Release seqproc 0.1.0

.cargo/config-baseline.toml

@@ -0,0 +1,2 @@
+# Architecture-neutral control used only by comprehensive compatibility and
+# byte-equivalence CI. It intentionally supplies no target-specific rustflags.

.cargo/config-release.toml

@@ -0,0 +1,24 @@
+# Fixed release targets: x86_64 artifacts require the psABI x86-64-v3 level;
+# aarch64 artifacts use one documented target per operating system. Local
+# development uses target-cpu=native from config.toml instead.
+
+[env]
+# Cargo does not pass target-specific rustflags to build scripts. This marker
+# lets build.rs record the fixed artifact CPU target and its expanded feature
+# set in `seqproc --version --verbose`.
+SEQPROC_RELEASE_BUILD = "1"
+
+[target.x86_64-unknown-linux-gnu]
+# Load-bearing: GNU ld preserves/merges the explicit x86 ISA-needed note in
+# the seqproc binary. rust-lld on the reviewed toolchain silently drops it.
+linker = "gcc"
+rustflags = ["-C", "target-cpu=x86-64-v3"]
+
+[target.x86_64-apple-darwin]
+rustflags = ["-C", "target-cpu=x86-64-v3"]
+
+[target.aarch64-unknown-linux-gnu]
+rustflags = ["-C", "target-cpu=neoverse-n1"]
+
+[target.aarch64-apple-darwin]
+rustflags = ["-C", "target-cpu=apple-a14"]

.cargo/config.local.example.toml

@@ -2,6 +2,14 @@
 # tuned to the local CPU:
 #
 #     cargo build --release --config .cargo/config.local.toml
+#
+# seqproc selects ANTISEQUENCE's architecture-tuned backend by default. To use
+# the generic compatibility control instead, override rustflags and features:
+#
+#     RUSTFLAGS="" cargo build --release --no-default-features \
+#       --features antisequence/baseline-simd
 
+# `config.toml` already uses this setting. Keep this example for packagers or
+# users who override the repository config and want an explicit local build.
 [build]
 rustflags = ["-C", "target-cpu=native"]

.cargo/config.toml

@@ -1,16 +1,6 @@
-# Portable release defaults. Never use target-cpu=native here: Cargo source
-# builds and downstream packagers honor this file, and native tuning can create
-# binaries that fail on older hosts. The release workflow applies the same
-# target floors through .github/build-setup.yml.
+# Local development follows the host CPU so profiling and ordinary source
+# builds exercise its best code generation. Release automation replaces this
+# file with config-release.toml, whose target floors are fixed and reproducible.
 
-[target.x86_64-unknown-linux-gnu]
-rustflags = ["-C", "target-cpu=x86-64-v3", "-C", "target-feature=+avx2"]
-
-[target.aarch64-unknown-linux-gnu]
-rustflags = ["-C", "target-cpu=neoverse-n1"]
-
-[target.aarch64-apple-darwin]
-rustflags = ["-C", "target-cpu=apple-a14"]
-
-[target.x86_64-apple-darwin]
-rustflags = ["-C", "target-cpu=x86-64-v3", "-C", "target-feature=+avx2"]
+[build]
+rustflags = ["-C", "target-cpu=native"]

.github/build-setup.yml

@@ -1,22 +1,8 @@
-- name: Use portable Cargo config for CI builds
+- name: Use fixed architecture-tuned Cargo config for release builds
   shell: bash
   run: |
     set -euo pipefail
-    cp .cargo/config-portable.toml .cargo/config.toml
-
-- name: Force AVX2 for x86_64 release artifacts
-  if: contains(join(matrix.targets, ','), 'x86_64-')
-  shell: bash
-  run: |
-    set -euo pipefail
-    echo 'RUSTFLAGS=-C target-cpu=x86-64-v3 -C target-feature=+avx2' >> "$GITHUB_ENV"
-
-- name: Show effective x86_64 rustflags
-  if: contains(join(matrix.targets, ','), 'x86_64-')
-  shell: bash
-  run: |
-    set -euo pipefail
-    echo "RUSTFLAGS=${RUSTFLAGS-}"
+    cp .cargo/config-release.toml .cargo/config.toml
 
 - name: Cache Rust release build
   uses: Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6 # v2.9.2

.github/workflows/actions.yml

@@ -3,9 +3,9 @@ name: Fast CI
 on:
   workflow_dispatch:
   push:
-    branches: ["main"]
+    branches: ["main", "dev"]
   pull_request:
-    branches: ["main"]
+    branches: ["main", "dev"]
 
 concurrency:
   group: fast-ci-${{ github.ref }}
@@ -22,7 +22,7 @@ jobs:
       CARGO_NET_RETRY: 10
       CARGO_TERM_COLOR: always
       RUST_BACKTRACE: 1
-      RUSTFLAGS: -D warnings
+      RUSTFLAGS: -C target-cpu=x86-64-v3 -D warnings
       RUSTDOCFLAGS: -D warnings
       RUSTUP_MAX_RETRIES: 10
 
@@ -48,6 +48,15 @@ jobs:
         run: >-
           cargo test --locked --workspace --lib --bins
           --test cli_workflow_tests
+          --test compile_tests
           --test error_handling_tests
+          --test error_contract_tests
+          --test anchor_set_tests
+          --test layout_algebra_tests
+          --test lexer_tests
           --test paper_chemistry_tests
+          --test diff_tests
           --no-fail-fast
+
+      - name: clippy (library, binaries, and tests)
+        run: cargo clippy --locked --workspace --lib --bins --tests

.github/workflows/comprehensive.yml

@@ -19,7 +19,7 @@ jobs:
       CARGO_NET_RETRY: 10
       CARGO_TERM_COLOR: always
       RUST_BACKTRACE: 1
-      RUSTFLAGS: -D warnings
+      RUSTFLAGS: -C target-cpu=x86-64-v3 -D warnings
       RUSTDOCFLAGS: -D warnings
       RUSTUP_MAX_RETRIES: 10
 
@@ -49,3 +49,54 @@ jobs:
 
       - name: doc
         run: cargo doc --locked --workspace --no-deps
+
+      - name: Inspect source package contents
+        # Full verification resolves the registry form of ANTISEQUENCE and is
+        # performed by bump_and_publish.sh after that crate is published.
+        run: cargo package --locked --list > /dev/null
+
+      - name: Verify x86-64-v3 release contract
+        run: |
+          cargo build --release --locked --bin seqproc
+          target/release/seqproc --version --verbose
+          target/release/seqproc --version --verbose | grep -F "compiler CPU target: x86-64-v3"
+          target/release/seqproc --version --verbose | grep -F "SIMD backend: x86-avx2"
+
+  baseline:
+    runs-on: ubuntu-latest
+    env:
+      RUSTFLAGS: -D warnings
+    steps:
+      - uses: actions/checkout@v6
+      - uses: dtolnay/rust-toolchain@stable
+      - uses: Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6 # v2.9.2
+        with:
+          shared-key: seqproc-comprehensive-baseline
+      - name: Test generic SSE2 compatibility backend
+        run: >-
+          cargo test --locked --lib --bins --no-default-features
+          --features antisequence/baseline-simd
+      - name: Compare baseline and release-SIMD FASTQ bytes
+        run: bash scripts/verify_simd_equivalence.sh
+
+  msrv:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+      - uses: dtolnay/rust-toolchain@master
+        with:
+          toolchain: "1.88"
+      - uses: Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6 # v2.9.2
+        with:
+          shared-key: seqproc-msrv-1.88
+      - run: cargo check --locked --all-targets
+
+  macos-arm64:
+    runs-on: macos-15
+    steps:
+      - uses: actions/checkout@v6
+      - uses: dtolnay/rust-toolchain@stable
+      - uses: Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6 # v2.9.2
+        with:
+          shared-key: seqproc-macos-arm64
+      - run: cargo test --locked --lib --bins

.github/workflows/release.yml

@@ -127,22 +127,10 @@ jobs:
             curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
             echo "$HOME/.cargo/bin" >> $GITHUB_PATH
           fi
-      - name: "Use portable Cargo config for CI builds"
+      - name: "Use fixed architecture-tuned Cargo config for release builds"
         run: |
           set -euo pipefail
-          cp .cargo/config-portable.toml .cargo/config.toml
-        shell: "bash"
-      - name: "Force AVX2 for x86_64 release artifacts"
-        if: "contains(join(matrix.targets, ','), 'x86_64-')"
-        run: |
-          set -euo pipefail
-          echo 'RUSTFLAGS=-C target-cpu=x86-64-v3 -C target-feature=+avx2' >> "$GITHUB_ENV"
-        shell: "bash"
-      - name: "Show effective x86_64 rustflags"
-        if: "contains(join(matrix.targets, ','), 'x86_64-')"
-        run: |
-          set -euo pipefail
-          echo "RUSTFLAGS=${RUSTFLAGS-}"
+          cp .cargo/config-release.toml .cargo/config.toml
         shell: "bash"
       - name: "Cache Rust release build"
         uses: "Swatinem/rust-cache@6323deb102c322ba6fcbdcafc7e3dddab59af2b6"
@@ -232,14 +220,24 @@ jobs:
           path: |
             ${{ steps.cargo-dist.outputs.paths }}
             ${{ env.BUILD_MANIFEST_NAME }}
+
+  custom-verify-release-provenance:
+    needs:
+      - plan
+      - build-local-artifacts
+    uses: ./.github/workflows/verify-release-provenance.yml
+    with:
+      plan: ${{ needs.plan.outputs.val }}
+    secrets: inherit
   # Determines if we should publish/announce
   host:
     needs:
       - plan
       - build-local-artifacts
       - build-global-artifacts
+      - custom-verify-release-provenance
     # Only run if we're "publishing", and only if plan, local and global didn't fail (skipped is fine)
-    if: ${{ always() && needs.plan.result == 'success' && needs.plan.outputs.publishing == 'true' && (needs.build-global-artifacts.result == 'skipped' || needs.build-global-artifacts.result == 'success') && (needs.build-local-artifacts.result == 'skipped' || needs.build-local-artifacts.result == 'success') }}
+    if: ${{ always() && needs.plan.result == 'success' && needs.plan.outputs.publishing == 'true' && (needs.build-global-artifacts.result == 'skipped' || needs.build-global-artifacts.result == 'success') && (needs.custom-verify-release-provenance.result == 'skipped' || needs.custom-verify-release-provenance.result == 'success') && (needs.build-local-artifacts.result == 'skipped' || needs.build-local-artifacts.result == 'success') }}
     env:
       GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
     runs-on: "ubuntu-22.04"

.github/workflows/verify-release-provenance.yml

@@ -0,0 +1,39 @@
+name: Verify release provenance
+
+on:
+  workflow_call:
+    inputs:
+      plan:
+        required: true
+        type: string
+
+permissions:
+  contents: read
+
+jobs:
+  verify-x86_64-linux:
+    runs-on: ubuntu-22.04
+    steps:
+      - uses: actions/checkout@v6
+      - name: Download cargo-dist Linux artifact
+        uses: actions/download-artifact@v8
+        with:
+          name: artifacts-build-local-x86_64-unknown-linux-gnu
+          path: artifacts
+      - name: Verify packaged binary CPU floor and SIMD provenance
+        shell: bash
+        run: |
+          set -euo pipefail
+          archive="$(find artifacts -type f \( -name '*x86_64-unknown-linux-gnu.tar.xz' -o -name '*x86_64-unknown-linux-gnu.tar.gz' \) -print -quit)"
+          test -n "$archive"
+          mkdir unpacked
+          tar -xf "$archive" -C unpacked
+          binary="$(find unpacked -type f -name seqproc -perm -111 -print -quit)"
+          test -n "$binary"
+
+          provenance="$($binary --version --verbose)"
+          printf '%s\n' "$provenance"
+          grep -F 'compiler CPU target: x86-64-v3' <<<"$provenance"
+          grep -F 'CPU floor: x86-64-v3 (AVX2)' <<<"$provenance"
+          grep -F 'SIMD backend: x86-avx2' <<<"$provenance"
+          python3 scripts/verify_x86_64_v3_elf.py "$binary"

CHANGELOG.md

@@ -0,0 +1,61 @@
+# Changelog
+
+All notable changes to seqproc are documented here. This project follows
+[Semantic Versioning](https://semver.org/).
+
+## [Unreleased]
+
+## [0.1.0] - 2026-08-22
+
+Initial public release. It includes the preprint functionality plus the
+post-preprint correctness, usability, and performance work reviewed for this
+release.
+
+- Make `seqproc run` require an exact, nonempty primary-output topology while
+  preserving the deprecated flag-only prefix behavior for one compatibility
+  cycle. Unassigned outputs likewise require exactly one target per input lane,
+  and primary outputs combined with demultiplexing are rejected rather than
+  silently ignored.
+- Treat only a closed stdout pipe as normal Unix early-consumer termination
+  (silent exit 0). ENOSPC, quota exhaustion, named-pipe/file failures, and all
+  other output errors remain nonzero. Output writers retain their target
+  identity through the type-erased execution graph so a file `EPIPE` cannot be
+  mistaken for stdout closure.
+- Report demultiplexed output topology from the effective writer graph: one
+  `path` sink per emitted read lane, without exposing data-dependent sample
+  filenames.
+- Propagate ANTISEQUENCE's single-use graph lifecycle and fallible finalization:
+  malformed input, output flush/footer failures, and repeated execution now
+  produce typed nonzero errors instead of panic, silent truncation, or false
+  success.
+- Fix mixed-length Hamming seed planning so a fast-path seed is used only when
+  it is guaranteed for every literal; this removes a silent false-negative
+  case while retaining safe indexed matching.
+- Make seqproc's executable default architecture-tuned while preserving
+  ANTISEQUENCE's library-safe baseline: distributed x86_64 binaries target
+  x86-64-v3/AVX2, aarch64 artifacts use fixed platform targets, and local
+  repository builds use `target-cpu=native`.
+- Add exact target-feature CPUID/XGETBV compatibility checks, Linux ELF loader
+  ISA notes, `--version --verbose` build provenance, and build/SIMD provenance
+  in summary schema 1.13.0. Release CI inspects the packaged cargo-dist binary
+  before hosting it.
+- Add comprehensive byte-equivalence gates between generic SSE2 and tuned
+  AVX2 builds.
+
+- EFGDL 2 document headers, named resources, fixed output sequences, output
+  header templates, indexed captures, layout choice/optional/repeat algebra,
+  anchor sets, and explicit ambiguity/position policies.
+- One to three separate input lanes, ordered shard lists, stdin/stdout,
+  interleaved FASTQ, gzip auto-detection, and optional accelerated gzip input.
+- Typed `RunConfig`, `RunReport`, and `SeqprocError` APIs plus `run`, `validate`,
+  and `explain` CLI subcommands; legacy flag-only invocation remains available
+  for one compatibility cycle.
+- Corrected SPLiT-seq FILTER/Hamming behavior, normalized duplicate whitelist
+  entries, deterministic ambiguity handling, and exhaustive input validation.
+- Worker-local statistics, versioned run summaries, proof-gated graph
+  optimization, dynamic batch planning, and optimized low-thread execution.
+- Reproducible multi-platform binary release automation with explicit CPU
+  floors.
+
+[Unreleased]: https://github.com/COMBINE-lab/seqproc/compare/v0.1.0...HEAD
+[0.1.0]: https://github.com/COMBINE-lab/seqproc/releases/tag/v0.1.0

CITATION.cff

@@ -0,0 +1,32 @@
+cff-version: 1.2.0
+message: "If you use seqproc, please cite this software and the associated preprint."
+title: "seqproc"
+type: software
+version: 0.1.0
+date-released: 2026-08-22
+license: BSD-3-Clause
+repository-code: "https://github.com/COMBINE-lab/seqproc"
+url: "https://combine-lab.github.io/seqproc"
+authors:
+  - family-names: Cape
+    given-names: Noah
+  - family-names: Fisher
+    given-names: Elan
+  - family-names: Liu
+    given-names: Daniel
+  - family-names: Patro
+    given-names: Rob
+preferred-citation:
+  type: article
+  title: "seqproc: An efficient, flexible, and concise tool for sequence geometry description and transformation"
+  authors:
+    - family-names: Cape
+      given-names: Noah
+    - family-names: Fisher
+      given-names: Elan
+    - family-names: Liu
+      given-names: Daniel
+    - family-names: Patro
+      given-names: Rob
+  doi: "10.64898/2026.07.28.741211"
+  year: 2026