all: replace publicly facing *rand.Rand with rand.Source

graph/community/bisect.go

@@ -62,7 +62,7 @@ func Weight(g ReducedGraph) float64 {
 // ModularScore returns a modularized scoring function for Profile based on the
 // graph g and the given score function. The effort parameter determines how
 // many attempts will be made to get an improved score for any given resolution.
-func ModularScore(g graph.Graph, score func(ReducedGraph) float64, effort int, src *rand.Rand) func(float64) (float64, Reduced) {
+func ModularScore(g graph.Graph, score func(ReducedGraph) float64, effort int, src rand.Source) func(float64) (float64, Reduced) {
 	return func(resolution float64) (float64, Reduced) {
 		max := math.Inf(-1)
 		var best Reduced
@@ -108,7 +108,7 @@ func WeightMultiplex(g ReducedMultiplex) float64 {
 // ModularMultiplexScore returns a modularized scoring function for Profile based
 // on the graph g and the given score function. The effort parameter determines how
 // many attempts will be made to get an improved score for any given resolution.
-func ModularMultiplexScore(g Multiplex, weights []float64, all bool, score func(ReducedMultiplex) float64, effort int, src *rand.Rand) func(float64) (float64, Reduced) {
+func ModularMultiplexScore(g Multiplex, weights []float64, all bool, score func(ReducedMultiplex) float64, effort int, src rand.Source) func(float64) (float64, Reduced) {
 	return func(resolution float64) (float64, Reduced) {
 		max := math.Inf(-1)
 		var best Reduced

graph/community/louvain_common.go

@@ -82,7 +82,7 @@ type ReducedGraph interface {
 //
 // graph.Undirect may be used as a shim to allow modularization of
 // directed graphs with the undirected modularity function.
-func Modularize(g graph.Graph, resolution float64, src *rand.Rand) ReducedGraph {
+func Modularize(g graph.Graph, resolution float64, src rand.Source) ReducedGraph {
 	switch g := g.(type) {
 	case graph.Undirected:
 		return louvainUndirected(g, resolution, src)
@@ -192,7 +192,7 @@ type ReducedMultiplex interface {
 //
 // graph.Undirect may be used as a shim to allow modularization of
 // directed graphs with the undirected modularity function.
-func ModularizeMultiplex(g Multiplex, weights, resolutions []float64, all bool, src *rand.Rand) ReducedMultiplex {
+func ModularizeMultiplex(g Multiplex, weights, resolutions []float64, all bool, src rand.Source) ReducedMultiplex {
 	if weights != nil && len(weights) != g.Depth() {
 		panic("community: weights vector length mismatch")
 	}

graph/community/louvain_directed.go

@@ -79,14 +79,14 @@ func qDirected(g graph.Directed, communities [][]graph.Node, resolution float64)
 // resolution using the Louvain algorithm. If src is nil, rand.Intn is used
 // as the random generator. louvainDirected will panic if g has any edge with negative
 // edge weight.
-func louvainDirected(g graph.Directed, resolution float64, src *rand.Rand) ReducedGraph {
+func louvainDirected(g graph.Directed, resolution float64, src rand.Source) ReducedGraph {
 	// See louvain.tex for a detailed description
 	// of the algorithm used here.
 
 	c := reduceDirected(g, nil)
 	rnd := rand.Intn
 	if src != nil {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 	for {
 		l := newDirectedLocalMover(c, c.communities, resolution)

graph/community/louvain_directed_multiplex.go

@@ -168,7 +168,7 @@ func (g DirectedLayers) Layer(l int) graph.Directed { return g[l] }
 // edge weight that does not sign-match the layer weight.
 //
 // graph.Undirect may be used as a shim to allow modularization of directed graphs.
-func louvainDirectedMultiplex(g DirectedMultiplex, weights, resolutions []float64, all bool, src *rand.Rand) *ReducedDirectedMultiplex {
+func louvainDirectedMultiplex(g DirectedMultiplex, weights, resolutions []float64, all bool, src rand.Source) *ReducedDirectedMultiplex {
 	if weights != nil && len(weights) != g.Depth() {
 		panic("community: weights vector length mismatch")
 	}
@@ -182,7 +182,7 @@ func louvainDirectedMultiplex(g DirectedMultiplex, weights, resolutions []float6
 	c := reduceDirectedMultiplex(g, nil, weights)
 	rnd := rand.Intn
 	if src != nil {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 	for {
 		l := newDirectedMultiplexLocalMover(c, c.communities, weights, resolutions, all)

graph/community/louvain_undirected.go

@@ -77,14 +77,14 @@ func qUndirected(g graph.Undirected, communities [][]graph.Node, resolution floa
 // weight.
 //
 // graph.Undirect may be used as a shim to allow modularization of directed graphs.
-func louvainUndirected(g graph.Undirected, resolution float64, src *rand.Rand) *ReducedUndirected {
+func louvainUndirected(g graph.Undirected, resolution float64, src rand.Source) *ReducedUndirected {
 	// See louvain.tex for a detailed description
 	// of the algorithm used here.
 
 	c := reduceUndirected(g, nil)
 	rnd := rand.Intn
 	if src != nil {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 	for {
 		l := newUndirectedLocalMover(c, c.communities, resolution)

graph/community/louvain_undirected_multiplex.go

@@ -165,7 +165,7 @@ func (g UndirectedLayers) Layer(l int) graph.Undirected { return g[l] }
 // edge weight that does not sign-match the layer weight.
 //
 // graph.Undirect may be used as a shim to allow modularization of directed graphs.
-func louvainUndirectedMultiplex(g UndirectedMultiplex, weights, resolutions []float64, all bool, src *rand.Rand) *ReducedUndirectedMultiplex {
+func louvainUndirectedMultiplex(g UndirectedMultiplex, weights, resolutions []float64, all bool, src rand.Source) *ReducedUndirectedMultiplex {
 	if weights != nil && len(weights) != g.Depth() {
 		panic("community: weights vector length mismatch")
 	}
@@ -179,7 +179,7 @@ func louvainUndirectedMultiplex(g UndirectedMultiplex, weights, resolutions []fl
 	c := reduceUndirectedMultiplex(g, nil, weights)
 	rnd := rand.Intn
 	if src != nil {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 	for {
 		l := newUndirectedMultiplexLocalMover(c, c.communities, weights, resolutions, all)

graph/graphs/gen/batagelj_brandes.go

@@ -21,7 +21,7 @@ import (
 // between nodes are formed with the probability, p. If src is not nil it is used
 // as the random source, otherwise rand.Float64 is used. The graph is constructed
 // in O(n+m) time where m is the number of edges added.
-func Gnp(dst GraphBuilder, n int, p float64, src *rand.Rand) error {
+func Gnp(dst GraphBuilder, n int, p float64, src rand.Source) error {
 	if p == 0 {
 		return nil
 	}
@@ -32,7 +32,7 @@ func Gnp(dst GraphBuilder, n int, p float64, src *rand.Rand) error {
 	if src == nil {
 		r = rand.Float64
 	} else {
-		r = src.Float64
+		r = rand.New(src).Float64
 	}
 
 	for i := 0; i < n; i++ {
@@ -87,7 +87,7 @@ func edgeNodesFor(i int) (v, w simple.Node) {
 // order n and size m. If src is not nil it is used as the random source,
 // otherwise rand.Intn is used. The graph is constructed in O(m) expected
 // time for m ≤ (n choose 2)/2.
-func Gnm(dst GraphBuilder, n, m int, src *rand.Rand) error {
+func Gnm(dst GraphBuilder, n, m int, src rand.Source) error {
 	if m == 0 {
 		return nil
 	}
@@ -108,7 +108,7 @@ func Gnm(dst GraphBuilder, n, m int, src *rand.Rand) error {
 	if src == nil {
 		rnd = rand.Intn
 	} else {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 
 	for i := 0; i < n; i++ {
@@ -153,7 +153,7 @@ func Gnm(dst GraphBuilder, n, m int, src *rand.Rand) error {
 // The graph is constructed in O(nd) time.
 //
 // The algorithm used is described in http://algo.uni-konstanz.de/publications/bb-eglrn-05.pdf
-func SmallWorldsBB(dst GraphBuilder, n, d int, p float64, src *rand.Rand) error {
+func SmallWorldsBB(dst GraphBuilder, n, d int, p float64, src rand.Source) error {
 	if d < 1 || d > (n-1)/2 {
 		return fmt.Errorf("gen: bad degree: d=%d", d)
 	}
@@ -171,8 +171,9 @@ func SmallWorldsBB(dst GraphBuilder, n, d int, p float64, src *rand.Rand) error
 		rnd = rand.Float64
 		rndN = rand.Intn
 	} else {
-		rnd = src.Float64
-		rndN = src.Intn
+		r := rand.New(src)
+		rnd = r.Float64
+		rndN = r.Intn
 	}
 
 	hasEdge := dst.HasEdgeBetween
@@ -298,15 +299,15 @@ func SmallWorldsBB(dst GraphBuilder, n, d int, p float64, src *rand.Rand) error
 // The graph is constructed in O(nd) — O(n+m) — time.
 //
 // The algorithm used is described in http://algo.uni-konstanz.de/publications/bb-eglrn-05.pdf
-func PowerLaw(dst graph.MultigraphBuilder, n, d int, src *rand.Rand) error {
+func PowerLaw(dst graph.MultigraphBuilder, n, d int, src rand.Source) error {
 	if d < 1 {
 		return fmt.Errorf("gen: bad minimum degree: d=%d", d)
 	}
 	var rnd func(int) int
 	if src == nil {
 		rnd = rand.Intn
 	} else {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 
 	m := make([]graph.Node, 2*n*d)
@@ -333,15 +334,15 @@ func PowerLaw(dst graph.MultigraphBuilder, n, d int, src *rand.Rand) error {
 // The graph is constructed in O(nd) — O(n+m) — time.
 //
 // The algorithm used is described in http://algo.uni-konstanz.de/publications/bb-eglrn-05.pdf
-func BipartitePowerLaw(dst graph.MultigraphBuilder, n, d int, src *rand.Rand) (p1, p2 []graph.Node, err error) {
+func BipartitePowerLaw(dst graph.MultigraphBuilder, n, d int, src rand.Source) (p1, p2 []graph.Node, err error) {
 	if d < 1 {
 		return nil, nil, fmt.Errorf("gen: bad minimum degree: d=%d", d)
 	}
 	var rnd func(int) int
 	if src == nil {
 		rnd = rand.Intn
 	} else {
-		rnd = src.Intn
+		rnd = rand.New(src).Intn
 	}
 
 	p := make([]graph.Node, 2*n)

graph/graphs/gen/duplication.go

@@ -29,7 +29,7 @@ type UndirectedMutator interface {
 // created with probability sigma. With the exception of the sigma parameter, this
 // corresponds to the completely correlated case in doi:10.1016/S0022-5193(03)00028-6.
 // If src is not nil it is used as the random source, otherwise rand.Float64 is used.
-func Duplication(dst UndirectedMutator, n int, delta, alpha, sigma float64, src *rand.Rand) error {
+func Duplication(dst UndirectedMutator, n int, delta, alpha, sigma float64, src rand.Source) error {
 	// As described in doi:10.1016/S0022-5193(03)00028-6 but
 	// also clarified in doi:10.1186/gb-2007-8-4-r51.
 
@@ -51,8 +51,9 @@ func Duplication(dst UndirectedMutator, n int, delta, alpha, sigma float64, src
 		rnd = rand.Float64
 		rndN = rand.Intn
 	} else {
-		rnd = src.Float64
-		rndN = src.Intn
+		r := rand.New(src)
+		rnd = r.Float64
+		rndN = r.Intn
 	}
 
 	nodes := dst.Nodes()

graph/graphs/gen/holme_kim.go

@@ -21,10 +21,10 @@ import (
 // additional edges joining existing nodes with probability proportional to the nodes'
 // degrees. The edges are formed as a triad with probability, p.
 // If src is not nil it is used as the random source, otherwise rand.Float64 and
-// rand.Intn are used.
+// rand.Intn are used for the random number generators.
 //
 // The algorithm is essentially as described in http://arxiv.org/abs/cond-mat/0110452.
-func TunableClusteringScaleFree(dst graph.UndirectedBuilder, n, m int, p float64, src *rand.Rand) error {
+func TunableClusteringScaleFree(dst graph.UndirectedBuilder, n, m int, p float64, src rand.Source) error {
 	if p < 0 || p > 1 {
 		return fmt.Errorf("gen: bad probability: p=%v", p)
 	}
@@ -40,8 +40,9 @@ func TunableClusteringScaleFree(dst graph.UndirectedBuilder, n, m int, p float64
 		rnd = rand.Float64
 		rndN = rand.Intn
 	} else {
-		rnd = src.Float64
-		rndN = src.Intn
+		r := rand.New(src)
+		rnd = r.Float64
+		rndN = r.Intn
 	}
 
 	// Initial condition.
@@ -116,11 +117,11 @@ func permute(n []graph.Node, rnd func(int) int) []graph.Node {
 // node having degree m-1. At each iteration of graph addition, one node is added
 // with m additional edges joining existing nodes with probability proportional
 // to the nodes' degrees. If src is not nil it is used as the random source,
-// otherwise rand.Float64 is used.
+// otherwise rand.Float64 is for the random number generator.
 //
 // The algorithm is essentially as described in http://arxiv.org/abs/cond-mat/0110452
 // after 10.1126/science.286.5439.509.
-func PreferentialAttachment(dst graph.UndirectedBuilder, n, m int, src *rand.Rand) error {
+func PreferentialAttachment(dst graph.UndirectedBuilder, n, m int, src rand.Source) error {
 	if n <= m {
 		return fmt.Errorf("gen: n <= m: n=%v m=%d", n, m)
 	}

graph/graphs/gen/small_world.go

@@ -24,7 +24,7 @@ import (
 // Manhattan distance between non-local nodes.
 //
 // The algorithm is essentially as described on p4 of http://www.cs.cornell.edu/home/kleinber/swn.pdf.
-func NavigableSmallWorld(dst GraphBuilder, dims []int, p, q int, r float64, src *rand.Rand) (err error) {
+func NavigableSmallWorld(dst GraphBuilder, dims []int, p, q int, r float64, src rand.Source) (err error) {
 	if p < 1 {
 		return fmt.Errorf("gen: bad local distance: p=%v", p)
 	}

optimize/cmaes.go

@@ -82,7 +82,7 @@ type CmaEsChol struct {
 	ForgetBest bool
 	// Src allows a random number generator to be supplied for generating samples.
 	// If Src is nil the generator in golang.org/x/math/rand is used.
-	Src *rand.Rand
+	Src rand.Source
 
 	// Fixed algorithm parameters.
 	dim                 int

stat/distmat/wishart.go

@@ -27,7 +27,7 @@ import (
 // See https://en.wikipedia.org/wiki/Wishart_distribution for more information.
 type Wishart struct {
 	nu  float64
-	src *rand.Rand
+	src rand.Source
 
 	dim     int
 	cholv   mat.Cholesky
@@ -43,7 +43,7 @@ type Wishart struct {
 // successful.
 //
 // NewWishart panics if nu <= d - 1 where d is the order of v.
-func NewWishart(v mat.Symmetric, nu float64, src *rand.Rand) (*Wishart, bool) {
+func NewWishart(v mat.Symmetric, nu float64, src rand.Source) (*Wishart, bool) {
 	dim := v.Symmetric()
 	if nu <= float64(dim-1) {
 		panic("wishart: nu must be greater than dim-1")

stat/distmv/dirichlet.go

@@ -27,15 +27,15 @@ import (
 type Dirichlet struct {
 	alpha []float64
 	dim   int
-	src   *rand.Rand
+	src   rand.Source
 
 	lbeta    float64
 	sumAlpha float64
 }
 
 // NewDirichlet creates a new dirichlet distribution with the given parameters alpha.
 // NewDirichlet will panic if len(alpha) == 0, or if any alpha is <= 0.
-func NewDirichlet(alpha []float64, src *rand.Rand) *Dirichlet {
+func NewDirichlet(alpha []float64, src rand.Source) *Dirichlet {
 	dim := len(alpha)
 	if dim == 0 {
 		panic(badZeroDimension)