humemdb 0.1.0.dev1__tar.gz → 0.1.0.dev2__tar.gz

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: humemdb
-Version: 0.1.0.dev1
+Version: 0.1.0.dev2
 Summary: Embedded multi-model database orchestration over SQLite, DuckDB, Cypher, and exact and ANN vector search
 Project-URL: Homepage, https://github.com/humemai/humemdb
 Project-URL: Documentation, https://docs.humem.ai/humemdb/
@@ -104,6 +104,8 @@ Current behavior is intentionally explicit:
 
 ## Install
 
+HumemDB supports Python 3.10 and newer.
+
 Install from source:
 
 ```bash
@@ -135,6 +137,21 @@ uv sync
 `uv lock` updates the lockfile with exact resolved versions. `uv sync` installs that
 exact environment.
 
+## Libraries HumemDB relies on
+
+HumemDB is a pure Python orchestration layer, but it relies on a small set of core
+Python libraries and embedded engines:
+
+- `sqlite3` from the Python standard library for the canonical local write path.
+- `duckdb` for analytical reads over the SQLite-backed source-of-truth database.
+- `numpy` for the exact in-memory vector search baseline.
+- `sqlglot[c]` for the current PostgreSQL-like SQL translation layer.
+- `lancedb` for benchmark work and future indexed ANN paths.
+- `threadpoolctl` for thread-pool coordination around compute-heavy dependencies.
+
+Those dependencies are part of the public runtime story. HumemDB does not try to hide
+them behind a fake "single engine" narrative.
+
 ## 🧠 What is supported today
 
 ### SQL
@@ -191,7 +208,7 @@ More examples live in [examples/](examples/) and in the docs site.
 - Docs: [docs.humem.ai/humemdb](https://docs.humem.ai/humemdb/)
 - Repository: [github.com/humemai/humemdb](https://github.com/humemai/humemdb)
 - Issues: [github.com/humemai/humemdb/issues](https://github.com/humemai/humemdb/issues)
-- Internal phase tracking: [things-to-do.md](things-to-do.md)
+- Internal roadmap notes: [things-to-do.md](things-to-do.md)
 
 ## 📦 Packaging
 
@@ -201,9 +218,15 @@ side.
 
 ## 🗺️ Planning
 
-The detailed internal roadmap and phase tracking now live in `things-to-do.md` instead
-of this README.
+Detailed internal roadmap notes now live in `things-to-do.md` instead of this README.
 
 ## 📄 License
 
-MIT.
+HumemDB's own source code is licensed under MIT. See [LICENSE](LICENSE).
+
+Third-party dependencies keep their own licenses. Installing HumemDB may also install
+third-party Python packages and, in some cases, their native wheels. Those components
+are not relicensed under MIT just because HumemDB depends on them.
+
+For the concrete dependency set, see [pyproject.toml](pyproject.toml) and
+[uv.lock](uv.lock).

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/README.md

@@ -52,6 +52,8 @@ Current behavior is intentionally explicit:
 
 ## Install
 
+HumemDB supports Python 3.10 and newer.
+
 Install from source:
 
 ```bash
@@ -83,6 +85,21 @@ uv sync
 `uv lock` updates the lockfile with exact resolved versions. `uv sync` installs that
 exact environment.
 
+## Libraries HumemDB relies on
+
+HumemDB is a pure Python orchestration layer, but it relies on a small set of core
+Python libraries and embedded engines:
+
+- `sqlite3` from the Python standard library for the canonical local write path.
+- `duckdb` for analytical reads over the SQLite-backed source-of-truth database.
+- `numpy` for the exact in-memory vector search baseline.
+- `sqlglot[c]` for the current PostgreSQL-like SQL translation layer.
+- `lancedb` for benchmark work and future indexed ANN paths.
+- `threadpoolctl` for thread-pool coordination around compute-heavy dependencies.
+
+Those dependencies are part of the public runtime story. HumemDB does not try to hide
+them behind a fake "single engine" narrative.
+
 ## 🧠 What is supported today
 
 ### SQL
@@ -139,7 +156,7 @@ More examples live in [examples/](examples/) and in the docs site.
 - Docs: [docs.humem.ai/humemdb](https://docs.humem.ai/humemdb/)
 - Repository: [github.com/humemai/humemdb](https://github.com/humemai/humemdb)
 - Issues: [github.com/humemai/humemdb/issues](https://github.com/humemai/humemdb/issues)
-- Internal phase tracking: [things-to-do.md](things-to-do.md)
+- Internal roadmap notes: [things-to-do.md](things-to-do.md)
 
 ## 📦 Packaging
 
@@ -149,9 +166,15 @@ side.
 
 ## 🗺️ Planning
 
-The detailed internal roadmap and phase tracking now live in `things-to-do.md` instead
-of this README.
+Detailed internal roadmap notes now live in `things-to-do.md` instead of this README.
 
 ## 📄 License
 
-MIT.
+HumemDB's own source code is licensed under MIT. See [LICENSE](LICENSE).
+
+Third-party dependencies keep their own licenses. Installing HumemDB may also install
+third-party Python packages and, in some cases, their native wheels. Those components
+are not relicensed under MIT just because HumemDB depends on them.
+
+For the concrete dependency set, see [pyproject.toml](pyproject.toml) and
+[uv.lock](uv.lock).

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/docs/development/documentation.md

@@ -22,7 +22,7 @@ The docs site is published into the shared `humemai-docs` repository under the
 
 `mike` manages versioned documentation aliases such as `latest`.
 
-## Writing rules for this phase
+## Writing rules
 
 - keep the docs aligned with the actual `v0` runtime surface
 - prefer clear boundaries over aspirational claims

humemdb-0.1.0.dev2/docs/development/licensing.md

@@ -0,0 +1,44 @@
+# Licensing
+
+## HumemDB license
+
+HumemDB's own source code is licensed under MIT.
+
+- Repository license file:
+  [LICENSE]({{ config.repo_url }}/blob/{{ config.extra.version_tag }}/LICENSE)
+- Package metadata:
+  [pyproject.toml]({{ config.repo_url }}/blob/{{ config.extra.version_tag }}/pyproject.toml)
+
+That MIT grant applies to HumemDB's code and documentation in this repository. It does
+not automatically relicense third-party dependencies.
+
+## Runtime libraries HumemDB relies on
+
+The current runtime depends on a small set of explicit Python libraries and embedded
+engines:
+
+- `sqlite3` from the Python standard library for the canonical local write path
+- `duckdb` for analytical reads
+- `numpy` for the exact in-memory vector search baseline
+- `sqlglot[c]` for the SQL translation layer
+- `lancedb` for benchmark work and future indexed ANN paths
+- `threadpoolctl` for thread-pool coordination
+
+HumemDB intentionally presents those libraries as part of the implementation story
+instead of pretending everything comes from one hidden backend.
+
+## Third-party licenses
+
+Third-party dependencies keep their own licenses and distribution terms.
+
+- Installing HumemDB may download third-party wheels or source distributions.
+- Some of those dependencies may include native code or platform-specific wheels.
+- Those components remain under their upstream licenses.
+
+For the concrete dependency set used by this project, see:
+
+- [pyproject.toml]({{ config.repo_url }}/blob/{{ config.extra.version_tag }}/pyproject.toml)
+- [uv.lock]({{ config.repo_url }}/blob/{{ config.extra.version_tag }}/uv.lock)
+
+If you need a formal license review for a deployment, review the resolved dependency
+set from the lockfile rather than only the top-level MIT license.

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/docs/development/release.md

@@ -1,6 +1,6 @@
 # Release Workflow
 
-Phase 6 hardens the public release path around two outputs:
+HumemDB's public release path currently centers on two outputs:
 
 - the PyPI package
 - the versioned docs site

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/docs/getting-started/installation.md

@@ -2,7 +2,7 @@
 
 ## Requirements
 
-- Python 3.12 or newer.
+- Python 3.10 or newer.
 - A normal local Python environment managed with `uv`.
 
 ## Install from source
@@ -43,11 +43,19 @@ uv sync --group docs
 
 ## Installed runtime dependencies
 
-- `duckdb`
-- `lancedb`
-- `numpy`
-- `sqlglot[c]`
-- `threadpoolctl`
+- `sqlite3` from the Python standard library for the canonical local write path.
+- `duckdb` for analytical reads.
+- `numpy` for the exact vector search baseline.
+- `sqlglot[c]` for PostgreSQL-like SQL translation.
+- `lancedb` for benchmark work and future indexed ANN paths.
+- `threadpoolctl` for thread-pool coordination.
 
 LanceDB is present for benchmark and later accelerated vector work, but the shipped
-default vector runtime today is still the exact SQLite plus NumPy path.
+default vector runtime today is still the exact SQLite plus NumPy path.
+
+## Licensing note
+
+HumemDB's own code is MIT-licensed, but installed dependencies keep their own licenses.
+See the project [LICENSE]({{ config.repo_url }}/blob/{{ config.extra.version_tag
+}}/LICENSE) for HumemDB itself and the lockfile plus package metadata for the concrete
+third-party dependency set.

{humemdb-0.1.0.dev1 → humemdb-0.1.0.dev2}/mkdocs.yml

@@ -96,4 +96,5 @@ nav:
       - Benchmarks: guide/benchmarks.md
   - Development:
       - Documentation: development/documentation.md
+      - Licensing: development/licensing.md
       - Release Workflow: development/release.md