mlschema 0.2.1__tar.gz → 0.3.0__tar.gz

{mlschema-0.2.1 → mlschema-0.3.0}/.gitignore

@@ -177,10 +177,23 @@ cython_debug/
 # Custom
 .vscode/
 .history/
-uv.lock
 
 # Graphify
 graphify-out/
 
 # Codex
 .codex/
+
+# Product sites
+!/landing/src/lib/
+!/landing/src/lib/**
+node_modules/
+.astro/
+.wrangler/
+playwright-report/
+test-results/
+
+# Local browser verification
+.playwright-cli/
+output/playwright/
+.vite/

{mlschema-0.2.1 → mlschema-0.3.0}/AUTHORS.md

@@ -1,6 +1,6 @@
 # Authors and Contributors
 
-MLSchema is made possible by the contributions of the following individuals and organizations.
+MLschema is made possible by the contributions of the following individuals and organizations.
 
 ## Core Maintainers
 
@@ -35,7 +35,7 @@ See [CONTRIBUTING.md](CONTRIBUTING.md) for contribution guidelines.
 
 ### Dependencies
 
-MLSchema is built on the shoulders of giants. We are grateful to the maintainers and contributors of:
+MLschema is built on the shoulders of giants. We are grateful to the maintainers and contributors of:
 
 - **pandas** - The foundational data analysis library
 - **Pydantic** - Data validation using Python type annotations
@@ -44,7 +44,7 @@ For a complete list of dependencies and their licenses, see [THIRD_PARTY_LICENSE
 
 ### Inspiration
 
-MLSchema draws inspiration from:
+MLschema draws inspiration from:
 
 - **JSON Schema** specification
 - **FastAPI** for its excellent use of Pydantic

mlschema-0.3.0/CONTRIBUTING.md

@@ -0,0 +1,80 @@
+# Contributing to MLschema
+
+Use Python 3.14 and uv for the library. Websites use Vite+ (`vp`), backed by pnpm 12.
+Supported Node.js versions: 22.18+, 24.11+, or 26+ (the corresponding major versions).
+
+## Library
+
+```bash
+uv sync --locked --group dev
+uv run pre-commit install
+uv run ruff check .
+uv run ruff format --check .
+uv run pyright
+uv run pytest
+uv build
+```
+
+Start with the tests for the behavior you change, then run the full suite. Test both
+success and error paths. Keep each source file at or below 300 lines. Update the
+affected documentation when contracts or known limitations change; include actual
+verification and any remaining blockers in the change description.
+
+The public API starts at `mlschema.infer_schema`. Builders inspect pandas columns;
+strict field models validate their output. Custom kinds pair a `BaseField` subclass
+with a builder through `kind(model=..., infer=...)`. Keep rules in their owning
+module so callers share them.
+
+Series columns use local ids and scalar-kind validators. Only the parent maps to
+a backend feature. `with_positional_mappings` converts named targets when a consumer
+needs input positions; it preserves the original field list.
+
+## Websites
+
+```bash
+vp install --frozen-lockfile
+vp run dev:landing
+vp run dev:docs
+vp run dev:playground
+vp run build
+```
+
+Landing, Astro Starlight documentation and the Python playground live in separate
+workspace packages. Use the eshkra product identity and original assets under
+`brand/`. Verify narrow/wide screens and both themes when changing presentation.
+
+Local URLs: landing `http://127.0.0.1:4330`, docs `http://127.0.0.1:4331`,
+playground `http://127.0.0.1:10011`. Start each dev task in its own terminal.
+Install the [Vite+ CLI](https://viteplus.dev/guide/) first, or bootstrap with
+`pnpm install --frozen-lockfile` and use `pnpm exec vp` for these commands.
+
+## Release
+
+`pyproject.toml` owns the Python package version. Update it and the documentation
+changelog together. Breaking contracts need migration notes. Keep `uv.lock` and
+`pnpm-lock.yaml` synchronized with dependency changes.
+
+CI tests Python on Ubuntu and Windows, checks types/formatting, builds distributions,
+checks their PyPI description and builds all sites. A successful push to `main`
+publishes a version absent from PyPI and creates its GitHub release. The verified
+site artifacts deploy to Cloudflare Workers from `main`.
+
+Repository secrets:
+
+- `PYPI_API_TOKEN`: scoped to the MLschema PyPI project.
+- `CLOUDFLARE_API_TOKEN`: permission to deploy Workers and their custom domains.
+- `CLOUDFLARE_ACCOUNT_ID`: the account containing the eshkra zone.
+
+Domains are declared in each site's `wrangler.jsonc`. Docs, landing and playground
+use `docs.mlschema.eshkra.dev`, `mlschema.eshkra.dev` and
+`playground.mlschema.eshkra.dev`. A manual CI run on `main` can redeploy sites.
+
+Publish MLschema 0.3.0 before updating downstream registry locks such as mlbundle's.
+To verify the coordinated change before publication, install the local package
+into the consumer environment with `uv pip install --python .venv --no-deps -e ../mlschema`
+and run its checks with `uv run --no-sync`. This is a local test setup; refresh the
+consumer's ordinary registry lock with `uv sync --all-extras` after publication.
+
+Open focused pull requests describing the behavior change and actual verification.
+Use Conventional Commit titles. Report security issues through the process in
+[SECURITY.md](https://github.com/UlloaSP/mlschema/blob/main/SECURITY.md).

mlschema-0.3.0/PKG-INFO

@@ -0,0 +1,208 @@
+Metadata-Version: 2.5
+Name: mlschema
+Version: 0.3.0
+Summary: Infer validated, frontend-ready field schemas from pandas DataFrames.
+Project-URL: Homepage, https://mlschema.eshkra.dev/
+Project-URL: Documentation, https://docs.mlschema.eshkra.dev/
+Project-URL: Playground, https://playground.mlschema.eshkra.dev/
+Project-URL: Repository, https://github.com/UlloaSP/mlschema
+Project-URL: Bug Tracker, https://github.com/UlloaSP/mlschema/issues
+Project-URL: Source Code, https://github.com/UlloaSP/mlschema
+Project-URL: Changelog, https://docs.mlschema.eshkra.dev/changelog/
+Project-URL: Discussions, https://github.com/UlloaSP/mlschema/discussions
+Project-URL: Funding, https://github.com/sponsors/UlloaSP
+Project-URL: Release Notes, https://github.com/UlloaSP/mlschema/releases
+Author-email: Pablo Ulloa Santín <pablo.ulloa.santin@udc.es>
+Maintainer-email: Pablo Ulloa Santín <pablo.ulloa.santin@udc.es>
+License: MIT
+License-File: AUTHORS.md
+License-File: LICENSE
+Keywords: data-contracts,data-science,data-validation,dataframe,field-schema,form-generation,frontend-forms,frontend-integration,json,json-schema,machine-learning,mlops,pandas,pydantic,schema-inference,validation
+Classifier: Development Status :: 4 - Beta
+Classifier: Framework :: Pydantic
+Classifier: Framework :: Pydantic :: 2
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Information Technology
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Natural Language :: English
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Scientific/Engineering :: Information Analysis
+Classifier: Topic :: Software Development :: Code Generators
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Utilities
+Classifier: Typing :: Typed
+Requires-Python: <3.15,>=3.14
+Requires-Dist: pandas<4.0.0,>=3.0.3
+Requires-Dist: pydantic<3.0.0,>=2.13.4
+Description-Content-Type: text/markdown
+
+<p align="center">
+  <img src="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/mlschema.png" alt="" height="112" />
+</p>
+
+<h1 align="center">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/wordmark-dark.svg" />
+    <img src="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/wordmark-light.svg" alt="MLschema" height="52" />
+  </picture>
+</h1>
+
+<p align="center">From pandas DataFrames to validated field schemas.</p>
+
+<p align="center">
+  <a href="https://mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/website-303039?style=flat&amp;logo=astro&amp;logoColor=ff6a1a" alt="Website" /></a>
+  <a href="https://docs.mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/docs-303039?style=flat&amp;logo=readthedocs&amp;logoColor=ff6a1a" alt="Documentation" /></a>
+  <a href="https://playground.mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/playground-303039?style=flat&amp;logo=python&amp;logoColor=ff6a1a" alt="Playground" /></a>
+  <a href="https://pypi.org/project/mlschema/"><img src="https://img.shields.io/pypi/v/mlschema?style=flat&amp;logo=pypi&amp;logoColor=white&amp;labelColor=303039&amp;color=b73d00" alt="PyPI version" /></a>
+  <a href="https://github.com/UlloaSP/mlschema"><img src="https://img.shields.io/badge/source-303039?style=flat&amp;logo=github&amp;logoColor=white" alt="Source on GitHub" /></a>
+  <a href="https://github.com/UlloaSP/mlschema/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/UlloaSP/mlschema/ci.yml?branch=main&amp;style=flat&amp;label=CI&amp;labelColor=303039" alt="CI status" /></a>
+  <a href="https://github.com/UlloaSP/mlschema/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-MIT-b73d00?style=flat&amp;labelColor=303039" alt="MIT license" /></a>
+  <a href="https://eshkra.dev"><img src="https://img.shields.io/badge/part_of-eshkra-b73d00?style=flat&amp;labelColor=303039" alt="Part of eshkra" /></a>
+</p>
+
+MLschema is a Python library for deriving validated field schemas from pandas DataFrames. It reads column names, dtypes, categories and nullability, then validates the inferred fields with Pydantic and returns a JSON-compatible list.
+
+Use the result with [MLform](https://mlform.eshkra.dev/), include it in [MLbundle](https://github.com/UlloaSP/mlbundle), or consume it in your own frontend. Start with built-in inference, refine individual fields, or add domain-specific builders and kinds.
+
+## Why infer a field schema?
+
+The data pipeline already knows whether an input is a number, a date or a category. Writing those definitions again in a frontend creates another place for the model and its inputs to drift apart.
+
+MLschema turns that information into a starting contract with explicit backend mappings. Your application supplies product-specific labels, bounds and defaults; the same validation path checks both inferred and customized fields. Inference runs locally and the resulting list is ordinary data, independent of a rendering framework.
+
+## Installation
+
+> MLschema is pre-1.0. Public contracts may change between minor releases. Requires Python 3.14 (`>=3.14,<3.15`); supported pandas 3 and Pydantic 2 versions are installed as dependencies.
+
+```bash
+uv add mlschema
+```
+
+Or install from PyPI with `pip install mlschema`.
+
+See the [release and upgrade notes](https://docs.mlschema.eshkra.dev/changelog/) when updating an existing integration. Version 0.3 changes the series field contract.
+
+## Quick start
+
+Create a DataFrame with the dtypes your inputs should use:
+
+```python
+import json
+
+import pandas as pd
+from mlschema import infer_schema
+
+frame = pd.DataFrame(
+    {
+        "age": [28, 41],
+        "plan": pd.Categorical(["free", "pro"]),
+        "active": [True, False],
+    }
+)
+
+fields = infer_schema(frame)
+print(json.dumps(fields, indent=2))
+```
+
+The result is a field list, ready to serialize:
+
+```json
+[
+  {"kind": "number", "label": "age", "required": true, "mappedTo": "age", "step": 1.0},
+  {"kind": "category", "label": "plan", "required": true, "mappedTo": "plan", "options": ["free", "pro"]},
+  {"kind": "boolean", "label": "active", "required": true, "mappedTo": "active"}
+]
+```
+
+Pass it to MLform as the schema's `fields` property. Your application owns reports, transport and model execution.
+
+## Choose your customization
+
+| API | Use it to |
+| --- | --- |
+| `infer_schema(frame)` | Infer fields using the built-in kinds. |
+| `overrides={...}` | Refine a column's label, bounds or other supported properties before validation. |
+| `builders=[...]` | Supply inference logic for an existing kind, before the built-in builders. |
+| `kinds=[kind(model=..., infer=...)]` | Add a strict field model and the builder that produces it. |
+| `with_positional_mappings(fields, feature_names)` | Copy a field list and convert mappings into an explicit model input order. |
+
+For example, refine a label and bounds, then adapt the mappings for a model that consumes a feature vector:
+
+```python
+from mlschema import with_positional_mappings
+
+fields = infer_schema(
+    frame,
+    overrides={"age": {"label": "Age", "min": 0, "max": 120}},
+)
+positional = with_positional_mappings(fields, ["plan", "age", "active"])
+```
+
+The `age` field keeps its displayed label `Age`; its target changes from `"age"` to position `1` in the returned copy. The helper rewrites mappings, preserving field order, series columns and one-hot grouping. It leaves the original list unchanged.
+
+All application-facing functions and extension types come from `mlschema`. Public inference exceptions come from `mlschema.core`; see the [Python API](https://docs.mlschema.eshkra.dev/reference/api/).
+
+## What is included?
+
+- Inference for text, numbers, categories, booleans, dates and two-axis series.
+- One-hot groups from named binary columns such as `color__red` and `color__blue`.
+- Explicit backend mappings: names for named columns, positions for positional columns.
+- Series output with ordered `columns[]`, compatible with MLform.
+- Strict field validation, overrides, custom builders and custom kinds.
+- A typed, function-first API that runs entirely in your Python process.
+
+DataFrame dtypes matter: categorical options come from `category`, and date inference accepts specific datetime dtypes. Columns unclaimed by earlier builders reach the text fallback. See the [field-kind reference](https://docs.mlschema.eshkra.dev/schema/field-kinds/) for supported inputs and limits.
+
+## Some notes
+
+MLschema owns field inference, validation and explicit mappings. Your application owns rendering, report definitions, model training and execution. The output is a field list, not a JSON Schema document or a complete MLform schema.
+
+Empty DataFrames are rejected, and invalid builder output or overrides fail validation. Inference does not establish that a dataset is suitable for a model. The returned list can contain category values, so treat it as part of your data contract when sharing it.
+
+## Documentation
+
+Full documentation lives at [docs.mlschema.eshkra.dev](https://docs.mlschema.eshkra.dev/).
+
+- [Installation](https://docs.mlschema.eshkra.dev/getting-started/installation/)
+- [Quick start](https://docs.mlschema.eshkra.dev/getting-started/quick-start/)
+- [Field contract](https://docs.mlschema.eshkra.dev/schema/overview/)
+- [Backend mappings](https://docs.mlschema.eshkra.dev/schema/mappings/)
+- [Custom builders and kinds](https://docs.mlschema.eshkra.dev/guides/extensions/)
+- [Python API](https://docs.mlschema.eshkra.dev/reference/api/)
+
+Try the examples in the [Python playground](https://playground.mlschema.eshkra.dev/). It runs the repository's Python code in an experimental browser environment; its dependency limits are documented in the [playground guide](https://github.com/UlloaSP/mlschema/blob/main/playground/README.md).
+
+## Contributing
+
+Python development uses [uv](https://docs.astral.sh/uv/), Ruff, Pyright and pytest:
+
+```bash
+uv sync --locked --group dev
+uv run ruff check .
+uv run ruff format --check .
+uv run pyright
+uv run pytest
+uv build
+```
+
+The landing page, documentation and playground use [Vite+](https://viteplus.dev/guide/) (`vp`) with a pnpm workspace. This tooling is separate from the Python package:
+
+```bash
+vp install --frozen-lockfile
+vp run build
+```
+
+Read [CONTRIBUTING.md](https://github.com/UlloaSP/mlschema/blob/main/CONTRIBUTING.md) for local development and release setup.
+
+Run `vp run dev:playground` to serve the local Python playground at `http://127.0.0.1:10011`. The [playground guide](https://github.com/UlloaSP/mlschema/blob/main/playground/README.md) covers examples and verification commands.
+
+## License
+
+MIT

mlschema-0.3.0/README.md

@@ -0,0 +1,163 @@
+<p align="center">
+  <img src="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/mlschema.png" alt="" height="112" />
+</p>
+
+<h1 align="center">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/wordmark-dark.svg" />
+    <img src="https://raw.githubusercontent.com/UlloaSP/mlschema/main/brand/wordmark-light.svg" alt="MLschema" height="52" />
+  </picture>
+</h1>
+
+<p align="center">From pandas DataFrames to validated field schemas.</p>
+
+<p align="center">
+  <a href="https://mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/website-303039?style=flat&amp;logo=astro&amp;logoColor=ff6a1a" alt="Website" /></a>
+  <a href="https://docs.mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/docs-303039?style=flat&amp;logo=readthedocs&amp;logoColor=ff6a1a" alt="Documentation" /></a>
+  <a href="https://playground.mlschema.eshkra.dev/"><img src="https://img.shields.io/badge/playground-303039?style=flat&amp;logo=python&amp;logoColor=ff6a1a" alt="Playground" /></a>
+  <a href="https://pypi.org/project/mlschema/"><img src="https://img.shields.io/pypi/v/mlschema?style=flat&amp;logo=pypi&amp;logoColor=white&amp;labelColor=303039&amp;color=b73d00" alt="PyPI version" /></a>
+  <a href="https://github.com/UlloaSP/mlschema"><img src="https://img.shields.io/badge/source-303039?style=flat&amp;logo=github&amp;logoColor=white" alt="Source on GitHub" /></a>
+  <a href="https://github.com/UlloaSP/mlschema/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/UlloaSP/mlschema/ci.yml?branch=main&amp;style=flat&amp;label=CI&amp;labelColor=303039" alt="CI status" /></a>
+  <a href="https://github.com/UlloaSP/mlschema/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-MIT-b73d00?style=flat&amp;labelColor=303039" alt="MIT license" /></a>
+  <a href="https://eshkra.dev"><img src="https://img.shields.io/badge/part_of-eshkra-b73d00?style=flat&amp;labelColor=303039" alt="Part of eshkra" /></a>
+</p>
+
+MLschema is a Python library for deriving validated field schemas from pandas DataFrames. It reads column names, dtypes, categories and nullability, then validates the inferred fields with Pydantic and returns a JSON-compatible list.
+
+Use the result with [MLform](https://mlform.eshkra.dev/), include it in [MLbundle](https://github.com/UlloaSP/mlbundle), or consume it in your own frontend. Start with built-in inference, refine individual fields, or add domain-specific builders and kinds.
+
+## Why infer a field schema?
+
+The data pipeline already knows whether an input is a number, a date or a category. Writing those definitions again in a frontend creates another place for the model and its inputs to drift apart.
+
+MLschema turns that information into a starting contract with explicit backend mappings. Your application supplies product-specific labels, bounds and defaults; the same validation path checks both inferred and customized fields. Inference runs locally and the resulting list is ordinary data, independent of a rendering framework.
+
+## Installation
+
+> MLschema is pre-1.0. Public contracts may change between minor releases. Requires Python 3.14 (`>=3.14,<3.15`); supported pandas 3 and Pydantic 2 versions are installed as dependencies.
+
+```bash
+uv add mlschema
+```
+
+Or install from PyPI with `pip install mlschema`.
+
+See the [release and upgrade notes](https://docs.mlschema.eshkra.dev/changelog/) when updating an existing integration. Version 0.3 changes the series field contract.
+
+## Quick start
+
+Create a DataFrame with the dtypes your inputs should use:
+
+```python
+import json
+
+import pandas as pd
+from mlschema import infer_schema
+
+frame = pd.DataFrame(
+    {
+        "age": [28, 41],
+        "plan": pd.Categorical(["free", "pro"]),
+        "active": [True, False],
+    }
+)
+
+fields = infer_schema(frame)
+print(json.dumps(fields, indent=2))
+```
+
+The result is a field list, ready to serialize:
+
+```json
+[
+  {"kind": "number", "label": "age", "required": true, "mappedTo": "age", "step": 1.0},
+  {"kind": "category", "label": "plan", "required": true, "mappedTo": "plan", "options": ["free", "pro"]},
+  {"kind": "boolean", "label": "active", "required": true, "mappedTo": "active"}
+]
+```
+
+Pass it to MLform as the schema's `fields` property. Your application owns reports, transport and model execution.
+
+## Choose your customization
+
+| API | Use it to |
+| --- | --- |
+| `infer_schema(frame)` | Infer fields using the built-in kinds. |
+| `overrides={...}` | Refine a column's label, bounds or other supported properties before validation. |
+| `builders=[...]` | Supply inference logic for an existing kind, before the built-in builders. |
+| `kinds=[kind(model=..., infer=...)]` | Add a strict field model and the builder that produces it. |
+| `with_positional_mappings(fields, feature_names)` | Copy a field list and convert mappings into an explicit model input order. |
+
+For example, refine a label and bounds, then adapt the mappings for a model that consumes a feature vector:
+
+```python
+from mlschema import with_positional_mappings
+
+fields = infer_schema(
+    frame,
+    overrides={"age": {"label": "Age", "min": 0, "max": 120}},
+)
+positional = with_positional_mappings(fields, ["plan", "age", "active"])
+```
+
+The `age` field keeps its displayed label `Age`; its target changes from `"age"` to position `1` in the returned copy. The helper rewrites mappings, preserving field order, series columns and one-hot grouping. It leaves the original list unchanged.
+
+All application-facing functions and extension types come from `mlschema`. Public inference exceptions come from `mlschema.core`; see the [Python API](https://docs.mlschema.eshkra.dev/reference/api/).
+
+## What is included?
+
+- Inference for text, numbers, categories, booleans, dates and two-axis series.
+- One-hot groups from named binary columns such as `color__red` and `color__blue`.
+- Explicit backend mappings: names for named columns, positions for positional columns.
+- Series output with ordered `columns[]`, compatible with MLform.
+- Strict field validation, overrides, custom builders and custom kinds.
+- A typed, function-first API that runs entirely in your Python process.
+
+DataFrame dtypes matter: categorical options come from `category`, and date inference accepts specific datetime dtypes. Columns unclaimed by earlier builders reach the text fallback. See the [field-kind reference](https://docs.mlschema.eshkra.dev/schema/field-kinds/) for supported inputs and limits.
+
+## Some notes
+
+MLschema owns field inference, validation and explicit mappings. Your application owns rendering, report definitions, model training and execution. The output is a field list, not a JSON Schema document or a complete MLform schema.
+
+Empty DataFrames are rejected, and invalid builder output or overrides fail validation. Inference does not establish that a dataset is suitable for a model. The returned list can contain category values, so treat it as part of your data contract when sharing it.
+
+## Documentation
+
+Full documentation lives at [docs.mlschema.eshkra.dev](https://docs.mlschema.eshkra.dev/).
+
+- [Installation](https://docs.mlschema.eshkra.dev/getting-started/installation/)
+- [Quick start](https://docs.mlschema.eshkra.dev/getting-started/quick-start/)
+- [Field contract](https://docs.mlschema.eshkra.dev/schema/overview/)
+- [Backend mappings](https://docs.mlschema.eshkra.dev/schema/mappings/)
+- [Custom builders and kinds](https://docs.mlschema.eshkra.dev/guides/extensions/)
+- [Python API](https://docs.mlschema.eshkra.dev/reference/api/)
+
+Try the examples in the [Python playground](https://playground.mlschema.eshkra.dev/). It runs the repository's Python code in an experimental browser environment; its dependency limits are documented in the [playground guide](https://github.com/UlloaSP/mlschema/blob/main/playground/README.md).
+
+## Contributing
+
+Python development uses [uv](https://docs.astral.sh/uv/), Ruff, Pyright and pytest:
+
+```bash
+uv sync --locked --group dev
+uv run ruff check .
+uv run ruff format --check .
+uv run pyright
+uv run pytest
+uv build
+```
+
+The landing page, documentation and playground use [Vite+](https://viteplus.dev/guide/) (`vp`) with a pnpm workspace. This tooling is separate from the Python package:
+
+```bash
+vp install --frozen-lockfile
+vp run build
+```
+
+Read [CONTRIBUTING.md](https://github.com/UlloaSP/mlschema/blob/main/CONTRIBUTING.md) for local development and release setup.
+
+Run `vp run dev:playground` to serve the local Python playground at `http://127.0.0.1:10011`. The [playground guide](https://github.com/UlloaSP/mlschema/blob/main/playground/README.md) covers examples and verification commands.
+
+## License
+
+MIT

{mlschema-0.2.1 → mlschema-0.3.0}/SECURITY.md

@@ -11,7 +11,7 @@ We release patches for security vulnerabilities for the following versions:
 
 ## Reporting a Vulnerability
 
-We take the security of MLSchema seriously. If you believe you have found a security vulnerability, please report it to us as described below.
+We take the security of MLschema seriously. If you believe you have found a security vulnerability, please report it to us as described below.
 
 ### Please Do NOT
 
@@ -39,7 +39,7 @@ We take the security of MLSchema seriously. If you believe you have found a secu
 
 ## Security Best Practices
 
-When using MLSchema, we recommend:
+When using MLschema, we recommend:
 
 ### Input Validation
 
@@ -83,7 +83,7 @@ df.columns = [col.replace(" ", "_").lower() for col in df.columns]
 
 ### Data Processing
 
-MLSchema processes pandas DataFrames to generate JSON schemas. It:
+MLschema processes pandas DataFrames to generate JSON schemas. It:
 
 - **Does not** modify the original DataFrame
 - **Does not** send data over the network
@@ -92,7 +92,7 @@ MLSchema processes pandas DataFrames to generate JSON schemas. It:
 
 ### Type Safety
 
-MLSchema uses Pydantic for validation, which provides:
+MLschema uses Pydantic for validation, which provides:
 
 - Type checking and validation
 - Protection against invalid data structures
@@ -100,7 +100,7 @@ MLSchema uses Pydantic for validation, which provides:
 
 ### Known Limitations
 
-- MLSchema assumes DataFrames come from trusted sources
+- MLschema assumes DataFrames come from trusted sources
 - Column names are used as-is in the output schema
 - Very large DataFrames may cause memory issues
 
@@ -128,11 +128,11 @@ Security updates are announced via:
 
 - GitHub Security Advisories: <https://github.com/UlloaSP/mlschema/security/advisories>
 - Release notes: <https://github.com/UlloaSP/mlschema/releases>
-- Changelog: [docs/changelog.md](docs/changelog.md)
+- Changelog: [release notes](https://docs.mlschema.eshkra.dev/changelog/)
 
 ## Dependency Security
 
-MLSchema depends on:
+MLschema depends on:
 
 ### Runtime Dependencies
 
@@ -179,7 +179,7 @@ When contributing code:
 
 ## Security-Related Configuration
 
-MLSchema has minimal configuration. Security-relevant settings:
+MLschema has minimal configuration. Security-relevant settings:
 
 ```python
 # Example: Safe usage pattern
@@ -200,7 +200,7 @@ schema = mls.build(dataframe)
 For security concerns:
 
 - **Email**: <pablo.ulloa.santin@udc.es>
-- **Subject**: `[SECURITY] MLSchema Vulnerability Report`
+- **Subject**: `[SECURITY] MLschema Vulnerability Report`
 
 For general questions:
 
@@ -209,14 +209,14 @@ For general questions:
 
 ## Attribution
 
-We appreciate responsible security researchers who help keep MLSchema safe. With your permission, we will:
+We appreciate responsible security researchers who help keep MLschema safe. With your permission, we will:
 
 - Credit you in the security advisory
 - Add your name to our hall of fame (if you wish)
 - Provide a reference for your responsible disclosure
 
-Thank you for helping keep MLSchema and its users safe! 🔒
+Thank you for helping keep MLschema and its users safe! 🔒
 
 ---
 
-**Last Updated**: June 1, 2026
+**Last Updated**: October 19, 2026