[WIP] Add approximation tensor network contraction subpackage #11
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
LinjianMa
changed the title
LinjianMa
commented
Nov 2, 2022
| end | ||
| end | ||
|
|
||
| # ctree: contraction tree |
There was a problem hiding this comment.
nested vector if itensors, e.g. [[A,B,C,], [D,E]]
LinjianMa
commented
Nov 2, 2022
|
|
||
| # ctree: contraction tree | ||
| # tn: vector of tensors representing a tensor network | ||
| # tn_tree: a dict maps each index tree in the tn to a tensor |
There was a problem hiding this comment.
used in the caching
LinjianMa
commented
Nov 3, 2022
| # contract_ig: the index group to be contracted next | ||
| # ig_tree: an index group with a tree hierarchy | ||
| function approximate_contract(ctree::Vector; kwargs...) | ||
| tn_leaves = get_leaves(ctree) |
There was a problem hiding this comment.
tn_leaves are all the groups
LinjianMa
commented
Nov 3, 2022
| # ig_tree: an index group with a tree hierarchy | ||
| function approximate_contract(ctree::Vector; kwargs...) | ||
| tn_leaves = get_leaves(ctree) | ||
| ctrees = topo_sort(ctree; leaves=tn_leaves) |
There was a problem hiding this comment.
ctrees: A list of contraction tree, it defines the contraction path
LinjianMa
commented
Nov 3, 2022
| ) | ||
| end | ||
| for c in ctrees | ||
| if ctree_to_igs[c] == [] |
There was a problem hiding this comment.
772-802 are to get caching information
LinjianMa
changed the title
…ntract, remove project_boundary
mtfishman
reviewed
Dec 23, 2022
mtfishman
reviewed
Dec 23, 2022
Comment on lines
1
to
100
|
|
||
| # merge two trees | ||
| # new tree: | ||
| # s | ||
| # / \ | ||
| # t1 t2 | ||
| function merge_tree(t1::Vector, t2::Vector; append=false) | ||
| if t2 == [] | ||
| return t1 | ||
| end | ||
| if t1 == [] | ||
| return t2 | ||
| end | ||
| if isleaf(t1) && isleaf(t2) | ||
| return [t1, t2] | ||
| end | ||
| if isleaf(t1) | ||
| return append ? [t1, t2...] : [t1, t2] | ||
| end | ||
| if isleaf(t2) | ||
| return append ? [t1..., t2] : [t1, t2] | ||
| end | ||
| return append ? [t1..., t2...] : [t1, t2] | ||
| end | ||
|
|
||
| function isleaf(tree::Vector) | ||
| if tree == [] | ||
| @info "tree is empty" | ||
| return false | ||
| end | ||
| if all(v -> !(v isa Vector), tree) | ||
| return true | ||
| end | ||
| return false | ||
| end | ||
|
|
||
| # get the subtree of tree that is in the subset | ||
| # example: | ||
| # subtree([[1, 2], [3, 4]], [1, 3]) = ([[1], [3]]) | ||
| function subtree(tree::Vector, subset::Union{Vector,Tuple}) | ||
| if tree == [] | ||
| return [] | ||
| end | ||
| if isleaf(tree) | ||
| return intersect(tree, subset) | ||
| end | ||
| tree = [subtree(i, subset) for i in tree] | ||
| tree = filter(t -> t != [], tree) | ||
| if length(tree) == 1 && tree[1] isa Vector | ||
| return tree[1] | ||
| end | ||
| return tree | ||
| end | ||
|
|
||
| # vectorize a tree | ||
| # example: [[1,2], [3,4]] = [1, 2, 3, 4] | ||
| function vectorize(tree) | ||
| @assert tree != [] | ||
| if !(tree isa Vector) | ||
| return [tree] | ||
| end | ||
| return mapreduce(vectorize, vcat, tree) | ||
| end | ||
|
|
||
| # example: [[[1,2], [3,4]], [[5,6], [7,8]]] = [[1,2], [3,4], [5,6], [7,8]] | ||
| function get_leaves(tree::Vector) | ||
| if !(tree isa Vector{<:Vector}) | ||
| return [tree] | ||
| end | ||
| return mapreduce(get_leaves, vcat, tree) | ||
| end | ||
|
|
||
| function line_to_tree(line::Vector) | ||
| if length(line) == 1 && line[1] isa Vector | ||
| return line[1] | ||
| end | ||
| if length(line) <= 2 | ||
| return line | ||
| end | ||
| return [line_to_tree(line[1:(end - 1)]), line[end]] | ||
| end | ||
|
|
||
| function topo_sort(tn; type=Vector, leaves=[]) | ||
| @timeit timer "topo_sort" begin | ||
| topo_order = [] | ||
| topo_sort_dfs!(tn, topo_order, leaves, type) | ||
| return topo_order | ||
| end | ||
| end | ||
|
|
||
| function topo_sort_dfs!(tn, topo_order, leaves, type) | ||
| #Post-order DFS | ||
| if (tn in leaves) || !(tn isa type) | ||
| return nothing | ||
| end | ||
| for subtn in tn | ||
| topo_sort_dfs!(subtn, topo_order, leaves, type) | ||
| end | ||
| return append!(topo_order, [tn]) | ||
| end |
There was a problem hiding this comment.
Can any of this make use of functionality from AbstractTrees.jl? It would be nice to use existing packages if possible so we don't have to maintain our own versions of existing functions.
There was a problem hiding this comment.
Thanks for the suggestion, I will keep this in mind when refactoring the code
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
mtfishman
reviewed
Jan 6, 2023
| end | ||
|
|
||
| # Note that the children ordering matters here. | ||
| mutable struct IndexAdjacencyTree |
There was a problem hiding this comment.
@LinjianMa could you explain what this type is for?
|
Will just close this one since updates are merged gradually in other PRs. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Still need to get some more tests working, but most of the parts are done.
Package transferred from https://github.com/ITensor/ITensorNetworkAD.jl