pto-core 0.1.6__tar.gz

pto_core-0.1.6/.clang-format

@@ -0,0 +1,30 @@
+# Shared clang-format defaults for pto-core's five independently-developed
+# modules. This is formatting policy, not code, so it does not conflict with
+# CLAUDE.md's "no shared layer" rule -- each module still fully owns its own
+# style via modules/<name>/.clang-format, which inherits this file and
+# overrides only what that module has always done differently (indent
+# width; see the per-module files for the rest).
+#
+# Deliberately conservative otherwise. This repository's headers carry long,
+# carefully written prose comments (see CLAUDE.md and any docs/AUDIT_*.md /
+# docs/TORTURE_*.md for why) and SIMD/BGZF code where a line break is
+# sometimes chosen for a reason a formatter can't see. ReflowComments and
+# SortIncludes are off so the tool only ever touches indentation, brace
+# placement and pointer/reference alignment -- never comment text or
+# include order.
+#
+# Vendored third-party code (modules/*/third_party/) is excluded entirely
+# via .pre-commit-config.yaml, not here: clang-format has no per-directory
+# disable, only per-directory style, and giving vendored code a style here
+# would imply an opinion about code this project does not own.
+BasedOnStyle: LLVM
+ColumnLimit: 100
+PointerAlignment: Left
+ReferenceAlignment: Left
+NamespaceIndentation: None
+ReflowComments: false
+SortIncludes: false
+AllowShortIfStatementsOnASingleLine: WithoutElse
+AllowShortLoopsOnASingleLine: true
+SpacesBeforeTrailingComments: 2
+AlwaysBreakTemplateDeclarations: Yes

pto_core-0.1.6/.dockerignore

@@ -0,0 +1,48 @@
+# Keep the TSan build context small and reproducible.
+# The image builds from source; it needs no build trees, no binary fixtures and
+# no local caches. modules/cuttag_profiler/data alone is ~100 MB of BAM.
+**/build
+**/build-*
+build-tsan
+.git
+**/.git
+**/__pycache__
+**/*.pyc
+**/.DS_Store
+**/.claude
+
+# Large binary test fixtures. cuttag_profiler's suites synthesise their own BAMs
+# via TempBam, so these are only needed for manual benchmarking.
+# modules/cuttag_profiler/data  (kept: deploy/ seeds demo fixtures from it)
+modules/cuttag_profiler/results
+modules/cuttag_profiler/docs/*.pdf
+
+# Notebook outputs and generated docs.
+**/*.ipynb_checkpoints
+
+# Demo outputs (regenerated by scripts/run_demo.sh).
+demo-out
+
+# Virtualenvs. Never needed in an image (both Dockerfiles install their own
+# dependencies) and easily the largest thing in a developer's tree after the
+# build directories -- pto-core/venv alone is ~165 MB.
+venv
+.venv
+**/venv
+**/.venv
+
+# Terraform provider binaries. `terraform init` leaves ~1.4 GB under
+# pto-cloud/infra/.terraform, and the MVP image builds with the REPOSITORY
+# ROOT as its context (mvp/docker-compose.yml sets `context: ../..`, because
+# it needs pto-core/ and pto-cloud/mvp/ which are siblings). Without this the
+# daemon is handed the whole provider cache on every build.
+#
+# State files are excluded for a different reason: the callback token is
+# generated by `random_password` and lives in state, so a state file must
+# never reach an image layer.
+**/.terraform
+**/*.tfstate
+**/*.tfstate.*
+
+# Test caches.
+**/.pytest_cache

pto_core-0.1.6/.gitignore

@@ -0,0 +1,32 @@
+# Build trees. CMake output is regenerable and host-specific; never commit it.
+build/
+build-*/
+cmake-build-*/
+
+# Demo and benchmark outputs (scripts/run_demo.sh, scripts/benchmark.sh).
+demo-out/
+results/
+
+# Datasets pulled or generated at run time. scanpy writes sc.datasets.* into
+# ./data at the repo root.
+#
+# Anchored with a leading slash on purpose: an unanchored `data/` also matches
+# modules/cuttag_profiler/data/, which holds the checked-in demo.bam that
+# scripts/run_demo.sh and deploy/ both depend on. (That module has its own
+# .gitignore excluding them already -- a separate decision about ~100 MB of
+# binary fixtures, not one to make accidentally from the root.)
+/data/
+*.h5ad
+
+# Python environments and caches.
+venv/
+.venv/
+__pycache__/
+*.py[cod]
+.pytest_cache/
+.ipynb_checkpoints/
+
+# Editor / OS detritus.
+.DS_Store
+*.swp
+compile_commands.json

pto_core-0.1.6/.pre-commit-config.yaml

@@ -0,0 +1,59 @@
+# Repository hygiene, not correctness -- these hooks catch trailing
+# whitespace, missing newlines, and formatting drift. They do not replace
+# `ctest`, the sanitizer builds, or `pytest`; see CONTRIBUTING.md for those.
+#
+# Install once per clone with:
+#   pip install pre-commit && pre-commit install
+#
+# Run against the whole tree (not just staged files) with:
+#   pre-commit run --all-files
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v6.0.0
+    hooks:
+      - id: trailing-whitespace
+        exclude: &third_party '^modules/[^/]+/third_party/'
+      - id: end-of-file-fixer
+        exclude: *third_party
+      - id: check-yaml
+      - id: check-added-large-files
+        # demo.bam (~8.6 MB) is a deliberate, already-reviewed checked-in
+        # fixture -- see modules/cuttag_profiler/.gitignore's own comment.
+        # This threshold is about catching an ACCIDENTAL large add (a stray
+        # build artifact, a BAM someone forgot to .gitignore), not that file.
+        args: ['--maxkb=1024']
+        exclude: '^modules/cuttag_profiler/data/'
+
+  # Python: packaging/pto/ (the pto-core CLI and API), scripts/, tests/validation/.
+  # Notebooks are excluded: ruff lints them by default, and E402 (import not
+  # at the top) is a false positive against normal notebook style, where a
+  # setup cell before the imports cell is the norm, not a hygiene lapse.
+  # `types: [python]` on ruff-format specifically (overriding its default,
+  # which also reaches into fenced ```python blocks in Markdown) because
+  # README.md and packaging/README.md carry Python examples with deliberate,
+  # non-default spacing for visual alignment -- confirmed with
+  # `ruff format --diff packaging/README.md` while writing this file.
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.16.8
+    hooks:
+      - id: ruff
+        args: ['--fix']
+        exclude: &py_exclude '^modules/[^/]+/third_party/|\.ipynb$'
+      - id: ruff-format
+        types: [python]
+        exclude: *py_exclude
+
+  # C++: each module carries its own modules/<name>/.clang-format, inheriting
+  # shared conservative settings from the root .clang-format (see that
+  # file's header comment) and overriding only indent width -- the five
+  # modules were developed independently and were never on one convention.
+  # Vendored code is excluded entirely: it isn't ours to reformat, and
+  # clang-format has no per-directory "leave alone", only per-directory
+  # style, so exclusion belongs here rather than in a third .clang-format
+  # pretending to have an opinion about code this project doesn't own.
+  - repo: https://github.com/pre-commit/mirrors-clang-format
+    rev: v23.1.1
+    hooks:
+      - id: clang-format
+        types_or: [c++, c]
+        exclude: *third_party

pto_core-0.1.6/CITATION.cff

@@ -0,0 +1,30 @@
+cff-version: 1.2.0
+message: >-
+  If you use this software, please cite it using the metadata below.
+title: "pto-core: High-Throughput Omics C++ Engine and Python Interface"
+abstract: >-
+  Five independently-developed C++20 engines for single-cell and epigenomic
+  data, vendored into one repository so they can be built, tested and
+  reviewed together: streaming FASTQ QC and adapter/quality trimming
+  (fastq_stream), streaming fragment QC with duplicate rate and FRiP
+  (genomic_toolkit), CUT&Tag/CUT&RUN reference-point signal profiling
+  (cuttag_profiler), a streaming epigenomic peak caller (peaks), and a
+  cache-aligned sparse-matrix engine with SIMD and HNSW k-NN for single-cell
+  analysis (scrna_matrix). Each tool runs on the machine that already holds
+  the sequencing data, with no service dependency and no data leaving the
+  host. The Python interface is distributed as the `pto-core` package
+  (`scrna-matrix` ships separately due to its OpenMP requirement).
+type: software
+authors:
+  - name: "pto-core contributors"
+repository-code: "https://github.com/chrisaperez/pto-core"
+url: "https://github.com/chrisaperez/pto-core"
+license: MIT
+version: 0.1.5
+date-released: "2026-09-11"
+keywords:
+  - bioinformatics
+  - genomics
+  - cpp
+  - single-cell
+  - hpc