Download data/lean/OAI/Algorithms/SubexponentialSAT/Subexp/AssignmentLabelSerialization.lean from introvoyz045/math: direct link, hf CLI and curl.
- Browser
- Download file 4.98 kB
-
https://huggingface.co/datasets/introvoyz045/math/resolve/main/data/lean/OAI/Algorithms/SubexponentialSAT/Subexp/AssignmentLabelSerialization.lean
- Command line
-
hf download hf://datasets/introvoyz045/math/data/lean/OAI/Algorithms/SubexponentialSAT/Subexp/AssignmentLabelSerialization.lean
-
curl -L -o AssignmentLabelSerialization.lean https://huggingface.co/datasets/introvoyz045/math/resolve/main/data/lean/OAI/Algorithms/SubexponentialSAT/Subexp/AssignmentLabelSerialization.lean
4.98 kB
| import OAI.Algorithms.SubexponentialSAT.LeanCompat | |
| import OAI.Algorithms.SubexponentialSAT.Subexp.AssignmentSerialization | |
| import OAI.Algorithms.SubexponentialSAT.Subexp.PackingVertexLabels | |
| namespace OAI.Algorithms | |
| /-! The vertex-label producer emits exactly the semantic bucket records, | |
| in ascending vertex order before its explicit accumulator reversal. -/ | |
| namespace SubexponentialSAT.Subexp.AssignmentLabelSerialization | |
| open SubexponentialSAT.MulticolorTypes SubexponentialSAT.SubexpTypeVariance SubexponentialSAT.SlotAssignment | |
| theorem labelRecords_finRange {D : ℕ} (rows : List (List Bool)) | |
| (plans : Fin D → List ℕ) : | |
| PackingVertexLabels.labelRecords rows ((List.finRange D).map plans) 0 = | |
| (List.finRange D).map (fun v => | |
| (GatherProgram.gathered rows (plans v), List.replicate v.val true)) := by | |
| unfold PackingVertexLabels.labelRecords | |
| rw [List.length_map, List.length_finRange, ← List.range_eq_range', | |
| ← List.map_coe_finRange D, List.zipWith_map, List.zipWith_self] | |
| variable {D E r Q Δ : ℕ} | |
| variable (P : SubexponentialSAT.Factor.Pattern (Fin D) (Fin E)) (slots : SlotMap P Δ) | |
| /-- Gathering every actual seed array at the supplied typed read plans | |
| produces exactly the type-key and unary-vertex records used by the bucket. -/ | |
| theorem labelRecords_eq_vertexRecords (ω : Assignment (Fin D) (Fin E) r Q Δ) : | |
| PackingVertexLabels.labelRecords | |
| (ReadPlan.seedRows (List.finRange D) (List.finRange E) ω) | |
| ((List.finRange D).map | |
| (ReadPlan.vertexPlan P slots (List.finRange D) (List.finRange E))) 0 = | |
| BucketProgram.vertexRecords (List.finRange D) (assignmentType P slots ω) | |
| CollisionProgram.encodeType (fun v => List.replicate v.val true) := by | |
| rw [labelRecords_finRange] | |
| unfold BucketProgram.vertexRecords | |
| apply List.map_congr_left | |
| intro v _ | |
| apply Prod.ext | |
| · exact (PackingVertexLabels.gathered_eq_readPlan _ _).trans | |
| (ReadPlan.gather_seedRows P slots (List.finRange D) (List.finRange E) | |
| (fun v => List.mem_finRange v) (fun e => List.mem_finRange e) ω v) | |
| · rfl | |
| /-- The physical accumulator stores the same records in reverse list order. -/ | |
| theorem labelBytes_eq_vertexRecords (ω : Assignment (Fin D) (Fin E) r Q Δ) : | |
| BucketProgram.encodeRecords | |
| ((PackingVertexLabels.labelRecords | |
| (ReadPlan.seedRows (List.finRange D) (List.finRange E) ω) | |
| ((List.finRange D).map | |
| (ReadPlan.vertexPlan P slots (List.finRange D) (List.finRange E))) 0).reverse) = | |
| BucketProgram.encodeRecords | |
| ((BucketProgram.vertexRecords (List.finRange D) (assignmentType P slots ω) | |
| CollisionProgram.encodeType (fun v => List.replicate v.val true)).reverse) := | |
| congrArg (fun records => BucketProgram.encodeRecords records.reverse) | |
| (labelRecords_eq_vertexRecords P slots ω) | |
| variable [NeZero r] | |
| omit [NeZero r] in | |
| /-- The producer can consume the rows already emitted by the natural | |
| assignment loop, without reserializing or reordering that assignment. -/ | |
| theorem natural_labelRecords_eq_vertexRecords (a : Word r Q) (tail : List (Word r Q)) : | |
| let indices := AssignmentEnumeration.seedIndices (List.finRange D) (List.finRange E) Δ | |
| let ω := NaturalAssignment.naturalAssignment P slots a tail (List.finRange D) indices | |
| PackingVertexLabels.labelRecords | |
| ((NaturalAssignment.naturalValues P slots a tail (List.finRange D) indices).reverse.map | |
| ReadPlan.encodeWord) | |
| ((List.finRange D).map | |
| (ReadPlan.vertexPlan P slots (List.finRange D) (List.finRange E))) 0 = | |
| BucketProgram.vertexRecords (List.finRange D) (assignmentType P slots ω) | |
| CollisionProgram.encodeType (fun v => List.replicate v.val true) := by | |
| dsimp only | |
| rw [AssignmentSerialization.naturalValues_reverse_map_eq_seedRows P slots a tail | |
| (List.finRange D) (List.finRange E) (List.nodup_finRange D) (List.nodup_finRange E)] | |
| exact labelRecords_eq_vertexRecords P slots _ | |
| omit [NeZero r] in | |
| theorem natural_labelBytes_eq_vertexRecords (a : Word r Q) (tail : List (Word r Q)) : | |
| let indices := AssignmentEnumeration.seedIndices (List.finRange D) (List.finRange E) Δ | |
| let ω := NaturalAssignment.naturalAssignment P slots a tail (List.finRange D) indices | |
| BucketProgram.encodeRecords | |
| ((PackingVertexLabels.labelRecords | |
| ((NaturalAssignment.naturalValues P slots a tail (List.finRange D) indices).reverse.map | |
| ReadPlan.encodeWord) | |
| ((List.finRange D).map | |
| (ReadPlan.vertexPlan P slots (List.finRange D) (List.finRange E))) 0).reverse) = | |
| BucketProgram.encodeRecords | |
| ((BucketProgram.vertexRecords (List.finRange D) (assignmentType P slots ω) | |
| CollisionProgram.encodeType (fun v => List.replicate v.val true)).reverse) := by | |
| exact congrArg (fun records => BucketProgram.encodeRecords records.reverse) | |
| (natural_labelRecords_eq_vertexRecords P slots a tail) | |
| end SubexponentialSAT.Subexp.AssignmentLabelSerialization | |
| end OAI.Algorithms | |