feat: GGUF v3 parser, IQ wire layouts, and T1 large MoE pilots (#7) - #44

Merged

rmems merged 32 commits into

mainfrom

feat/gguf-parser-0.2.0

Aug 7, 2026

Commits on Jul 24, 2026

feat: enhance GGUF parser with full GGML type support and metadata helpers
rmems
committed
07f5940
fix: treat GGUF wire type 31 as Q4_0_4_4, not IQ3_M
rmems
committed
39c7d29
docs: charter + #7 GGUF extraction source and wire-type notes
rmems
committed
f2a5e55
style: rustfmt cursor/layout after #7 work
rmems
committed
6678b42
feat: complete GGUF IQ wire layouts and file_type quant fallback
rmems
committed
5cc0a2d

Commits on Aug 2, 2026

test: harden real_gguf MoE pilots with EXPECT_MOE and samples
rmems
committed
4823f97
test: gate real_gguf default env asserts on unset vars
rmems
committed
e74ab09
docs: document large MoE T1 RAM budget and pilot env vars
rmems
committed
9d666e9
chore: release 0.2.0 prep — MSRV 1.97.1, inspect_gguf, docs
rmems
committed
2890444
chore: sync Cargo.lock package version to 0.2.0
rmems
committed
c59c306
docs: clarify GGUF wire types are metadata only (no GGML compute)
rmems
committed
6403b32
fix: address PR #44 review — Result export, MSRV pin, quant layout
rmems
committed
7120d77

Commits on Aug 3, 2026

fix: tighten review follow-ups — layout-only negative check, restore DType helpers
rmems
committed
d776365

Commits on Aug 7, 2026

refactor: split complex methods/tests flagged by CodeScene/Codacy
rmems
committed
09d6e33
fix(parser): reject malicious metadata and negative signed counts

devin-ai-integration[bot]
and
rmems
committed
957d172
docs(tensor): clarify has_known_byte_layout and fix intra-doc link

devin-ai-integration[bot]
and
rmems
committed
62a6234
test(examples): cover f32 metadata and validate quantized expert bytes

devin-ai-integration[bot]
and
rmems
committed
8766e5c
chore(ci): pin toolchain in Docker and tighten MSRV version check

devin-ai-integration[bot]
and
rmems
committed
91eafe0
test(common): align synthetic tensor payloads to 32 bytes

devin-ai-integration[bot]
and
rmems
committed
ca78eb1
chore(tests): exercise env helpers and improve example UX

devin-ai-integration[bot]
and
rmems
committed
5fe851c
fix(layout): preserve numeric signedness and skip arrays iteratively

devin-ai-integration[bot]
and
rmems
committed
9a48852
test(real_gguf): exercise env helpers without global mutation

devin-ai-integration[bot]
and
rmems
committed
99fea0d
fix(inspect_gguf): handle --help/--version/-- and scope Dockerfile toolchain

devin-ai-integration[bot]
and
rmems
committed
f82d6b2
fix(layout): consult signed_numerics in quantization() and add test

devin-ai-integration[bot]
and
rmems
committed
661f141
ci: retrigger Kilo Code Review after rate limit

devin-ai-integration[bot]
and
rmems
committed
82870cb
test(cursor): verify iterative skip of deeply nested GGUF arrays

devin-ai-integration[bot]
and
rmems
committed
ade7893
style: rustfmt float assertion; pin Dockerfile CMD to RUST_VERSION toolchain

devin-ai-integration[bot]
and
rmems
committed
a5fd1c3
refactor(real_gguf): bound directory scan, default max to 1, split scan_one_pilot

devin-ai-integration[bot]
and
rmems
committed
82a563e
fixup(docker): exec-form CMD + ignore rust-toolchain.toml; sync REVIEW.md default cap

devin-ai-integration[bot]
and
rmems
committed
500c911
fix(codacy): drop cache shared-key, split inspect main, refactor read_numeric, fix md style

devin-ai-integration[bot]
and
rmems
committed
b184207
ci: remove shared-key from rust-cache to avoid gitleaks false positive

devin-ai-integration[bot]
and
rmems
committed
d0a1cfb
refactor(tests): reduce assertion blocks and cyclomatic complexity flagged by CodeScene

devin-ai-integration[bot]
and
rmems
committed
ac8aca8