pen-stack 3.1.0__py3-none-any.whl

pen_stack-3.1.0.dist-info/METADATA

@@ -0,0 +1,451 @@
+Metadata-Version: 2.4
+Name: pen-stack
+Version: 3.1.0
+Summary: Open infrastructure for genome writing: the Writable Genome atlas, the Writer Atlas, and the Write Planner.
+Author-email: Anees Ahmed Mahaboob Ali <ahmedaneesm@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/ahmedanees-m/pen-stack
+Project-URL: Repository, https://github.com/ahmedanees-m/pen-stack
+Project-URL: Documentation, https://github.com/ahmedanees-m/pen-stack/tree/main/docs
+Project-URL: Changelog, https://github.com/ahmedanees-m/pen-stack/blob/main/CHANGELOG.md
+Project-URL: Issues, https://github.com/ahmedanees-m/pen-stack/issues
+Project-URL: Benchmark, https://github.com/ahmedanees-m/pen-stack/tree/main/benchmarks/genome_writing_bench
+Keywords: genome-writing,genome-editing,bridge-recombinase,safe-harbor,writable-genome,writer-atlas,write-planner,bioinformatics
+Classifier: Development Status :: 4 - Beta
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.26
+Requires-Dist: pandas>=2.2
+Requires-Dist: pyarrow>=16
+Requires-Dist: scipy>=1.13
+Requires-Dist: scikit-learn>=1.5
+Requires-Dist: lightgbm>=4.5
+Requires-Dist: pydantic>=2.8
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: click>=8.1
+Requires-Dist: requests>=2.32
+Requires-Dist: tqdm>=4.66
+Provides-Extra: models
+Requires-Dist: xgboost>=2.1; extra == "models"
+Requires-Dist: torch>=2.3; extra == "models"
+Requires-Dist: statsmodels>=0.14; extra == "models"
+Provides-Extra: bio
+Requires-Dist: biopython>=1.84; extra == "bio"
+Requires-Dist: pysam>=0.22; extra == "bio"
+Requires-Dist: pyBigWig>=0.3; extra == "bio"
+Requires-Dist: pybedtools>=0.10; extra == "bio"
+Requires-Dist: ViennaRNA>=2.6; extra == "bio"
+Provides-Extra: bridge
+Requires-Dist: bridgernadesigner>=0.1.1; extra == "bridge"
+Requires-Dist: ViennaRNA>=2.6; extra == "bridge"
+Requires-Dist: pysam>=0.22; extra == "bridge"
+Provides-Extra: server
+Requires-Dist: fastapi>=0.111; extra == "server"
+Requires-Dist: uvicorn>=0.30; extra == "server"
+Provides-Extra: services
+Requires-Dist: litellm>=1.44; extra == "services"
+Requires-Dist: paper-qa>=5.10; extra == "services"
+Requires-Dist: streamlit>=1.37; extra == "services"
+Requires-Dist: fastmcp>=2.3; extra == "services"
+Requires-Dist: beautifulsoup4>=4.12; extra == "services"
+Provides-Extra: orchestrate
+Requires-Dist: paramiko>=3.4; extra == "orchestrate"
+Requires-Dist: scp>=0.15; extra == "orchestrate"
+Requires-Dist: rich>=13.7; extra == "orchestrate"
+Requires-Dist: python-dotenv>=1.0; extra == "orchestrate"
+Provides-Extra: docs
+Requires-Dist: mkdocs>=1.6; extra == "docs"
+Requires-Dist: mkdocs-material>=9.5; extra == "docs"
+Provides-Extra: dev
+Requires-Dist: pytest>=8; extra == "dev"
+Requires-Dist: ruff>=0.5; extra == "dev"
+Requires-Dist: pytest-cov>=5; extra == "dev"
+Dynamic: license-file
+
+<div align="center">
+
+# PEN-STACK
+
+### The Writable Genome - open infrastructure for genome *writing*
+
+*Editing tools tell you **how** to change a base. PEN-STACK tells you **where** in the genome you can safely
+and durably write new DNA, **which enzyme** can write it there, and **how** to design the write end-to-end.*
+
+[![PyPI](https://img.shields.io/pypi/v/pen-stack.svg)](https://pypi.org/project/pen-stack/)
+[![CI](https://github.com/ahmedanees-m/pen-stack/actions/workflows/ci.yml/badge.svg)](https://github.com/ahmedanees-m/pen-stack/actions/workflows/ci.yml)
+[![Publish](https://github.com/ahmedanees-m/pen-stack/actions/workflows/publish.yml/badge.svg)](https://github.com/ahmedanees-m/pen-stack/actions/workflows/publish.yml)
+[![coverage](https://raw.githubusercontent.com/ahmedanees-m/pen-stack/main/.github/badges/coverage.svg)](https://github.com/ahmedanees-m/pen-stack/actions/workflows/ci.yml)
+[![codecov](https://codecov.io/gh/ahmedanees-m/pen-stack/branch/main/graph/badge.svg)](https://codecov.io/gh/ahmedanees-m/pen-stack)
+[![License: MIT](https://img.shields.io/badge/License-MIT-informational.svg)](LICENSE)
+[![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
+[![Version](https://img.shields.io/badge/version-3.1.0-blue.svg)](CHANGELOG.md)
+[![Tests](https://img.shields.io/badge/tests-115%20passing-success.svg)](tests/)
+[![Lint: ruff](https://img.shields.io/badge/lint-ruff-purple.svg)](https://github.com/astral-sh/ruff)
+[![Runtime: Docker](https://img.shields.io/badge/runtime-docker-2496ED.svg)](docker/)
+[![Validation: pre-registered](https://img.shields.io/badge/validation-pre--registered-critical.svg)](prereg/)
+[![Genome-Writing Bench v0.1](https://img.shields.io/badge/benchmark-Genome--Writing%20Bench%20v0.1-6f42c1.svg)](benchmarks/genome_writing_bench/)
+
+**Built on five prior, separately published repositories:**
+
+[![genome-atlas](https://img.shields.io/badge/built_on-genome--atlas-1f6feb.svg)](https://github.com/ahmedanees-m/genome-atlas)
+[![mech-class](https://img.shields.io/badge/built_on-mech--class-1f6feb.svg)](https://github.com/ahmedanees-m/mech-class)
+[![pen-score](https://img.shields.io/badge/built_on-pen--score-1f6feb.svg)](https://github.com/ahmedanees-m/pen-score)
+[![pen-assemble](https://img.shields.io/badge/built_on-pen--assemble-1f6feb.svg)](https://github.com/ahmedanees-m/pen-assemble)
+[![pen-compare](https://img.shields.io/badge/built_on-pen--compare-1f6feb.svg)](https://github.com/ahmedanees-m/pen-compare)
+
+</div>
+
+---
+
+## What is PEN-STACK?
+
+PEN-STACK is a single, installable, pre-registered computational stack that builds the reference and design
+layer the genome-**writing** era lacks. It consolidates five earlier research projects into one citable
+package, then adds the two reference maps and the design engine the field was missing.
+
+Genome **editing** changes a base or short stretch in place. Genome **writing** installs *new* information -
+inserting genes, flipping or excising kilobases, placing programmable landing pads. Writing is the harder,
+less-tooled, and more clinically transformative modality, and it is gated by questions that today have no
+canonical answer.
+
+## The problem, and the gaps PEN-STACK closes
+
+Two questions gate every genome-writing project, and before PEN-STACK no resource answered them together:
+
+| Gap | The problem today | What PEN-STACK provides |
+|---|---|---|
+| **Where can you write?** | Each lab re-derives an ad-hoc "safe harbour" shortlist from inconsistent criteria; published lists range from ~2,000 sites to 25, none predict expression durability from a learned model, none are writer-aware, most cover one cell type. | **The Writable Genome** - a learned, cell-type-aware, writer-aware atlas scoring every locus for *safety* (genotoxicity risk) x *durability* (will the cassette stay expressed) x *reachability* (which enzyme can engage it). |
+| **What can write there, and how well?** | Enzyme capabilities are scattered across papers; no catalogue places all genome-writing families on common, measured axes with their targeting requirements. | **The Writer Atlas** - 33,370 enzyme systems across 8 families on common measured axes, joined to the Writable Genome by a bidirectional cross-link. |
+| **How do I design the actual write?** | Destination, enzyme, cargo and delivery are interdependent and goal-dependent; no tool optimises them jointly. | **The Write Planner** - inverse design that, given a goal and an `edit_intent`, returns ranked, traceable site x writer x cargo x delivery plans. |
+| **Where might my bridge-recombinase design go off-target?** | Bridge recombinases are the most programmable writers, but had no genome-wide off-target screening tool (a "CRISPOR" equivalent); their developers list this as future work. | **The bridge off-target engine** (`pen-bridge`) - measured-data-validated screening that *nominates and ranks candidate off-target locations* (a screen, not a per-site risk calculator). |
+
+Everything is built on bulk-downloadable public data, runs on a single GPU, and is validated **blind** against
+a pre-registered, honest baseline before release.
+
+## What is new in v3.1
+
+v3.1 hardens the honesty of the planning benchmark, surrounds the models with strong baselines, adds a
+predicted-structure safety axis, and ships the first benchmark and grounded agent for the genome-*writing*
+side of the field. Every workstream is pre-registered (`prereg/ws_*.yaml`, SHA-locked) and reports its
+honest negatives, not just its wins.
+
+| Workstream | What it adds | Honest headline result |
+|---|---|---|
+| **A - De-circularized benchmark** (gate) | retires the circular targeted-intent recovery@k; the headline is now blind safe-harbour discovery | blind GSH discovery **AUROC 0.92** vs safety-only 0.50 |
+| **B - Strong baselines + safety metric switch** | endogenous-expression baseline, multi-mark ablation, published GSH rule-set; safe-harbour discrimination is now the primary safety metric | learned writability **0.92 (95% CI 0.82-0.98)** vs GSH distance-rule 0.38 (delta CI excludes zero); the circular `genotoxic_cis` AUROC demoted to a labeled diagnostic |
+| **C - AlphaGenome integration** | predicted sequence tracks + a predicted **3D structural-risk** axis (Hi-C contact-map deltas) via the hosted AlphaGenome API | per-track transfers well (HepG2 ATAC 0.91), but the *composite* score degrades from predicted tracks, so the measured atlas stays the backbone (flagged) |
+| **D - Cargo Polish** | scores the *insert* for silencing/instability triggers (CpG islands, GC, cryptic splice, MFE, silencers) | directional: high-CpG bacterial cassette 0.75 vs CpG-depleted 0.0, every flag carries a fix |
+| **E - Genome-Writing Bench v0.1 + PEN-Agent** | the first benchmark for the writing side, plus a grounded agent that cannot fabricate | planner beats the naive baseline 3/3; a real LLM agent reaches the planner's numbers only by grounding (0 fabricated) |
+| **F - Local recalibration / private-data adaptation** | recalibrate or fine-tune the released models on your own assays, in-container, behind a validation gate | the adapted model activates only if it beats the released model AND a no-skill baseline; the released model is provably unchanged |
+| **G - Multiplex + guide QC** | a pairwise translocation-risk screen for multi-edit plans, and a bridge-RNA guide ranker | DSB-free recombinase plans carry ~zero translocation risk by construction; known-bad guides are retrospectively down-ranked |
+
+The **Genome-Writing Bench** (workstream E) is v3.1's adoption vehicle: a one-command, SHA-locked, leaderboard
+benchmark with deterministic scorers and no circular labels. See
+[`benchmarks/genome_writing_bench/`](benchmarks/genome_writing_bench/) and `docs/positioning.md`.
+
+## Architecture
+
+```
+                           +-------------------------------------------+
+                           |            WRITE PLANNER (engine)         |
+                           |   inverse design: destination x writer    |
+                           |   x cargo/guide x delivery -> ranked plan  |
+                           +----------------^-------------^-------------+
+                                            |             |
+                  +-------------------------+--+      +---+------------------------+
+                  |    WRITABLE GENOME (B)     |      |      WRITER ATLAS (A)      |
+                  |    flagship reference      |<---->|   companion reference      |
+                  |                            | reach|                            |
+                  |  - Safety layer (learned)  | ability  - Family targeting KB    |
+                  |  - Durability layer (learned)|      |  - Measured scoring axes   |
+                  |  - Reachability layer  -----+------+  - Mechanism classifier    |
+                  |  -> writability profile    |      |  - DMS variant model        |
+                  +-------------------------^--+      +---+------------------------+
+                                            |             |
+         +----------------------------------+-------------+-------------------------+
+         |                       DATA FOUNDATION (bulk-downloadable)                |
+         |  hg38 . ENCODE/Roadmap chromatin . Hi-C/LADs . TRIP position effects .   |
+         |  RID/VISDB/MLV integration sites . clinical genotoxic CIS . COSMIC .     |
+         |  DepMap . gnomAD . GTEx . UniProt . Pfam/InterPro . bridge-recombinase   |
+         |  off-target + DMS (Perry 2025)                                           |
+         +-------------------------------------------------------------------------+
+
+   Platform services (on top of the validated core): PEN-MONITOR (Europe PMC living database),
+   grounded RAG/Q&A, a tool-using agent + MCP server, and a Streamlit web app.
+```
+
+## How it works
+
+PEN-STACK is organised as **two reference layers + one engine + a services layer**.
+
+| Component | Module | Role | Status |
+|---|---|---|---|
+| **Writable Genome** (flagship) | `pen_stack.wgenome` | learned per-locus safety x durability x reachability | Paper 1 |
+| **Writer Atlas** (companion) | `pen_stack.atlas`, `.mech`, `.score` | cross-family enzyme catalogue + Writer-Targeting KB | Paper 2 |
+| **Cross-link** | `pen_stack.atlas.crosslink` | bidirectional writer to locus queries | Paper 2 |
+| **Write Planner** (engine) | `pen_stack.planner` | inverse design, `edit_intent`-conditioned | Paper 3 |
+| **Agentic platform** | `pen_stack.agent` | goal to cited, auditable plan; MCP server; one-command deploy | Paper 3 |
+| **Bridge off-target engine** | `pen_stack.bridge` | "CRISPOR for bridge recombinases" + guide QC (v3.1) | Paper 4 |
+| **Genome-Writing Bench** (v3.1) | `benchmarks/`, `bench/run.py` | first writing-side benchmark; deterministic scorers, leaderboard | M2 |
+| **PEN-Agent** (v3.1) | `pen_stack.agent.pen_agent` | grounded write-planning state machine; zero fabrication | M2 |
+| **3D structural risk** (v3.1) | `pen_stack.wgenome.structure3d` | AlphaGenome contact-map deltas as a safety axis | M1 |
+| **Cargo Polish** (v3.1) | `pen_stack.planner.cargo_polish` | cargo-sequence silencing-risk scan | M1 |
+| **Local adaptation** (v3.1) | `pen_stack.adapt` | gated recalibration / fine-tuning on private data | M1 |
+| **Multiplex risk** (v3.1) | `pen_stack.planner.multiplex` | pairwise translocation-risk screen for multi-edit plans | M3 |
+| **Platform services** | `monitor`, `rag`, `ui`, `server` | living database, grounded RAG, web app, REST API | - |
+
+### Headline results (all blind / pre-registered)
+
+- **Paper 1 (Writable Genome):** a genome-wide atlas of 3,031,030 loci x 3 cell types (K562, HepG2, CD34+
+  HSPC) recovers validated safe harbours as highly writable and clinical genotoxic loci as non-writable,
+  blind. Durability transfers mouse to human (Spearman rho = 0.42).
+- **Paper 2 (Writer Atlas):** 33,370 enzyme systems across 8 families on common measured axes; mechanism
+  classifier agrees with the audited labels on the curated core (1.00); cross-link validated on AAVS1.
+- **Paper 3 / v3.1 (Write Planner + de-circularized benchmark):** the honest headline is **blind
+  safe-harbour site discovery** - run genome-wide (so no on-target identity term fires), the planner's
+  writability separates held-out, DOI-validated safe harbours from matched-context controls at **AUROC
+  0.92** (safety-only baseline 0.50). Writer-family recovery@1 = 1.0 vs prevalence 0.25 across 4 families.
+  The earlier "recovery@10 = 1.00, McNemar p" result for *targeted* intents was definitional, not
+  predictive (an on-target identity term dominates the score), so it is now reported only as a
+  specification-compliance correctness table - see `docs/benchmark_circularity.md`. A tool-using agent
+  never fabricates a number (every value traces to a validated tool call).
+- **Paper 4 (Bridge off-target engine):** to our knowledge the first measured-data-validated tool that
+  **nominates and ranks candidate off-target *locations*** for bridge recombinases. On the measured Perry
+  2025 data (6,856 real off-targets) the per-position profile confirms the central core (positions 7-9) is
+  the specificity determinant, and the model ranks real off-targets above core-disrupted decoys at AUROC
+  0.77 vs 0.62 for Hamming. Stated plainly: it is a **screening tool, not a quantitative safety
+  calculator**, it does not quantify how much recombination occurs at each site (sequence-risk vs measured
+  magnitude, rho approximately 0.30). A first-of-its-kind beachhead for a genuinely unoccupied gap, not a
+  Nature-tier breakthrough; the Writable Genome (Paper 1) remains the flagship novelty.
+
+## The Genome-Writing Bench (v3.1, M2)
+
+The first benchmark for the **writing** side of genome engineering - *where* to write, *what* writer to use,
+*how* to design the cargo, and *what off-target / structural risk* a write carries - complementing the many
+editing-side (Cas9 / base / prime) benchmarks. Six tasks, each with a deterministic scorer and a documented
+ground-truth source; **no task is scored against a circular label** (it inherits the de-circularization gate).
+
+```bash
+python bench/run.py --agent          # one command -> out/bench_results.json + a leaderboard
+docker compose run --rm bench python bench/run.py --agent   # same, on the clean image
+```
+
+| Solver | Beats naive on | No-fabrication | Note |
+|---|---|---|---|
+| deterministic planner | 3/3 grounded tasks | n/a | the validated planning tools (reference) |
+| naive baseline | - | n/a | safety-only / prevalence / Hamming |
+| **LLM agent** (PEN-Agent) | = planner (grounded) | **PASS** | a real LLM drives the tools; reaches the planner only by grounding every value, 0 fabricated |
+
+Per-task (planner vs naive): site selection **0.92** vs 0.50, writer recovery **1.0** vs 0.25, off-target
+**0.77** vs 0.62, intent 7/7, no-fabrication **PASS** (a hard gate). **PEN-Agent** (`pen_stack.agent`) is a
+grounded write-planning state machine - goal to site to writer to cargo (with Cargo Polish) to off-target
+to 3D structural risk to report - that copies every number from a validated tool with provenance and refuses
+or degrades rather than invent. See [`benchmarks/genome_writing_bench/`](benchmarks/genome_writing_bench/),
+`docs/agent.md`, and the leaderboard submission guide.
+
+## How PEN-STACK connects to the prior repositories
+
+PEN-STACK v3.0 consolidates and re-grounds five earlier projects. Their genuinely reusable assets are
+imported here; the originals are archived read-only for provenance and DOI stability. This is what makes
+PEN-STACK "the thing you cite instead of rebuilding the pipeline."
+
+```
+  genome-atlas  --+  18-family InterPro-audited Pfam whitelist (v1.2.1)  -->  WT-KB + mechanism classifier
+  mech-class  ----+  multi-source mechanism classifier                   -->  family / mechanism calls
+  pen-score  -----+- 9 scoring axes (dsb/cargo/deliv/immuno/prog/...)     -->  re-grounded therapeutic axes
+  pen-assemble  --+  IS110 ortholog / design set                         -->  part of the 1,058-entity universe
+  pen-compare  ---+  unified_editor_universe.parquet (1,058) + scorecard  -->  canonical universe + scorecard
+```
+
+| Prior repo | Pinned version | What v3.0 reuses | What changed |
+|---|---|---|---|
+| [genome-atlas](https://github.com/ahmedanees-m/genome-atlas) | v0.7.2 | the audited 18-family Pfam backbone - spine of the WT-KB and the at-scale mechanism classifier | GraphSAGE link-prediction framing retired |
+| [mech-class](https://github.com/ahmedanees-m/mech-class) | v0.5.4 | the mechanism classifier (Pfam + RHEA + CRISPRcasdb + UniProt) | reused as the family/mechanism caller |
+| [pen-score](https://github.com/ahmedanees-m/pen-score) | v0.1.3 | the scoring axes (deliv / immuno / cargo, ...) | prog/cargo re-grounded; hand-set overrides removed |
+| [pen-assemble](https://github.com/ahmedanees-m/pen-assemble) | v0.5.2 | the ortholog sequence set | de-novo chimera generation retired -> DMS-grounded point-variant proposal |
+| [pen-compare](https://github.com/ahmedanees-m/pen-compare) | v0.1.0 | the 1,058-entity universe + scorecard scaffold + tests | circular 5-gate "certification" -> descriptive scorecard with blind concordance |
+
+**One canonical assembly path** (`pen_stack/atlas/universe.py::assemble`) feeds the classifier, the scorer,
+and the scorecard identical metadata, so the cross-module inconsistency in the prior pipelines cannot recur.
+
+## Repository structure
+
+```
+pen-stack/
+├── pen_stack/                        the installable package
+│   ├── wgenome/                      Writable Genome (Paper 1)
+│   │   ├── features.py               unified feature matrix (accessibility + histones + safety + integration)
+│   │   ├── safety.py                 calibrated genotoxicity-risk model (chrom-block CV + baseline)
+│   │   ├── durability.py             conditional chromatin->expression model (TRIP-trained, transferable)
+│   │   ├── writability.py            decomposable safety x durability x reachability integration
+│   │   └── export_tracks.py          BigWig / BED atlas export
+│   ├── atlas/                        Writer Atlas + WT-KB + cross-link (Papers 1-2)
+│   │   ├── schema.py                 pydantic WriterEntry (enforces >=1 DOI per row)
+│   │   ├── build_wtkb.py             Writer-Targeting Knowledge Base builder (8 families, tiered)
+│   │   ├── expand.py                 ortholog ingestion -> atlas.parquet (33,370 systems)
+│   │   ├── crosslink.py              writers_for_locus / loci_for_writer / loci_for_gene
+│   │   ├── variant_propose.py        DMS-grounded point-mutation proposal (no chimeras)
+│   │   ├── universe.py               THE canonical universe assembly (1,058 entities)
+│   │   └── scorecard.py              descriptive scorecard + blind concordance
+│   ├── mech/                         mechanism classification at scale (audited 18-family whitelist v1.2.1)
+│   ├── score/                        re-grounded axes + therapeutic-readiness scoring
+│   ├── planner/                      Write Planner (Paper 3): optimize / cargo / cargo_polish / multiplex / pipeline
+│   ├── bridge/                       bridge off-target engine (Paper 4): offtarget / fold_qc / guide_qc / pipeline / cli
+│   ├── agent/                        agentic platform: tools / orchestrator / pen_agent / mcp_server / guardrails
+│   ├── adapt/                        local recalibration / private-data adaptation behind a gate (v3.1, WS-F)
+│   ├── monitor/                      PEN-MONITOR living database (Europe PMC)
+│   ├── rag/                          grounded, cited Q&A (hybrid LLM: Ollama primary, Nemotron fallback)
+│   ├── validate/                     benchmarks: blind_gsh_discovery / durability_baselines / seq_vs_measured / agent_eval / adapt_demo
+│   ├── data/                         ingestion (genome, chromatin, integration, TRIP, safety annotations)
+│   ├── server/api.py                 FastAPI REST (atlas, crosslink, writable, plan, bridge, ask)
+│   ├── ui/app.py                     Streamlit web app (11 pages)
+│   └── cli.py                        unified CLI
+├── benchmarks/genome_writing_bench/  Genome-Writing Bench v0.1 (tasks / harness / solvers / LEADERBOARD / SHAs)
+├── bench/run.py                      one-command bench entrypoint (--agent, --verify)
+├── scripts/                          reproducible pipeline drivers (p1_*, p2_*, p4_*, ws_*_report)
+├── configs/                          pinned datasets + thresholds + curation (YAML)
+├── prereg/                           SHA-locked success criteria (paper1..4 + ws_a..ws_g + locks)
+├── data/curated/                     small committed tables (universe, gene coords, measured bridge profile)
+├── tests/unit/                       unit + regression + blind-validation suite
+├── docs/                             mkdocs site (cards, tutorials, INFRA, DEPLOY, MCP)
+├── docker/                           CUDA image + UI image + pinned requirements
+├── tools/penctl.py                   laptop<->VM orchestrator (paramiko SSH/SFTP, Docker-only)
+├── docker-compose.yml                one-command self-hostable platform
+└── pyproject.toml  CITATION.cff  CHANGELOG.md  LICENSE
+```
+
+> **Data policy.** Large artifacts (3 M-row atlases, BigWig tracks, models) and any third-party copyrighted
+> data are *not* committed - they are released via Zenodo (DOI) or fetched from the original source, and are
+> reproducible by re-running the scripts. Only small curated tables and derived products live in git.
+
+## Installation and quick start
+
+**From PyPI** (the library, CLI, agent, and pure-logic tools):
+
+```bash
+pip install pen-stack            # core
+pip install "pen-stack[models,bio,bridge,server,services]"   # full stack
+```
+
+The wheel ships the importable package and the command-line tools. The **full data pipeline** (the 3 M-row
+atlases, BigWig tracks, and curated configs) is distributed via the cloned repo + Zenodo, per the data
+policy below; point an installed copy at a checkout with `export PEN_STACK_HOME=/path/to/pen-stack` to use
+the config-driven features. Most users who want the whole pipeline clone the repo:
+
+```bash
+git clone https://github.com/ahmedanees-m/pen-stack.git && cd pen-stack
+pip install -e ".[dev]"                                   # core + tests
+pip install -e ".[models,bio,bridge,server,services]"     # full stack
+pytest -q                                                 # 115 tests
+pen-stack info                                            # stack status
+python bench/run.py --agent                               # run the Genome-Writing Bench (under 5 min)
+```
+
+A five-minute quickstart that runs a bench task end-to-end is in [`docs/quickstart.md`](docs/quickstart.md).
+
+Query the stack:
+
+```bash
+pen-stack atlas --coverage                                # Writer Atlas coverage (33,370 systems x 8 families)
+pen-stack writable --gene CCR5 --ct k562                  # rank writable loci near a gene
+pen-stack crosslink --chrom chr19 --bin 55090             # which writers reach AAVS1
+pen-stack plan --gene TRAC --intent knock_in_with_disruption --cargo-bp 2000   # inverse-design plans
+pen-bridge design --target ACGTGTCTACGTGA --donor TTGCATCTAGGCAC               # bridge design + off-target + QC
+pen-stack monitor --back-test                             # PEN-MONITOR living-database scan
+```
+
+Self-host the whole platform (API + web app + agent + MCP + LLM), one command:
+
+```bash
+docker compose up -d
+docker compose exec ollama ollama pull qwen2.5:7b-instruct   # first run only (local fallback model)
+# Web app :8501  .  API :8000 (/plan, /bridge/design, /ask)  .  MCP :8765   (see docs/DEPLOY.md)
+```
+
+**LLM backend (hybrid, non-load-bearing).** Services (agent, RAG, PEN-MONITOR) use one switch in
+`configs/llm.yaml`. On the compute tier (the GPU VM) the default is the **local Ollama model**
+(`qwen2.5:7b-instruct`, free, private, tool-calling verified) with **automatic fallback** to the hosted
+**NVIDIA Nemotron** (free, no local resources), then to a deterministic no-LLM path. A cooldown cache and
+bounded timeouts mean an absent or slow provider degrades in seconds rather than stalling. The LLM is
+non-load-bearing - every number and citation comes from a validated tool - so the choice never affects
+scientific reproducibility, only orchestration quality. Set `NVIDIA_API_KEY` (or a gitignored
+`configs/nvidia_api_key.txt`) for the hosted fallback; a low-RAM laptop with no GPU uses it automatically.
+The core scientific compute uses no LLM at all.
+
+## The web platform
+
+`pen_stack/ui/app.py` is a single Streamlit app over the whole stack (11 pages):
+
+- **Writable Genome** - Overview, Forward query (gene to writability/safety/durability), Site finder
+  (inverse), Atlas browser, Validation dashboard, Cross-cell-type transfer.
+- **Writer Atlas** - family coverage and measured-axis comparison.
+- **Write Planner** - goal + `edit_intent` to ranked, traceable plans.
+- **Bridge design** - design a bridge RNA, fold/cross-loop QC, genome-wide off-target scan.
+- **Ask** - grounded, cited Q&A (numbers from validated tools).
+- **Agent** - a goal to a cited, auditable end-to-end plan.
+
+## Data sources (all public)
+
+hg38 (UCSC); ENCODE / Roadmap chromatin (ATAC/DNase + histone marks; K562, HepG2, CD34+ progenitor, mouse
+ES-Bruce4); GENCODE v46; COSMIC Cancer Gene Census v104; DepMap Public 26Q1; LaFave 2014 (NHGRI GeIST) MLV
+integrations; VISDB; TRIP / Akhtar 2013 (GEO GSE49806/49807); UniProt orthologs; Pfam/InterPro; Europe PMC;
+Addgene; Perry 2025 bridge-recombinase off-target + DMS data (Science adz0276; copyrighted - kept local,
+only derived products released). Every accession and DOI is pinned in `configs/datasets.yaml` and
+independently verified.
+
+## Validation philosophy
+
+- **Pre-register before training.** Success criteria, baselines and held-out sets are SHA-locked in
+  `prereg/` (paper1..4) before any model sees test data.
+- **Always report an honest baseline** (oncogene-distance for safety; H3K9me3/LAD for durability;
+  intent-blind ranking for the Planner; Hamming for the bridge engine).
+- **Blind external concordance** - recover validated safe harbours, clinical genotoxic loci, documented
+  writes, and measured off-targets the model never trained on.
+- **Report failure honestly** - cross-cell-type degradation, small benchmark N, and the limits of
+  sequence-only off-target magnitude prediction are quantified results, not footnotes.
+- **Grounded services** - every quantitative answer comes from a validated tool call (never a language
+  model); the living database never auto-edits the atlas; clinical directives are refused.
+
+## Papers and phases
+
+| # | Title | Phase | Status |
+|---|---|---|---|
+| 1 (flagship) | The Writable Genome: a predictive, writer-aware atlas of safe & durable insertion sites | 1 | complete |
+| 2 (platform) | PEN-STACK: unified open infrastructure for non-destructive genome writing | 2 | complete |
+| 3 (capstone) | The Write Planner: end-to-end inverse design of genomic writes | 3 | complete |
+| 4 (beachhead) | Genome-wide off-target prediction for RNA-guided bridge recombinases | 1.5 | complete |
+| M1 (v3.1) | Writable Genome hardened: strong baselines, AlphaGenome sequence + 3D structural-risk axis | v3.1 B,C,D,F | complete |
+| M2 (v3.1) | The Genome-Writing Bench + PEN-Agent: the writing-side benchmark and a grounded agent | v3.1 E | complete |
+| M3 (v3.1) | Multiplex translocation-risk + bridge-RNA guide QC | v3.1 G | complete |
+
+The v3.1 cycle (workstreams A-H) is recorded in `CHANGELOG.md`, `docs/positioning.md`, and the SHA-locked
+`prereg/ws_*.yaml`; preprint drafts are in `manuscripts/`.
+
+Per-phase build records, execution summaries, and Zenodo deposit packages are kept alongside the program
+plan. Data releases are deposited on Zenodo (one per paper).
+
+## Citation
+
+```bibtex
+@software{penstack2026,
+  author  = {Mahaboob Ali, Anees Ahmed},
+  title   = {PEN-STACK: open infrastructure for genome writing (The Writable Genome)},
+  year    = {2026},
+  version = {3.1.0},
+  url     = {https://github.com/ahmedanees-m/pen-stack}
+}
+```
+
+**Author:** Anees Ahmed Mahaboob Ali, VIT University, Vellore. MIT licensed.
+
+*Decision-support, not a clinical directive - every score is traceable to public data and a pre-registered
+model.*

pen_stack-3.1.0.dist-info/RECORD

@@ -0,0 +1,96 @@
+pen_stack/__init__.py,sha256=KCoEBgVyUpSQazTpaMXKHTTn_j70m6ki2xxRK4WQEC8,87
+pen_stack/_resources.py,sha256=8A5SDvkx64AbSRnlyancO1SpZp5NNHd2hbrO0LrGteI,1578
+pen_stack/cli.py,sha256=FLy8pwSQBnAqRJciS3_U_T31b91dKLjo3wEm0k4S6Ug,5745
+pen_stack/adapt/__init__.py,sha256=zRwPjyDVQcK0hXoH4_qkotQfrOlJna9GnBd_3SUsIY0,791
+pen_stack/adapt/finetune.py,sha256=EtjZFtBPYvBbH9g0oKWPAS2xCRVdHhqVtk7c5havQWM,1469
+pen_stack/adapt/ingest.py,sha256=JjQx8zNe0H9hmeOAGrpq-JFCuVjoULFzXWLgy01cXyU,4447
+pen_stack/adapt/pipeline.py,sha256=zs_fDAiv0LtuTTxwoIeCR9GjhvodviSpp1bwYQlNHb0,5389
+pen_stack/adapt/recalibrate.py,sha256=MGdDWeROqUmcpz6mT-Pmlp3NPmUMs52zRNFm55-twtI,2403
+pen_stack/adapt/report.py,sha256=ebYbtlMQ5wuJvNYwdeIHvgGZdWsVG3SbcI-mfIWN9QU,6343
+pen_stack/agent/__init__.py,sha256=J05GUK7ymg7jY1fTvJ9lUgTGW-y4_7hBXuLIHDpt7bk,57
+pen_stack/agent/guardrails.py,sha256=PF8hxBLbAEtU6vk0p3CiBARHw4cjEHYRF8kPWDuLTeQ,2644
+pen_stack/agent/mcp_server.py,sha256=HYh-H3a29SuW7yqGEh3tDoPeMpLFSYIKOT2yM1dMj4U,1980
+pen_stack/agent/orchestrator.py,sha256=BgCHijM1DjhtfwrM7F1VeS_akoU1QVAlJ_f6hd4RFmE,5585
+pen_stack/agent/pen_agent.py,sha256=Q_hChhFgsVjk4MrP0UiVAIm6bX2WstCuX15No01PIxg,9097
+pen_stack/agent/tools.py,sha256=CSvGexFtM6imFXjTjaMknt_MMwn3D6XaV7kxi7g0qmQ,6617
+pen_stack/atlas/__init__.py,sha256=-XGrYvi4u2qIM5FTgukHweLjCDe3SQipKs7_HvxTKx4,57
+pen_stack/atlas/build_wtkb.py,sha256=mBvm7swjqhFBkhrTKigT0ffLnzk-_d3h6qk9K0eoFHc,3177
+pen_stack/atlas/crosslink.py,sha256=nJWvpk_0AtrUIZbBZ1C25Az1EgTQQuPwaYHpDwDYT3c,6432
+pen_stack/atlas/expand.py,sha256=5SZt6tS3SLt9uZwEQ3YlYdoIzSIQ5MRX9QASttODRbw,8898
+pen_stack/atlas/schema.py,sha256=VLAItD-uGjrsySscBNnjr_lvbf06lByK7gribD15ogs,2410
+pen_stack/atlas/scorecard.py,sha256=cMkuUH3Ty_n8cSWL5EUfrMRouKskpkdx6Se3xyrGTPA,6136
+pen_stack/atlas/universe.py,sha256=BgwMjFAPC5W5rCprPcte84qXIn3phlBCVBNltL2tRkY,3218
+pen_stack/atlas/variant_propose.py,sha256=q5EJtKzrg2c7OQVTppdKn9Q0UwfJ_VTsJa3-tJBiXH0,8073
+pen_stack/bridge/__init__.py,sha256=QOESbsA8NIBPC6dSFS5qBtVnLJ1Ipu3_dPJWQ_5CYcQ,58
+pen_stack/bridge/activity.py,sha256=x4H2MoHw4sGp0o3Fe2xiDIjve7aGHNjVqbnJkgouI6s,2682
+pen_stack/bridge/cli.py,sha256=qGTjYoW2KBvs6B2T5gx92ofLlONzOo_2xZR-uHyJoak,3072
+pen_stack/bridge/fold_qc.py,sha256=CyX-zeoZPd6MRNfarQM4hbP8zw_TnHmMwenkFZHL5vE,2507
+pen_stack/bridge/guide_qc.py,sha256=Qdc2rrz2DK8ODZKl9iii1LiSk6ETAk7cJwfLhqUrGnk,4491
+pen_stack/bridge/ingest.py,sha256=43lk_6KrabZVMxpXN1EzB5fcUDtGfOHdPn8FSpq98Wc,6095
+pen_stack/bridge/offtarget.py,sha256=jKUHog0J9Pwf9rtEKgVau_H6wyU1Gq7Ns32ae5uqb_I,6425
+pen_stack/bridge/ortholog_screen.py,sha256=HHXq3WONaZgOmZicYzP56epy6L0vhBUhGBFcf98KIYI,3380
+pen_stack/bridge/pipeline.py,sha256=-ebgqppDGSYLJ2t91IKifj96BGB4lfxf7BsUlsON2GU,3830
+pen_stack/data/__init__.py,sha256=jpgYY9g7VwyFZXIifhlssZRt6b-ZUdxKR0dsmL_muBo,56
+pen_stack/data/encode.py,sha256=n9sW_lJOJq5-niflG_Owd10PdZFePphcGrcno1s5T6c,3146
+pen_stack/data/genome.py,sha256=X5KM1I6cfNM3wMBJVaesXMBgNEPeeSXQFIVpyMkqQhY,2419
+pen_stack/data/ingest_chromatin.py,sha256=J4U1K_Y_U5dQRv6t5YKGw2mpW8JoHEMvFK5cew4j37w,4425
+pen_stack/data/ingest_integration.py,sha256=QOVMbcrmcvELywA0vr_lSY13zNx7PdgpR3lmelm-WAk,5014
+pen_stack/data/ingest_safety_annot.py,sha256=XMqXKImy9RjhwBq_N-V5S7b-e6MhNN9N3Xq5yBZDju0,6818
+pen_stack/data/ingest_trip.py,sha256=IeH0TXEWHL3z1ahID235HZO_KHdUF8g7fpVtFpGJhV4,3733
+pen_stack/mech/__init__.py,sha256=GAILAvLwDv8InNGMIWNRiFIQaXUOXrv5m0ssDgSguWw,56
+pen_stack/mech/classify_atlas.py,sha256=OOxPCUj2qT9WMbsXjFx7CXNExFSSYvMlRBlruSWFHU0,3271
+pen_stack/mech/whitelist.py,sha256=OPKvqXy5_D5iHqCL84dlxfREqHlQOTPjNzNYZb-cBHs,3329
+pen_stack/monitor/__init__.py,sha256=6tXqDuWVhNxOulYgpoDs03Y8dUZshstJ774kcbzJp_M,59
+pen_stack/monitor/europepmc.py,sha256=igxEZycVFEJVyXk9DNJKHYXaOHrWOKuwRn4zjVh99-Q,1374
+pen_stack/monitor/run.py,sha256=WlQsy7HGu41HOy-O3EXe8DswcDTT2hElRgHzarg1x1k,2420
+pen_stack/monitor/triage.py,sha256=clt0hTN-ZHRu2s_sQqwwifMBblzamGW9Abec_4wdoZc,2630
+pen_stack/planner/__init__.py,sha256=rqXwN3F-726t6f1lFVTfozKlwo1A3SOOiI8TDyI-1QM,59
+pen_stack/planner/cargo.py,sha256=sWPXpWKaz47OX7ylqhwjt78GtExBVYn7Owit-4i_SO4,2744
+pen_stack/planner/cargo_polish.py,sha256=P_o4FFT_E3NEFf7I2JBG3KNEuwIA_SXb477IgAhtk94,6209
+pen_stack/planner/delivery.py,sha256=yVk79c05qM359LQPV_vE1UU_NpPXX2AlSkq6Mm9ARKg,1293
+pen_stack/planner/multiplex.py,sha256=zNWeGbJSrwEnFrGIl0SBMk0OfyfPt09TWZCfx1pdNOg,5960
+pen_stack/planner/optimize.py,sha256=m8KtxFc7b1EuoVlhl7Sz20LSb0Gc9R_0rpZ-1UY1Xes,7330
+pen_stack/planner/pipeline.py,sha256=PCxpxaKVhUwrAVmA1NFC6NWOIwtBssIKMLm8YRO4xyI,4162
+pen_stack/planner/report.py,sha256=HsQ6EsyYwEd_406y5WTr_MIxmclz-KAQPYYk-Uj_37M,1335
+pen_stack/rag/__init__.py,sha256=FR6Du4rVBnTVtrFXKCB-jhS-A_Z5L8nDTbbHyc4BSRo,55
+pen_stack/rag/index.py,sha256=gUdOystdW9QYhnpdccq4CblUOXRCW9DeZ1s9c7otBAU,2400
+pen_stack/rag/llm.py,sha256=-yfILBK78P43Gt5-GX9583G95N4gS0sMNfR5InePcwI,8072
+pen_stack/rag/qa.py,sha256=P9cau4k55QTy4pExbVxulDHaihe8wFgcu4THpq5ap5o,5348
+pen_stack/score/__init__.py,sha256=xxvDK6A2MiR3f9V8a7-LX0yxyq_s7yIgiKbF5iuQqok,57
+pen_stack/score/recalibrate.py,sha256=G-YpuE-AOZuXTmQwop3ZhqWFgsVztC9a3SUXtlQxZZE,3100
+pen_stack/score/therapeutic.py,sha256=iDsbLmoTLQ2MLmCwoGowbtNRkWJsSn6YSQNvcUrNwws,3620
+pen_stack/server/__init__.py,sha256=zg0WieD0JB731SfZeMR73PgJOXTPmwMax1_x_JSx3CE,58
+pen_stack/server/api.py,sha256=NdPX34VMoqKyJxs3U79-WhZSiFE988sLOZ_UGAkO_2Q,6323
+pen_stack/ui/__init__.py,sha256=rsfdyJ87h36XaCqoUvhBm1x5IigUNEuGacxkZTsjkTE,54
+pen_stack/ui/app.py,sha256=S1ZzQ0losBkRUxYmI3GJlFA1TpUDYndKP99TMPTNKKw,30815
+pen_stack/validate/__init__.py,sha256=BEu8bTzGd42CFvtpaOag9_SU8VLQzN_pd0mkxZzu_xo,60
+pen_stack/validate/adapt_demo.py,sha256=DMFVU2OD5TE91evE0H2r4MH4HrrbXyI20hACaM5VU7I,3925
+pen_stack/validate/agent_eval.py,sha256=RwBW2kdNP4pSeDWOjQ_Cw9FVsgUdzaukkTm7REgA0eI,5994
+pen_stack/validate/blind_gsh_discovery.py,sha256=NV1QVLMUsuNdIjM4TVOWdcXfReaGuacFpSI_fZuLxVs,8154
+pen_stack/validate/cargo_directionality.py,sha256=Tlo673ySL3vGENryyv7JLq--UWLP_ioy_iP08BL8iWk,2851
+pen_stack/validate/durability_baselines.py,sha256=LRKJLsDByLnFkwopXC84HBuibrOElKBncZbhqCeLGOs,8422
+pen_stack/validate/forward_hypotheses.py,sha256=0uRhd7JcDgNitqzzo5SGSCiqhpA6OrvdN4xqT_ykpJg,5007
+pen_stack/validate/guide_qc_demo.py,sha256=mRNr9JhGGp73js3Wj8gL-f6Hfi3rgiZZ1x1Zv3eM9AQ,2781
+pen_stack/validate/intent_specification.py,sha256=OuoMYVR9Si1i0IOAm51Qx3NFbqurJlFFJcPi0qIGRRg,3626
+pen_stack/validate/paper3_benchmark.py,sha256=xV9jAJpUBpVFozjX4OVHntaNAzpT1vgyqRgTtCKSu7s,8364
+pen_stack/validate/paper4_real_validation.py,sha256=FoVtYFgsgtYEkoed-3c2mIp9wuKGiehJOGBubc1EFB4,6970
+pen_stack/validate/paper4_validation.py,sha256=MmiDLu4GV-ouoNAj1QWgGCu5vcRDPpPogwWbREJYLJc,3904
+pen_stack/validate/seq_vs_measured.py,sha256=LF342iF4aS48FK6dZW72iw2IAMqma-63UbD74Oudq2A,7288
+pen_stack/validate/within_locus_ranking.py,sha256=ft2cAjjBR0Z92r7xEL1hoV5hBeazK9UeWSwwYxP89XM,3350
+pen_stack/validate/writer_recovery.py,sha256=Lv5OaN-BW6vlnhxDp5zizW_9n0xB7XvIhjvN2EgKspA,4208
+pen_stack/wgenome/__init__.py,sha256=enDHFWKOmEOoyO7BHh0MeM7WrZqyow3Y_EWQtoS6ONE,59
+pen_stack/wgenome/chromatin_seq.py,sha256=BTlZ79DhM95My9WEO03Bfx__20W84VHaOxpibpb3Hg4,4095
+pen_stack/wgenome/durability.py,sha256=zxfRNCm0Ba-o7uZRfS5IvHMWLVA4Wj-2LsY4Hgva8Pw,5273
+pen_stack/wgenome/export_tracks.py,sha256=TbFIsieSXmC7oTkwD0LdsGQPEKlCu0KJETDavh_GTks,2037
+pen_stack/wgenome/features.py,sha256=2IuAPB6PQbrOIwAKMIC4DTiv6op8r20sJYLs9WUuksw,3850
+pen_stack/wgenome/gsh_baseline.py,sha256=24RWbtEs-X59iN5a91xfKJ-m5zkQQ3u7qRlV9qgq7gE,5939
+pen_stack/wgenome/providers.py,sha256=0Vc7D2duge86YFK4Ub-WLwQess8LtgTMzGGcBmslylI,11703
+pen_stack/wgenome/safety.py,sha256=j4cOxWEDgk0J7aXQgSRsFGq2kYGgXpiiqG-YVwODcI0,3204
+pen_stack/wgenome/structure3d.py,sha256=jt0Cofbkjv_Rm4GP7nVX4BDQ1nGM-c8txl0Op-krEfE,9250
+pen_stack/wgenome/writability.py,sha256=m6RlG01r0w_UmbV08puBqNwmbwbm4OvzHenFLR0GGtc,3424
+pen_stack-3.1.0.dist-info/licenses/LICENSE,sha256=vxjNCQMsjGGvVRpJfJOg6RWHgXzyTwmxYUu0WF5jUbI,1081
+pen_stack-3.1.0.dist-info/METADATA,sha256=C5ubcAyyQQnJRLfj05zkUWmOZK_-IuVZxk_r56FSdxE,32315
+pen_stack-3.1.0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+pen_stack-3.1.0.dist-info/entry_points.txt,sha256=5FKPTemiVvUgmeecTWOrvwO-fiuHDpJ9y0ITK5BIO4E,88
+pen_stack-3.1.0.dist-info/top_level.txt,sha256=0jV2NCFBddUK9spN6ubdoD8w27XBdWg3sMZb-HM648Y,10
+pen_stack-3.1.0.dist-info/RECORD,,

pen_stack-3.1.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (82.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

pen_stack-3.1.0.dist-info/entry_points.txt

@@ -0,0 +1,3 @@
+[console_scripts]
+pen-bridge = pen_stack.bridge.cli:main
+pen-stack = pen_stack.cli:main

pen_stack-3.1.0.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Anees Ahmed Mahaboob Ali
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

pen_stack-3.1.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+pen_stack