photoprism-client-go/vendor/github.com/golang/geo/s2/cell_index.go

499 lines
16 KiB
Go

// Copyright 2020 Google Inc. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package s2
import (
"sort"
)
const (
// A special label indicating that the ContentsIterator done is true.
cellIndexDoneContents = -1
)
// cellIndexNode represents a node in the CellIndex. Cells are organized in a
// tree such that the ancestors of a given node contain that node.
type cellIndexNode struct {
cellID CellID
label int32
parent int32
}
// newCellIndexNode returns a node with the appropriate default values.
func newCellIndexNode() cellIndexNode {
return cellIndexNode{
cellID: 0,
label: cellIndexDoneContents,
parent: -1,
}
}
// A rangeNode represents a range of leaf CellIDs. The range starts at
// startID (a leaf cell) and ends at the startID field of the next
// rangeNode. contents points to the node of the CellIndex cellTree
// representing the cells that overlap this range.
type rangeNode struct {
startID CellID // First leaf cell contained by this range.
contents int32 // Contents of this node (an index within the cell tree).
}
// CellIndexIterator is an iterator that visits the entire set of indexed
// (CellID, label) pairs in an unspecified order.
type CellIndexIterator struct {
// TODO(roberts): Implement
}
// NewCellIndexIterator creates an iterator for the given CellIndex.
func NewCellIndexIterator(index *CellIndex) *CellIndexIterator {
return &CellIndexIterator{}
}
// CellIndexRangeIterator is an iterator that seeks and iterates over a set of
// non-overlapping leaf cell ranges that cover the entire sphere. The indexed
// (CellID, label) pairs that intersect the current leaf cell range can be
// visited using CellIndexContentsIterator (see below).
type CellIndexRangeIterator struct {
rangeNodes []rangeNode
pos int
nonEmpty bool
}
// NewCellIndexRangeIterator creates an iterator for the given CellIndex.
// The iterator is initially *unpositioned*; you must call a positioning method
// such as Begin() or Seek() before accessing its contents.
func NewCellIndexRangeIterator(index *CellIndex) *CellIndexRangeIterator {
return &CellIndexRangeIterator{
rangeNodes: index.rangeNodes,
}
}
// NewCellIndexNonEmptyRangeIterator creates an iterator for the given CellIndex.
// The iterator is initially *unpositioned*; you must call a positioning method such as
// Begin() or Seek() before accessing its contents.
func NewCellIndexNonEmptyRangeIterator(index *CellIndex) *CellIndexRangeIterator {
return &CellIndexRangeIterator{
rangeNodes: index.rangeNodes,
nonEmpty: true,
}
}
// StartID reports the CellID of the start of the current range of leaf CellIDs.
//
// If done is true, this returns the last possible CellID. This property means
// that most loops do not need to test done explicitly.
func (c *CellIndexRangeIterator) StartID() CellID {
return c.rangeNodes[c.pos].startID
}
// LimitID reports the non-inclusive end of the current range of leaf CellIDs.
//
// This assumes the iterator is not done.
func (c *CellIndexRangeIterator) LimitID() CellID {
return c.rangeNodes[c.pos+1].startID
}
// IsEmpty reports if no (CellID, label) pairs intersect this range.
// Also returns true if done() is true.
func (c *CellIndexRangeIterator) IsEmpty() bool {
return c.rangeNodes[c.pos].contents == cellIndexDoneContents
}
// Begin positions the iterator at the first range of leaf cells (if any).
func (c *CellIndexRangeIterator) Begin() {
c.pos = 0
for c.nonEmpty && c.IsEmpty() && !c.Done() {
c.pos++
}
}
// Prev positions the iterator at the previous entry and reports whether it was not
// already positioned at the beginning.
func (c *CellIndexRangeIterator) Prev() bool {
if c.nonEmpty {
return c.nonEmptyPrev()
}
return c.prev()
}
// prev is used to position the iterator at the previous entry without checking
// if nonEmpty is true to prevent unwanted recursion.
func (c *CellIndexRangeIterator) prev() bool {
if c.pos == 0 {
return false
}
c.pos--
return true
}
// Prev positions the iterator at the previous entry, and reports whether it was
// already positioned at the beginning.
func (c *CellIndexRangeIterator) nonEmptyPrev() bool {
for c.prev() {
if !c.IsEmpty() {
return true
}
}
// Return the iterator to its original position.
if c.IsEmpty() && !c.Done() {
c.Next()
}
return false
}
// Next advances the iterator to the next range of leaf cells.
//
// This assumes the iterator is not done.
func (c *CellIndexRangeIterator) Next() {
c.pos++
for c.nonEmpty && c.IsEmpty() && !c.Done() {
c.pos++
}
}
// Advance reports if advancing would leave it positioned on a valid range. If
// the value would not be valid, the positioning is not changed.
func (c *CellIndexRangeIterator) Advance(n int) bool {
// Note that the last element of rangeNodes is a sentinel value.
if n >= len(c.rangeNodes)-1-c.pos {
return false
}
c.pos += n
return true
}
// Finish positions the iterator so that done is true.
func (c *CellIndexRangeIterator) Finish() {
// Note that the last element of rangeNodes is a sentinel value.
c.pos = len(c.rangeNodes) - 1
}
// Done reports if the iterator is positioned beyond the last valid range.
func (c *CellIndexRangeIterator) Done() bool {
return c.pos >= len(c.rangeNodes)-1
}
// Seek positions the iterator at the first range with startID >= target.
// Such an entry always exists as long as "target" is a valid leaf cell.
//
// Note that it is valid to access startID even when done is true.
func (c *CellIndexRangeIterator) Seek(target CellID) {
c.pos = sort.Search(len(c.rangeNodes), func(i int) bool {
return c.rangeNodes[i].startID > target
}) - 1
// Ensure we don't go beyond the beginning.
if c.pos < 0 {
c.pos = 0
}
// Nonempty needs to find the next non-empty entry.
for c.nonEmpty && c.IsEmpty() && !c.Done() {
// c.Next()
c.pos++
}
}
// CellIndexContentsIterator is an iterator that visits the (CellID, label) pairs
// that cover a set of leaf cell ranges (see CellIndexRangeIterator). Note that
// when multiple leaf cell ranges are visited, this iterator only guarantees that
// each result will be reported at least once, i.e. duplicate values may be
// suppressed. If you want duplicate values to be reported again, be sure to call
// Clear first.
//
// In particular, the implementation guarantees that when multiple leaf
// cell ranges are visited in monotonically increasing order, then each
// (CellID, label) pair is reported exactly once.
type CellIndexContentsIterator struct {
// The maximum index within the cellTree slice visited during the
// previous call to StartUnion. This is used to eliminate duplicate
// values when StartUnion is called multiple times.
nodeCutoff int32
// The maximum index within the cellTree visited during the
// current call to StartUnion. This is used to update nodeCutoff.
nextNodeCutoff int32
// The value of startID from the previous call to StartUnion.
// This is used to check whether these values are monotonically
// increasing.
prevStartID CellID
// The cell tree from CellIndex
cellTree []cellIndexNode
// A copy of the current node in the cell tree.
node cellIndexNode
}
// NewCellIndexContentsIterator returns a new contents iterator.
//
// Note that the iterator needs to be positioned using StartUnion before
// it can be safely used.
func NewCellIndexContentsIterator(index *CellIndex) *CellIndexContentsIterator {
it := &CellIndexContentsIterator{
cellTree: index.cellTree,
prevStartID: 0,
nodeCutoff: -1,
nextNodeCutoff: -1,
node: cellIndexNode{label: cellIndexDoneContents},
}
return it
}
// Clear clears all state with respect to which range(s) have been visited.
func (c *CellIndexContentsIterator) Clear() {
c.prevStartID = 0
c.nodeCutoff = -1
c.nextNodeCutoff = -1
c.node.label = cellIndexDoneContents
}
// CellID returns the current CellID.
func (c *CellIndexContentsIterator) CellID() CellID {
return c.node.cellID
}
// Label returns the current Label.
func (c *CellIndexContentsIterator) Label() int32 {
return c.node.label
}
// Next advances the iterator to the next (CellID, label) pair covered by the
// current leaf cell range.
//
// This requires the iterator to not be done.
func (c *CellIndexContentsIterator) Next() {
if c.node.parent <= c.nodeCutoff {
// We have already processed this node and its ancestors.
c.nodeCutoff = c.nextNodeCutoff
c.node.label = cellIndexDoneContents
} else {
c.node = c.cellTree[c.node.parent]
}
}
// Done reports if all (CellID, label) pairs have been visited.
func (c *CellIndexContentsIterator) Done() bool {
return c.node.label == cellIndexDoneContents
}
// StartUnion positions the ContentsIterator at the first (cell_id, label) pair
// that covers the given leaf cell range. Note that when multiple leaf cell
// ranges are visited using the same ContentsIterator, duplicate values
// may be suppressed. If you don't want this behavior, call Reset() first.
func (c *CellIndexContentsIterator) StartUnion(r *CellIndexRangeIterator) {
if r.StartID() < c.prevStartID {
c.nodeCutoff = -1 // Can't automatically eliminate duplicates.
}
c.prevStartID = r.StartID()
contents := r.rangeNodes[r.pos].contents
if contents <= c.nodeCutoff {
c.node.label = cellIndexDoneContents
} else {
c.node = c.cellTree[contents]
}
// When visiting ancestors, we can stop as soon as the node index is smaller
// than any previously visited node index. Because indexes are assigned
// using a preorder traversal, such nodes are guaranteed to have already
// been reported.
c.nextNodeCutoff = contents
}
// CellIndex stores a collection of (CellID, label) pairs.
//
// The CellIDs may be overlapping or contain duplicate values. For example, a
// CellIndex could store a collection of CellUnions, where each CellUnion
// gets its own non-negative int32 label.
//
// Similar to ShapeIndex and PointIndex which map each stored element to an
// identifier, CellIndex stores a label that is typically used to map the
// results of queries back to client's specific data.
//
// The zero value for a CellIndex is sufficient when constructing a CellIndex.
//
// To build a CellIndex where each Cell has a distinct label, call Add for each
// (CellID, label) pair, and then Build the index. For example:
//
// // contents is a mapping of an identifier in my system (restaurantID,
// // vehicleID, etc) to a CellID
// var contents = map[int32]CellID{...}
//
// for key, val := range contents {
// index.Add(val, key)
// }
//
// index.Build()
//
// There is also a helper method that adds all elements of CellUnion with the
// same label:
//
// index.AddCellUnion(cellUnion, label)
//
// Note that the index is not dynamic; the contents of the index cannot be
// changed once it has been built. Adding more after calling Build results in
// undefined behavior of the index.
//
// There are several options for retrieving data from the index. The simplest
// is to use a built-in method such as IntersectingLabels (which returns
// the labels of all cells that intersect a given target CellUnion):
//
// labels := index.IntersectingLabels(targetUnion);
//
// Alternatively, you can use a ClosestCellQuery which computes the cell(s)
// that are closest to a given target geometry.
//
// For example, here is how to find all cells that are closer than
// distanceLimit to a given target point:
//
// query := NewClosestCellQuery(cellIndex, opts)
// target := NewMinDistanceToPointTarget(targetPoint);
// for result := range query.FindCells(target) {
// // result.Distance() is the distance to the target.
// // result.CellID() is the indexed CellID.
// // result.Label() is the label associated with the CellID.
// DoSomething(targetPoint, result);
// }
//
// Internally, the index consists of a set of non-overlapping leaf cell ranges
// that subdivide the sphere and such that each range intersects a particular
// set of (cellID, label) pairs.
//
// Most clients should use either the methods such as VisitIntersectingCells
// and IntersectingLabels, or a helper such as ClosestCellQuery.
type CellIndex struct {
// A tree of (cellID, label) pairs such that if X is an ancestor of Y, then
// X.cellID contains Y.cellID. The contents of a given range of leaf
// cells can be represented by pointing to a node of this tree.
cellTree []cellIndexNode
// The last element of rangeNodes is a sentinel value, which is necessary
// in order to represent the range covered by the previous element.
rangeNodes []rangeNode
}
// Add adds the given CellID and Label to the index.
func (c *CellIndex) Add(id CellID, label int32) {
if label < 0 {
panic("labels must be non-negative")
}
c.cellTree = append(c.cellTree, cellIndexNode{cellID: id, label: label, parent: -1})
}
// AddCellUnion adds all of the elements of the given CellUnion to the index with the same label.
func (c *CellIndex) AddCellUnion(cu CellUnion, label int32) {
if label < 0 {
panic("labels must be non-negative")
}
for _, cell := range cu {
c.Add(cell, label)
}
}
// Build builds the index for use. This method should only be called once.
func (c *CellIndex) Build() {
// To build the cell tree and leaf cell ranges, we maintain a stack of
// (CellID, label) pairs that contain the current leaf cell. This struct
// represents an instruction to push or pop a (cellID, label) pair.
//
// If label >= 0, the (cellID, label) pair is pushed on the stack.
// If CellID == SentinelCellID, a pair is popped from the stack.
// Otherwise the stack is unchanged but a rangeNode is still emitted.
// delta represents an entry in a stack of (CellID, label) pairs used in the
// construction of the CellIndex structure.
type delta struct {
startID CellID
cellID CellID
label int32
}
deltas := make([]delta, 0, 2*len(c.cellTree)+2)
// Create two deltas for each (cellID, label) pair: one to add the pair to
// the stack (at the start of its leaf cell range), and one to remove it from
// the stack (at the end of its leaf cell range).
for _, node := range c.cellTree {
deltas = append(deltas, delta{
startID: node.cellID.RangeMin(),
cellID: node.cellID,
label: node.label,
})
deltas = append(deltas, delta{
startID: node.cellID.RangeMax().Next(),
cellID: SentinelCellID,
label: -1,
})
}
// We also create two special deltas to ensure that a RangeNode is emitted at
// the beginning and end of the CellID range.
deltas = append(deltas, delta{
startID: CellIDFromFace(0).ChildBeginAtLevel(maxLevel),
cellID: CellID(0),
label: -1,
})
deltas = append(deltas, delta{
startID: CellIDFromFace(5).ChildEndAtLevel(maxLevel),
cellID: CellID(0),
label: -1,
})
sort.Slice(deltas, func(i, j int) bool {
// deltas are sorted first by startID, then in reverse order by cellID,
// and then by label. This is necessary to ensure that (1) larger cells
// are pushed on the stack before smaller cells, and (2) cells are popped
// off the stack before any new cells are added.
if si, sj := deltas[i].startID, deltas[j].startID; si != sj {
return si < sj
}
if si, sj := deltas[i].cellID, deltas[j].cellID; si != sj {
return si > sj
}
return deltas[i].label < deltas[j].label
})
// Now walk through the deltas to build the leaf cell ranges and cell tree
// (which is essentially a permanent form of the "stack" described above).
c.cellTree = nil
c.rangeNodes = nil
contents := int32(-1)
for i := 0; i < len(deltas); {
startID := deltas[i].startID
// Process all the deltas associated with the current startID.
for ; i < len(deltas) && deltas[i].startID == startID; i++ {
if deltas[i].label >= 0 {
c.cellTree = append(c.cellTree, cellIndexNode{
cellID: deltas[i].cellID,
label: deltas[i].label,
parent: contents})
contents = int32(len(c.cellTree) - 1)
} else if deltas[i].cellID == SentinelCellID {
contents = c.cellTree[contents].parent
}
}
c.rangeNodes = append(c.rangeNodes, rangeNode{startID, contents})
}
}
// TODO(roberts): Differences from C++
// IntersectingLabels
// VisitIntersectingCells
// CellIndexIterator