the range of symbols[] and [0..N), and
|
|
|
|
the first occurrences of values in symbols'[i] come in consecutive
|
|
|
|
increasing order.
|
|
|
|
Returns N, the number of unique values in symbols[]. */
|
|
|
|
|
2019-03-15 22:05:31 +03:00
|
|
|
var histogramReindexLiteral_kInvalidIndex uint32 = BROTLI_UINT32_MAX
|
2019-03-07 01:55:38 +03:00
|
|
|
|
2019-03-16 03:24:40 +03:00
|
|
|
func histogramReindexLiteral(out []histogramLiteral, symbols []uint32, length uint) uint {
|
2019-03-07 01:55:38 +03:00
|
|
|
var new_index []uint32 = make([]uint32, length)
|
|
|
|
var next_index uint32
|
2019-03-16 03:24:40 +03:00
|
|
|
var tmp []histogramLiteral
|
2019-03-07 01:55:38 +03:00
|
|
|
var i uint
|
|
|
|
for i = 0; i < length; i++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
new_index[i] = histogramReindexLiteral_kInvalidIndex
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
next_index = 0
|
|
|
|
for i = 0; i < length; i++ {
|
2019-03-15 22:05:31 +03:00
|
|
|
if new_index[symbols[i]] == histogramReindexLiteral_kInvalidIndex {
|
2019-03-07 01:55:38 +03:00
|
|
|
new_index[symbols[i]] = next_index
|
|
|
|
next_index++
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* TODO: by using idea of "cycle-sort" we can avoid allocation of
|
|
|
|
tmp and reduce the number of copying by the factor of 2. */
|
2019-03-16 03:24:40 +03:00
|
|
|
tmp = make([]histogramLiteral, next_index)
|
2019-03-07 01:55:38 +03:00
|
|
|
|
|
|
|
next_index = 0
|
|
|
|
for i = 0; i < length; i++ {
|
|
|
|
if new_index[symbols[i]] == next_index {
|
|
|
|
tmp[next_index] = out[symbols[i]]
|
|
|
|
next_index++
|
|
|
|
}
|
|
|
|
|
|
|
|
symbols[i] = new_index[symbols[i]]
|
|
|
|
}
|
|
|
|
|
|
|
|
new_index = nil
|
|
|
|
for i = 0; uint32(i) < next_index; i++ {
|
|
|
|
out[i] = tmp[i]
|
|
|
|
}
|
|
|
|
|
|
|
|
tmp = nil
|
|
|
|
return uint(next_index)
|
|
|
|
}
|
|
|
|
|
2019-03-16 03:24:40 +03:00
|
|
|
func clusterHistogramsLiteral(in []histogramLiteral, in_size uint, max_histograms uint, out []histogramLiteral, out_size *uint, histogram_symbols []uint32) {
|
2019-03-07 01:55:38 +03:00
|
|
|
var cluster_size []uint32 = make([]uint32, in_size)
|
|
|
|
var clusters []uint32 = make([]uint32, in_size)
|
|
|
|
var num_clusters uint = 0
|
|
|
|
var max_input_histograms uint = 64
|
|
|
|
var pairs_capacity uint = max_input_histograms * max_input_histograms / 2
|
2019-03-15 22:05:31 +03:00
|
|
|
var pairs []histogramPair = make([]histogramPair, (pairs_capacity + 1))
|
2019-03-07 01:55:38 +03:00
|
|
|
var i uint
|
|
|
|
|
|
|
|
/* For the first pass of clustering, we allow all pairs. */
|
|
|
|
for i = 0; i < in_size; i++ {
|
|
|
|
cluster_size[i] = 1
|
|
|
|
}
|
|
|
|
|
|
|
|
for i = 0; i < in_size; i++ {
|
|
|
|
out[i] = in[i]
|
2019-03-10 04:11:32 +03:00
|
|
|
out[i].bit_cost_ = populationCostLiteral(&in[i])
|
2019-03-07 01:55:38 +03:00
|
|
|
histogram_symbols[i] = uint32(i)
|
|
|
|
}
|
|
|
|
|
|
|
|
for i = 0; i < in_size; i += max_input_histograms {
|
|
|
|
var num_to_combine uint = brotli_min_size_t(in_size-i, max_input_histograms)
|
|
|
|
var num_new_clusters uint
|
|
|
|
var j uint
|
|
|
|
for j = 0; j < num_to_combine; j++ {
|
|
|
|
clusters[num_clusters+j] = uint32(i + j)
|
|
|
|
}
|
|
|
|
|
2019-03-15 22:05:31 +03:00
|
|
|
num_new_clusters = histogramCombineLiteral(out, cluster_size, histogram_symbols[i:], clusters[num_clusters:], pairs, num_to_combine, num_to_combine, max_histograms, pairs_capacity)
|
2019-03-07 01:55:38 +03:00
|
|
|
num_clusters += num_new_clusters
|
|
|
|
}
|
|
|
|
{
|
|
|
|
/* For the second pass, we limit the total number of histogram pairs.
|
|
|
|
After this limit is reached, we only keep searching for the best pair. */
|
|
|
|
var max_num_pairs uint = brotli_min_size_t(64*num_clusters, (num_clusters/2)*num_clusters)
|
|
|
|
if pairs_capacity < (max_num_pairs + 1) {
|
|
|
|
var _new_size uint
|
|
|
|
if pairs_capacity == 0 {
|
|
|
|
_new_size = max_num_pairs + 1
|
|
|
|
} else {
|
|
|
|
_new_size = pairs_capacity
|
|
|
|
}
|
2019-03-15 22:05:31 +03:00
|
|
|
var new_array []histogramPair
|
2019-03-07 01:55:38 +03:00
|
|
|
for _new_size < (max_num_pairs + 1) {
|
|
|
|
_new_size *= 2
|
|
|
|
}
|
2019-03-15 22:05:31 +03:00
|
|
|
new_array = make([]histogramPair, _new_size)
|
2019-03-07 01:55:38 +03:00
|
|
|
if pairs_capacity != 0 {
|
|
|
|
copy(new_array, pairs[:pairs_capacity])
|
|
|
|
}
|
|
|
|
|
|
|
|
pairs = new_array
|
|
|
|
pairs_capacity = _new_size
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Collapse similar histograms. */
|
2019-03-15 22:05:31 +03:00
|
|
|
num_clusters = histogramCombineLiteral(out, cluster_size, histogram_symbols, clusters, pairs, num_clusters, in_size, max_histograms, max_num_pairs)
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
pairs = nil
|
|
|
|
cluster_size = nil
|
|
|
|
|
|
|
|
/* Find the optimal map from original histograms to the final ones. */
|
2019-03-15 22:05:31 +03:00
|
|
|
histogramRemapLiteral(in, in_size, clusters, num_clusters, out, histogram_symbols)
|
2019-03-07 01:55:38 +03:00
|
|
|
|
|
|
|
clusters = nil
|
|
|
|
|
|
|
|
/* Convert the context map to a canonical form. */
|
2019-03-15 22:05:31 +03:00
|
|
|
*out_size = histogramReindexLiteral(out, histogram_symbols, in_size)
|
2019-03-07 01:55:38 +03:00
|
|
|
}
|