update

ad.go

@@ -0,0 +1,7 @@
+package main
+
+type Ad interface {
+	Features() []Feature
+	Weight(Feature) float32
+	Bid() float32
+}

feature.go

@@ -0,0 +1,36 @@
+package main
+
+var endOfPosting PagePostingItem = PagePostingItem{
+	PID: lastPID,
+}
+
+type PagePostingItem struct {
+	PID    int64
+	Weight float32
+}
+
+type PagePostingList struct {
+	// the end of posting items must be endOfPostingItem
+	Items []PagePostingItem
+	// meta info
+	MaxWeight float32
+}
+
+type Feature string // 特征名称，包含特征值？
+
+type FeatureCursor struct {
+	feature Feature
+	posting *PagePostingList
+	current int
+}
+
+func (c *FeatureCursor) CurrentPage() PagePostingItem {
+	return c.posting.Items[c.current]
+}
+
+func (c *FeatureCursor) Beyond(PID int64) {
+	// page IDs are sorted in non-decreasing order
+	for c.CurrentPage().PID < PID {
+		c.current++
+	}
+}

forecast.go

@@ -0,0 +1,133 @@
+package main
+
+import "sort"
+
+type ImpressionForecastor struct {
+	pages          map[int64]PageMeta
+	featureToPages map[Feature]*PagePostingList
+}
+
+func (i *ImpressionForecastor) forecast(ad Ad) int {
+	impressions := 0
+
+	searcher := &Searcher{
+		pages:         i.pages,
+		feature2Pages: i.featureToPages,
+		currentPage:   0,
+		ad:            ad,
+	}
+
+	searcher.initCursors()
+	pageID := searcher.nextCandidate()
+	for pageID < lastPID {
+		p := i.pages[pageID]
+		score := i.score(ad, p)
+		if score > p.minScore {
+			impressions += p.impressions
+		}
+		pageID = searcher.nextCandidate()
+	}
+
+	return impressions
+}
+
+func (i *ImpressionForecastor) score(ad Ad, page PageMeta) float32 {
+	pageWeight := make(map[Feature]float32, len(page.features))
+	for i, f := range page.features {
+		pageWeight[f] = page.weights[i]
+	}
+
+	sim := float32(0)
+	for _, f := range ad.Features() {
+		sim += pageWeight[f] * ad.Weight(f)
+	}
+
+	return sim * ad.Bid()
+}
+
+type Searcher struct {
+	pages         map[int64]PageMeta
+	feature2Pages map[Feature]*PagePostingList
+
+	currentPage int64
+	ad          Ad
+	features    []*FeatureCursor
+}
+
+func (s *Searcher) initCursors() {
+	s.currentPage = 0
+
+	ad := s.ad
+	features := ad.Features()
+	cursors := make([]*FeatureCursor, len(features))
+	for i, f := range features {
+
+		cursors[i] = &FeatureCursor{
+			feature: f,
+			posting: s.feature2Pages[f],
+			current: 0,
+		}
+	}
+
+	s.features = cursors
+}
+
+func (s *Searcher) nextCandidate() int64 {
+	for {
+
+		s.sortFeatures(s.features)
+		pivot := s.findPivotFeature(s.features)
+
+		if pivot == -1 {
+			return lastPID
+		}
+		pivotPID := s.features[pivot].CurrentPage().PID
+		if pivotPID == lastPID {
+			return lastPID
+		}
+
+		if pivotPID == s.currentPage {
+			f := s.pickFeature(s.features[0 : pivot+1])
+			s.features[f].Beyond(pivotPID + 1)
+		} else {
+			if s.features[0].CurrentPage().PID == pivotPID {
+				s.currentPage = pivotPID
+				return s.currentPage
+			} else {
+				f := s.pickFeature(s.features[0 : pivot+1])
+				s.features[f].Beyond(pivotPID)
+			}
+		}
+	}
+}
+
+func (s *Searcher) sortFeatures(features []*FeatureCursor) {
+	sort.Slice(features, func(i, j int) bool {
+		return features[i].CurrentPage().PID < features[j].CurrentPage().PID
+	})
+}
+
+func (s *Searcher) findPivotFeature(cursors []*FeatureCursor) int {
+	n := len(cursors)
+	ub := float32(0)
+	ad := s.ad
+
+	minScoreP := s.pages[s.currentPage].minScore
+	threshold := minScoreP / ad.Bid()
+
+	for i := 0; i < n; i++ {
+
+		f := cursors[i].feature
+		maxWeight := s.feature2Pages[f].MaxWeight
+		ub += ad.Weight(f) * maxWeight
+		if ub >= threshold {
+			return i
+		}
+	}
+
+	return -1
+}
+
+func (s *Searcher) pickFeature(cursors []*FeatureCursor) int {
+	return 0
+}

go.mod

@@ -0,0 +1,3 @@
+module github.com/csimplestring/ca-imp-forecast
+
+go 1.17

go.sum

@@ -0,0 +1,10 @@
+github.com/allegro/bigcache/v3 v3.0.0/go.mod h1:t5TAJn1B9qvf/VlJrSM1r6NlFAYoFDubYUsCuIO9nUQ=
+github.com/csimplestring/go-left-right v0.0.3/go.mod h1:ASS5AW8dwk3v6VxpOPQ6wI/N68ibyAMAPW1VTAWQpx8=
+github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/linxGnu/go-adder v0.2.0/go.mod h1:t9zC6P+vQpG5iEtw4sinwsVqe52AuFvL4P1dVXHD46c=
+github.com/philpearl/stringbank v1.1.0/go.mod h1:0V0f9Ba79DpIl4FTfotL+7IJ+etELdRQIcHJY2nX/+w=
+github.com/philpearl/symboltab v1.1.4/go.mod h1:vOdgHoGZAUqGg1apOcrqjuAxFERPUJ03PLfL6pvxNBg=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
+github.com/valyala/fastrand v1.0.0/go.mod h1:HWqCzkrkg6QXT8V2EXWvXCoow7vLwOFN002oeRzjapQ=

page.go

@@ -0,0 +1,15 @@
+package main
+
+import (
+	"math"
+)
+
+// PageID is generated by increasing order of minScore: smaller ID, smaller score
+const lastPID = math.MaxInt64
+
+type PageMeta struct {
+	impressions int
+	minScore    float32
+	features    []Feature
+	weights     []float32
+}