hypersteeldb 0.5.3

A database that compiles questions instead of guessing answers: typed vocabulary discovered from your documents, queries type-checked before they run, roaring-bitmap set algebra over reified hyperedges, and Dempster-Shafer evidence with an explicit conflict guard.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14

# SteelDB ships its models in-repo (works offline out of the box). They're large binaries — track them
# with git-lfs so the repo stays clone-able. Run `git lfs install` once before cloning/committing.
models/** filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.onnx.data filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
*.f32 filter=lfs diff=lfs merge=lfs -text
*.gguf filter=lfs diff=lfs merge=lfs -text
# keep small text configs as normal text (not LFS) for readable diffs
models/**/*.json -filter -diff -merge text
models/**/MODELS.md -filter -diff -merge text
# ...but the big tokenizer vocab JSON is a binary-ish blob — keep it in LFS (rule after wins)
models/**/tokenizer.json filter=lfs diff=lfs merge=lfs -text
models/**/tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text