litelink 0.1.0__tar.gz

litelink-0.1.0/.gitignore

@@ -0,0 +1,38 @@
+# Python-generated files
+__pycache__/
+*.py[oc]
+build/
+dist/
+wheels/
+*.egg-info
+
+# Virtual environments
+.venv
+
+# Tool caches
+.pytest_cache/
+.ruff_cache/
+.mypy_cache/
+
+# Local capture roots: a SQLite buffer, a file:// warehouse, and the two
+# catalog databases (docs/SPEC.md §1). Never commit a database.
+litelink-data/
+litelink-ws/
+*.db
+*.db-wal
+*.db-shm
+
+# Fetched binaries: the litestream sidecar (`just litestream`). Not a project
+# dependency — a separate process, pinned rather than vendored.
+.bin/
+
+# Working plans. Scaffolding for a change rather than documentation of the
+# system — what survives a plan belongs in SPEC or RUNTIME, and what does not
+# is answered by the commit messages.
+docs/plans/
+
+# Claude work artifacts
+.claude/worktrees/
+/.roborev/
+.env
+src/litelink/.bin/

litelink-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

litelink-0.1.0/NOTICE

@@ -0,0 +1,4 @@
+litelink
+Copyright 2026 nhobin219
+
+This product includes software developed by nhobin219.

litelink-0.1.0/PKG-INFO

@@ -0,0 +1,310 @@
+Metadata-Version: 2.5
+Name: litelink
+Version: 0.1.0
+Summary: Durable append-only capture into Iceberg tables. Embedded and local-first.
+Project-URL: Source, https://github.com/nhobin219/litelink
+Project-URL: Specification, https://github.com/nhobin219/litelink/blob/main/docs/SPEC.md
+Author-email: Nico Bautista Hobin <nico@penca.io>
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Keywords: append-only,capture,iceberg,parquet,sqlite
+Classifier: Development Status :: 3 - Alpha
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Database :: Database Engines/Servers
+Classifier: Typing :: Typed
+Requires-Python: <3.14,>=3.11
+Requires-Dist: duckdb<1.5.6,>=1.5.5
+Requires-Dist: pyarrow>=17.0.0
+Requires-Dist: pyiceberg[pyarrow,sql-sqlite]>=0.11.1
+Requires-Dist: xxhash>=3.6.0
+Description-Content-Type: text/markdown
+
+<p align="center">
+  <img src="docs/assets/litelink-logo.svg" alt="litelink" width="330">
+</p>
+
+[![CI](https://github.com/nhobin219/litelink/actions/workflows/ci.yml/badge.svg)](https://github.com/nhobin219/litelink/actions/workflows/ci.yml)
+[![license](https://img.shields.io/badge/license-Apache%20v2-blue)](LICENSE)
+[![Python](https://img.shields.io/badge/python-3.11%20%7C%203.12%20%7C%203.13-blue)](pyproject.toml)
+[![Iceberg](https://img.shields.io/badge/Apache%20Iceberg-v2-4B8BBE)](https://iceberg.apache.org/)
+
+# Durable append-only capture into Iceberg tables
+
+**Embedded and local-first.**
+
+## Introduction
+
+litelink is a Python library for the thing every capture pipeline hand-rolls badly: getting a
+stream of observations onto disk durably, into well-sized Parquet, and eventually into object
+storage — without a daemon, a broker, or a catalog service. `append()` returns once the row is
+durable, and a query a moment later sees it.
+
+```
+SQLite buffer          durable on commit. unsealed rows only.
+      │  seal at target_seal_size
+      ▼
+local Iceberg table    a rolling window. reads land here.
+      │  sync: upload data files, register into the archive
+      ▼
+remote Iceberg table   full history, on S3.
+```
+
+Reads span all three tiers, and the catalog is a SQLite file rather than a service, so **no
+read on the hot path touches the network.** Every other machine reads the archive instead,
+with any Iceberg engine and nothing from litelink.
+
+It exists because doing this by hand goes wrong the same way every time: one production
+capture system had 125,884 objects, 62.5% of them under 16 KiB, Parquet files at 2 rows each,
+a compaction routine nothing ever scheduled, and an in-memory buffer a `SIGKILL` emptied.
+Durable capture, file sizing and tiering are each easy alone and nobody's job together.
+
+**Status: early.** All three tiers work, and a log survives losing its machine. Read
+[what it is not](#what-it-is-not) and [not implemented yet](#not-implemented-yet) first.
+
+## Install
+
+```bash
+pip install litelink        # or: uv add litelink
+```
+
+That is the whole of it. The wheel carries what the library shells out to — a
+checksum-verified litestream, and the DuckDB `iceberg`, `avro` and `httpfs` extensions built
+for the DuckDB it pins — so a machine with no egress, nothing on `PATH` and no DuckDB
+extension cache still reads, writes and restores. Verified that way in CI, not assumed.
+
+That costs about 124 MB per wheel, and it buys the failure mode you do not want: a missing
+binary discovered during a failover, or an extension that autoinstalls fine on your laptop
+and cannot on the box that matters.
+
+```bash
+python -m litelink          # PASS/FAIL per requirement, non-zero exit if not ready
+```
+
+Platform wheels are published for Linux and macOS on x86-64 and arm64. Anywhere else, pip
+builds from the sdist — which produces a working pure-Python wheel with no binaries — and you
+supply litestream and the DuckDB extensions yourself; `python -m litelink` tells you which are
+missing and how to get them.
+
+## Quick start
+
+```bash
+git clone https://github.com/nhobin219/litelink && cd litelink
+just bootstrap         # uv sync + git hooks + DuckDB extensions
+just demo-websocket    # capture a live public feed, one process, ~30 seconds
+```
+
+For working ON litelink you need [`uv`](https://docs.astral.sh/uv/) and
+[`just`](https://github.com/casey/just); for using it you need neither. Either way the demo
+needs nothing else: no producer, no credentials, no maintainer, no container. That demo is
+[`examples/websocket.py`](examples/websocket.py), and this is its shape:
+
+```python
+import litelink
+import pyarrow as pa
+
+# A trade feed: durable the moment it arrives, queryable a moment later.
+schema = pa.schema([
+    pa.field("trade_id", pa.int64()),
+    pa.field("event_ts", pa.int64()),    # microseconds, as the exchange sends them
+    pa.field("price", pa.float64()),
+    pa.field("amount", pa.float64()),
+    pa.field("side", pa.int64()),        # 0 buy, 1 sell
+])
+
+# new() takes the shape, fixed at creation. open() takes none of it — schema,
+# sort order, config and archive all come from the log itself.
+log = litelink.new("data", "trades", schema=schema, sort_by=("event_ts",))
+log.append({"trade_id": 624438572, "event_ts": 1787772776240000,
+            "price": 78501.62, "amount": 0.0076, "side": 0})  # durable on return
+
+# extend() commits the whole group in ONE transaction — one fsync for the batch,
+# not one per row. That call size is the write throughput lever, and a call-site
+# choice: no LogConfig setting tunes it.
+log.extend(group_of_rows)                          # append(row) is extend([row])
+
+log = litelink.open("data", "trades")
+reader = litelink.open("data", "trades", read_only=True)         # alongside a live writer, no write surface
+
+recent = log.scan(where="event_ts > 1787772776000000").read_all()
+log.maintain()                                     # compact, evict, expire
+```
+
+That is the whole API surface for local capture. Everything below is optional, and every
+public call is in [`docs/API.md`](docs/API.md) — one page, forty-one of them, and most
+deployments use six.
+
+## More demos
+
+A synthetic feed you can drive as hard as you like, with one process per storage role:
+
+```bash
+just demo-capture      # append continuously — the hot path, and nothing else
+just demo-maintain     # in another terminal: seal, compact, evict, expire
+just demo-tail         # in a third: watch where the rows are
+```
+
+To add the archive tier, against a local S3-compatible store or a real bucket:
+
+```bash
+just rustfs            # object storage in one container
+just demo-archive      # capture, with an archive configured
+just demo-maintain     # also pushes to it, and evicts what it has pushed
+
+cp .env.example .env   # or: set LITELINK_DEMO_ARCHIVE=s3://your-bucket/prefix
+```
+
+Credentials are not in that file unless you put them there — the library reads them from the
+environment through the ordinary AWS chain, so a profile, instance metadata or SSO all work
+untouched, and a log directory never carries a key with it.
+
+To survive losing the machine, ship the SQLite WAL alongside:
+
+```bash
+just demo-replicate    # generates litestream.yml from the log, runs it
+```
+
+`litelink.restore(root, name, archive=...)` then rebuilds the log on another box, reserving an
+offset window so nothing the dead machine served is reissued. Verified against a local
+S3-compatible store and against AWS. See [`examples/`](examples/).
+
+### Run the sidecar as its own process
+
+**litelink emits the config; your supervisor runs the binary.** `replication_config()` writes
+a `litestream.yml` describing which databases carry the log's state, and systemd, Kubernetes
+or anything else runs `litestream replicate -config …` beside the writer. litelink never
+starts it and never supervises it, and that is a design commitment rather than an omission:
+
+- **It keeps the network out of the write path.** A sidecar reads the WAL from outside the
+  process. If litelink owned the replicator, a stalled upload could push back on `append`,
+  and "durable when it returns, with no network in the write path" is the property the whole
+  design rests on.
+- **The lifecycles are the wrong way round otherwise.** A writer that owns its replicator
+  kills it at exactly the moment you need the last frames shipped. A separate process keeps
+  draining what is already on disk.
+- **§1 allows one writer per log.** Two `Log` handles each spawning a replicator would run
+  two litestream instances against one database, which litestream forbids.
+
+litelink *does* run litestream in one place — `restore`, which shells out to it once and waits.
+That is a batch call, not a supervised daemon, and it is why the binary ships in the wheel
+rather than being left to PATH: the alternative is discovering it is missing during a
+failover.
+
+Check the machine before you need it:
+
+```bash
+python -m litelink                             # extensions, litestream
+python -m litelink s3://bucket/prefix trades   # ...and the archive is reachable
+```
+
+Run that as the user and from the process manager that will own the log. A systemd user unit
+does not inherit a login shell's PATH, so `which litestream` succeeding in your terminal says
+nothing about the unit that will actually perform the restore.
+
+## Reading it from another machine
+
+The demos above are the *writer's* read. Everywhere else reads the archive, which is an
+ordinary Iceberg table publishing `version-hint.text` at every commit — so an engine pointed
+at the prefix resolves the current metadata itself, with no catalog service, no `archive.db`,
+no local root and no litelink install:
+
+```sql
+SELECT count(*), max(litelink_offset)
+FROM iceberg_scan('s3://bucket/prefix/litelink/trades',
+                  version_name_format = '%s%s.metadata.json');
+```
+
+`litelink_offset` is monotonic and never reused, so a reader keeps the highest one it has seen
+and asks for what came after — which is how you follow an archive that `sync` is publishing
+into. The extensions, the credential shapes, why `version_name_format` is not optional, and
+the polling pattern in full are in [`docs/API.md`](docs/API.md).
+
+That read is only as fresh as the last `sync`. When you have a WAL sidecar running
+(`wal_replication`), `litelink.follow` does better: it restores the writer's buffer alongside the
+archive and merges them, so the reader sees down to the replication lag instead.
+
+```python
+with litelink.follow("trades", archive="s3://bucket/prefix", s3=opts) as reader:
+    reader.coverage()   # Coverage(archive=(1, 1928), buffered=(1929, 2100), gap=None, ...)
+    reader.scan(where="side = 0", columns=["event_ts", "price"])
+```
+
+It never writes anything the primary shares and cannot append — a `LogHandle` has no write
+surface at all, rather than one that raises. It is a **snapshot, not a subscription**:
+refreshing means assembling another one, and the root it builds is a temporary directory
+removed on close. `coverage()` is how it stays honest about what it can and cannot serve.
+See [§3b](docs/SPEC.md) for why a gap in that report is not necessarily loss.
+
+## How it works
+
+- **Iceberg is used, not reimplemented.** Manifests, per-file column statistics, schema with
+  field IDs, and atomic snapshot commits all come from it.
+- **The library owns exactly one column**, `litelink_offset` — monotonic, never reused. It is
+  the boundary mechanism between tiers. Everything else is the caller's schema.
+- **Sync is a watermark, not CDC.** There are no updates or deletes to replicate.
+- **Parts are sealed once and never rewritten.** Rewriting a growing partition costs ~144x
+  write amplification and buys nothing, because the local WAL already made the row durable.
+- **Read boundaries are derived from committed table state**, never from a stored flag — so
+  no seal window can double-count or drop.
+- **The seal cut is chosen by the appender**, in the transaction that crosses
+  `target_seal_size`, and queued. A sealer that falls behind therefore writes several
+  correctly-sized files rather than one oversized one.
+- **Sizing is two targets, not one.** A seal wants to be small, because the buffer is what a
+  hot read scans; a file wants to be large, because per-file overhead dominates scans and
+  uploads. Compaction bridges them, on local disk, at 8× the seal size by default.
+
+Read performance is the cost of reading Parquet, plus ~4 ms of fixed overhead. The numbers
+behind all of this are [`docs/SPEC.md`](docs/SPEC.md) §7 and §12; how the pieces run is
+[`docs/RUNTIME.md`](docs/RUNTIME.md); `just bench` is the same measurement on your hardware.
+
+## What it is not
+
+An OLTP or key-value store. A point lookup is ~1,600x slower than an indexed row store, and no
+configuration closes that gap. It is a **local, in-process, real-time analytics store**: data
+is durable at commit and queryable immediately, so freshness is sub-second *with* durability —
+but "real-time" means fresh, not point-lookup fast.
+
+Nor is it an unbounded local archive. **Keeping everything on one machine — `archive=None`
+with no `local_retention` — degrades as the table grows.** A seal's cost tracks what the
+table's metadata holds, and a residue grows with the file count: compaction never revisits a
+file already at the target size, and only eviction removes one. With no retention set nothing
+evicts, and the seal is on the write path, so the cost lands on appends. Configure a
+retention, or an archive to evict into — [`docs/SPEC.md`](docs/SPEC.md) §13.7.
+
+## Not implemented yet
+
+**Schema evolution** ([`docs/SPEC.md`](docs/SPEC.md) §9) and **blob fields** (§15) are
+specified and unbuilt, and are what the code lacks against its own design. `add_column`,
+`rename_column` and `drop_column` exist and raise; `binary` columns are refused outright,
+because §15 has large payloads bypass the buffer rather than travel through it.
+
+Still open: payload encoding, local-disk backpressure, bulk ingest, and extension provisioning
+for embedders. All four in [`docs/SPEC.md`](docs/SPEC.md) §13.
+
+## Documentation
+
+- [`docs/API.md`](docs/API.md) — every public call, on one page
+- [`docs/SPEC.md`](docs/SPEC.md) — the design, and in places still ahead of the code
+- [`docs/RUNTIME.md`](docs/RUNTIME.md) — writer and maintainer, threads, processes, what crosses between them
+- [`examples/`](examples/) — the websocket capture, and the synthetic feed with one process per role
+- [`benchmarks/`](benchmarks/) — the harness, including what litelink costs over raw SQLite
+- [`CONTRIBUTING.md`](CONTRIBUTING.md) — setup, the gates, and what a good PR here looks like
+- [`SECURITY.md`](SECURITY.md) — what to report privately, and what is a known limit instead
+
+## Development
+
+```bash
+just bootstrap          # uv sync + git hooks + DuckDB extensions
+just check              # lint + format-check + typecheck + tests, same as CI
+just --list             # the rest
+```
+
+`just bootstrap` provisions the `iceberg`, `avro` and `httpfs` DuckDB extensions, which are
+downloaded rather than bundled — see [`docs/SPEC.md`](docs/SPEC.md) §7. Tooling is uv + ruff +
+[ty](https://github.com/astral-sh/ty) + pytest. Commits follow
+[Conventional Commits](https://www.conventionalcommits.org), enforced by a `commit-msg` hook;
+[`CONTRIBUTING.md`](CONTRIBUTING.md) has the types, scopes and style gates.
+
+## License
+
+Apache License 2.0 — see [LICENSE](LICENSE) and [NOTICE](NOTICE).