Combining compression and deduplication?

[gcflymoto]

Hello. Dobry Wieczor. I am researching how to implement a series of steps to deduplicate files by their compressed versions while matching on their original content hash. There are two big goals

1. Skip any necessity to decompress the files in treeA first to run fclones to match them to treeB
2. Skip the computationally intensive steps of compression in treeB and directly deduplicate it with t

Here is a visual example:

treeA with HashCacheA.txt	Hash	Compress?	Deduplicate?
dir1/file1.txt » Size N	Calculate+Store Hash X in hash cache	Compress » file1.txt.zst	No
dir2/file2.txt » Size M	Calculate+Store Hash Y in hash cache	Compress » file2.txt.zst	No
dir3/file3.txt » Size P	Calculate+Store Hash Z in hash cache	Compress » file3.txt.zst	No
dir4/file4.txt » Size N	Calculate+Store Hash X in hash cache	Hash matches treeA/dir1/file1.txt » Skip Compression	Deduplicate with a hard-link to treeA/dir1/file1.txt.zst directly

Later, a new treeB is processed with some files matching and some not.

treeB with HashCacheB.txt	Hash	Compress?	Deduplicate?
dir1/file1.txt » Size N	Calculate & Store Hash X in HashCacheB.txt	Hash matches treeA/dir1/file1.txt » Skip Compression	Deduplicate with a hard-link to treeA/dir1/file1.txt.zst directly
dir2/file2.txt » Size MM	Calculate & Store Hash YY in HashCacheB.txt	Size does not match Compress » file2.txt.zst	No
dir3/file3.txt » Size P	Calculate & Store Hash ZZ in HashCacheB.txt	Hash does not match Compress » file3.txt.zst	No

Later, a new treeC is processed with some files matching in either A, B, or both. Combine the data from HashA and HashB during skip compression and deduplication steps.

Is there any way to use fclones for this (or any other deduplication tools?) to develop this custom system?