json2vec 0.4.8__tar.gz → 0.4.10__tar.gz

{json2vec-0.4.8/src/json2vec.egg-info → json2vec-0.4.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: json2vec
-Version: 0.4.8
+Version: 0.4.10
 Summary: Schema-first PyTorch models for hierarchical / nested / sequence data structures
 License-Expression: Apache-2.0
 Requires-Python: >=3.12
@@ -28,13 +28,7 @@ Requires-Dist: uvicorn>=0.38.0; extra == "serving"
 Provides-Extra: text
 Requires-Dist: transformers>=4.55.0; extra == "text"
 Provides-Extra: docs
-Requires-Dist: fastapi>=0.124.0; extra == "docs"
-Requires-Dist: mkdocs-material>=9.6; extra == "docs"
-Requires-Dist: mkdocs-jupyter>=0.26.3; extra == "docs"
-Requires-Dist: mkdocstrings[python]>=0.27; extra == "docs"
-Requires-Dist: orjson>=3.10.0; extra == "docs"
-Requires-Dist: pydantic-settings>=2.10.1; extra == "docs"
-Requires-Dist: uvicorn>=0.38.0; extra == "docs"
+Requires-Dist: marimo>=0.23.8; extra == "docs"
 Dynamic: license-file
 
 <h1 align="center"><code>json2vec</code></h1>
@@ -42,7 +36,7 @@ Dynamic: license-file
 <p align="center">
   <img alt="Python 3.12+" src="https://img.shields.io/badge/python-3.12%2B-3776AB?logo=python&amp;logoColor=white" />
   <a href="LICENSE"><img alt="Apache-2.0 license" src="https://img.shields.io/badge/license-Apache--2.0-2E8B57" /></a>
-  <a href="https://json2vec.github.io/json2vec/"><img alt="Documentation" src="https://img.shields.io/badge/docs-MkDocs-526CFE?logo=materialformkdocs&amp;logoColor=white" /></a>
+  <a href="https://json2vec.github.io/json2vec/"><img alt="Documentation" src="https://img.shields.io/badge/docs-Quarto-39729E?logo=quarto&amp;logoColor=white" /></a>
   <!-- discord-invite:start -->
   <a href="https://discord.gg/DVyZUkvTFA"><img alt="Discord channel invite" src="https://img.shields.io/badge/discord-join%20the%20channel-5865F2?logo=discord&amp;logoColor=white" /></a>
   <!-- discord-invite:end -->
@@ -63,7 +57,8 @@ A `json2vec` schema is both a data contract and an architecture blueprint.
 
 - Leaf fields such as `Number`, `Category`, `Set`, `Entity`, `Text`, and
   `Vector` become datatype-specific tensorfields.
-- `Array` nodes become local context encoders for repeated child objects.
+- `Branch` nodes define shared contexts for child fields, with optional local
+  attention and pooling before the representation flows upward.
 - Targets, masks, pruning, and embeddings are configured on the same schema
   tree.
 - Prediction output is keyed by schema address, so decoded values and
@@ -76,24 +71,23 @@ inference, and serving.
 ## A Model From A Nested Record
 
 ```python
-import json2vec as j2v
-
-model = j2v.Model.from_schema(
-    j2v.Category("customer_tier", max_vocab_size=16),
-    j2v.Array(
-        j2v.Category("sku", max_vocab_size=2048),
-        j2v.Number("quantity"),
-        j2v.Number("price"),
-        name="line_items",
-        max_length=32,
-        embed=True,
-    ),
-    j2v.Category("returned", target=True, max_vocab_size=2),
+import json2vec as jv
+
+model = jv.Model.from_tree(
     name="order",
     d_model=64,
     n_layers=2,
     n_heads=4,
     embed=True,
+    customer_tier=jv.Category(size=16),
+    line_items=jv.Branch(
+        length=32,
+        embed=True,
+        sku=jv.Category(size=2048),
+        quantity=jv.Number,
+        price=jv.Number,
+    ),
+    returned=jv.Category(target=True, size=2),
 )
 ```
 
@@ -116,8 +110,8 @@ to emit embeddings at configured addresses.
 
 ## Train With Lightning
 
-`j2v.Model` is a LightningModule. `j2v.PolarsDataModule` and
-`j2v.StreamingDataModule` are LightningDataModule implementations. The schema
+`jv.Model` is a LightningModule. `jv.PolarsDataModule` and
+`jv.StreamingDataModule` are LightningDataModule implementations. The schema
 defines the model tree, typed losses, prediction outputs, and embeddings;
 Lightning runs `fit`, `validate`, `test`, and `predict`.
 
@@ -126,23 +120,23 @@ import lightning.pytorch as lit
 import polars as pl
 import torch
 
-import json2vec as j2v
+import json2vec as jv
 
 records = pl.read_ndjson("docs/data/iris.jsonl").head(36)
 
-model = j2v.Model.from_schema(
-    j2v.Number("sepal_length"),
-    j2v.Number("petal_length"),
-    j2v.Category("species", target=True, max_vocab_size=4, topk=[2]),
+model = jv.Model.from_tree(
     d_model=16,
     n_layers=1,
     n_heads=4,
     batch_size=8,
     embed=True,
     optimizer=lambda module: torch.optim.AdamW(module.parameters(), lr=1e-2),
+    sepal_length=jv.Number,
+    petal_length=jv.Number,
+    species=jv.Category(target=True, size=4, topk=[2]),
 )
 
-datamodule = j2v.PolarsDataModule(
+datamodule = jv.PolarsDataModule(
     model=model,
     train=records,
     validate=records,
@@ -170,7 +164,7 @@ trainer.fit(model=model, datamodule=datamodule)
 For larger jobs, the same model can run through normal Lightning callbacks,
 checkpointing, precision settings, device placement, and distributed
 strategies. See
-[Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning/).
+[Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning.html).
 
 ## Predict And Embed
 
@@ -179,8 +173,8 @@ For small interactive batches, call `model.predict(...)` with raw dictionaries.
 ```python
 predictions = model.predict(records.to_dicts()[:3])
 
-species = predictions[j2v.Address("record", "species")]
-record = predictions[j2v.Address("record")]
+species = predictions[jv.Address("record", "species")]
+record = predictions[jv.Address("record")]
 
 print(species["content"]["value"])
 print(species["content"]["probability"])
@@ -188,10 +182,10 @@ print(record["embedding"])
 ```
 
 For larger offline jobs, configure a `predict` split on a data module and attach
-`j2v.Writer` to Lightning's prediction loop.
+`jv.Writer` to Lightning's prediction loop.
 
 ```python
-writer = j2v.Writer("predictions")
+writer = jv.Writer("predictions")
 
 trainer = lit.Trainer(
     accelerator="cpu",
@@ -199,7 +193,7 @@ trainer = lit.Trainer(
     logger=False,
 )
 
-predict_datamodule = j2v.PolarsDataModule(
+predict_datamodule = jv.PolarsDataModule(
     model=model,
     predict=records.drop("species"),
     num_workers=0,
@@ -213,8 +207,8 @@ trainer.predict(model=model, datamodule=predict_datamodule)
 `Writer` creates rank-partitioned Parquet files such as
 `predictions/rank-0.parquet`. Use a postprocessor when downstream systems need
 flat columns, renamed addresses, redacted payloads, or fewer fields. See
-[Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference/)
-and [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors/).
+[Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference.html)
+and [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors.html).
 
 ## Learning Modes
 
@@ -254,7 +248,7 @@ Choose the data module by where the records live:
 arguments are compiled regular expressions matched against discovered file
 paths.
 
-See [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules/)
+See [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules.html)
 for split configuration, sharding, sampling, buffers, and preprocessors.
 
 ## What Makes This Different
@@ -265,7 +259,7 @@ for split configuration, sharding, sampling, buffers, and preprocessors.
   missing-state handling, masking, decoding, loss, metrics, and output writing.
 - **Unified training roles:** `target=True`, `p_prune`, and `p_mask` all use the
   same reconstruction path.
-- **Embedding trees:** embeddings can come from the root, arrays, or selected
+- **Embedding trees:** embeddings can come from the root, branches, or selected
   leaves.
 - **Schema evolution:** fields can be added, removed, updated, reset, or
   temporarily overridden after construction.
@@ -318,55 +312,53 @@ uv sync --extra docs
 
 The `text` extra installs Hugging Face `transformers`. The `serving` extra
 installs FastAPI-backed deployment dependencies. The `docs` extra installs the
-MkDocs toolchain.
+Python packages used by the Quarto docs.
 
 ## Documentation Map
 
 Start with:
 
-- [Getting Started](https://json2vec.github.io/json2vec/getting-started/)
-- [AI / Expert Quickstart](https://json2vec.github.io/json2vec/ai-quickstart/)
-- [Model Tree](https://json2vec.github.io/json2vec/core-concepts/model-tree/)
-- [Query Paths](https://json2vec.github.io/json2vec/core-concepts/querypaths/)
-- [Built-In Data Types](https://json2vec.github.io/json2vec/core-concepts/data-types/)
-- [Learning Modes & Embeddings](https://json2vec.github.io/json2vec/core-concepts/embeddings/)
-- [Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning/)
-- [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules/)
-- [Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference/)
-- [API Reference](https://json2vec.github.io/json2vec/reference/api/)
+- [Getting Started](https://json2vec.github.io/json2vec/getting-started.html)
+- [AI / Expert Quickstart](https://json2vec.github.io/json2vec/ai-quickstart.html)
+- [Model Tree](https://json2vec.github.io/json2vec/core-concepts/model-tree.html)
+- [Query Paths](https://json2vec.github.io/json2vec/core-concepts/querypaths.html)
+- [Built-In Data Types](https://json2vec.github.io/json2vec/core-concepts/data-types.html)
+- [Learning Modes & Embeddings](https://json2vec.github.io/json2vec/core-concepts/embeddings.html)
+- [Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning.html)
+- [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules.html)
+- [Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference.html)
 
 Tutorials and guides:
 
-- [Hello World](https://json2vec.github.io/json2vec/tutorials/hello-world/)
-- [Supervised Tabular Training](https://json2vec.github.io/json2vec/tutorials/supervised-tabular-training/)
-- [Masked Pretraining](https://json2vec.github.io/json2vec/tutorials/pretraining/)
-- [Nested Supervised Training](https://json2vec.github.io/json2vec/tutorials/nested-supervised-training/)
-- [Serving](https://json2vec.github.io/json2vec/tutorials/serving/)
-- [Preprocessors](https://json2vec.github.io/json2vec/guides/preprocessors/)
-- [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors/)
-- [Field Importance](https://json2vec.github.io/json2vec/guides/field-importance/)
-- [Field Stacking](https://json2vec.github.io/json2vec/guides/field-stacking/)
-- [Custom Data Types](https://json2vec.github.io/json2vec/data-types/tensorfields/)
-- [Device Tenure Case Study](https://json2vec.github.io/json2vec/case-studies/device-tenure/)
-- [Whitepaper](https://json2vec.github.io/json2vec/whitepaper.pdf)
+- [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors.html)
+- [Field Importance](https://json2vec.github.io/json2vec/guides/field-importance.html)
+- [Field Stacking](https://json2vec.github.io/json2vec/guides/field-stacking.html)
+- [Branch](https://json2vec.github.io/json2vec/data-types/branch.html)
+- [Number](https://json2vec.github.io/json2vec/data-types/number.html)
+- [Category](https://json2vec.github.io/json2vec/data-types/category.html)
+- [Set](https://json2vec.github.io/json2vec/data-types/set.html)
+- [Entity](https://json2vec.github.io/json2vec/data-types/entity.html)
+- [DateParts](https://json2vec.github.io/json2vec/data-types/dateparts.html)
+- [Vector](https://json2vec.github.io/json2vec/data-types/vector.html)
+- [Text](https://json2vec.github.io/json2vec/data-types/text.html)
+- [Device Tenure Case Study](https://json2vec.github.io/json2vec/case-studies/device-tenure.html)
 
 Build the docs locally with:
 
 ```bash
-uv run --extra docs mkdocs build --strict
+make render
 ```
 
 ## Repository Layout
 
 - `src/json2vec/architecture`: model assembly, attention, pooling, and routing
-- `src/json2vec/data`: dataset fetch/read/process/batch/encode pipeline
+- `src/json2vec/data`: dataset fetch/read/process/batch/encode pipeline and preprocessor exports
 - `src/json2vec/inference`: serving and prediction callbacks
 - `src/json2vec/logging`: runtime logging callbacks
-- `src/json2vec/preprocessors`: preprocessor registry
 - `src/json2vec/structs`: pydantic config models, enums, and tree nodes
 - `src/json2vec/tensorfields`: tensorfield plugin system and built-in fields
 - `tests/`: package test suite
-- `docs/`: tutorials, guides, diagrams, and whitepaper source
+- `docs/`: Quarto project, pages, guides, stylesheets, and sample data
 
 ## Development
 

{json2vec-0.4.8 → json2vec-0.4.10}/README.md

@@ -3,7 +3,7 @@
 <p align="center">
   <img alt="Python 3.12+" src="https://img.shields.io/badge/python-3.12%2B-3776AB?logo=python&amp;logoColor=white" />
   <a href="LICENSE"><img alt="Apache-2.0 license" src="https://img.shields.io/badge/license-Apache--2.0-2E8B57" /></a>
-  <a href="https://json2vec.github.io/json2vec/"><img alt="Documentation" src="https://img.shields.io/badge/docs-MkDocs-526CFE?logo=materialformkdocs&amp;logoColor=white" /></a>
+  <a href="https://json2vec.github.io/json2vec/"><img alt="Documentation" src="https://img.shields.io/badge/docs-Quarto-39729E?logo=quarto&amp;logoColor=white" /></a>
   <!-- discord-invite:start -->
   <a href="https://discord.gg/DVyZUkvTFA"><img alt="Discord channel invite" src="https://img.shields.io/badge/discord-join%20the%20channel-5865F2?logo=discord&amp;logoColor=white" /></a>
   <!-- discord-invite:end -->
@@ -24,7 +24,8 @@ A `json2vec` schema is both a data contract and an architecture blueprint.
 
 - Leaf fields such as `Number`, `Category`, `Set`, `Entity`, `Text`, and
   `Vector` become datatype-specific tensorfields.
-- `Array` nodes become local context encoders for repeated child objects.
+- `Branch` nodes define shared contexts for child fields, with optional local
+  attention and pooling before the representation flows upward.
 - Targets, masks, pruning, and embeddings are configured on the same schema
   tree.
 - Prediction output is keyed by schema address, so decoded values and
@@ -37,24 +38,23 @@ inference, and serving.
 ## A Model From A Nested Record
 
 ```python
-import json2vec as j2v
-
-model = j2v.Model.from_schema(
-    j2v.Category("customer_tier", max_vocab_size=16),
-    j2v.Array(
-        j2v.Category("sku", max_vocab_size=2048),
-        j2v.Number("quantity"),
-        j2v.Number("price"),
-        name="line_items",
-        max_length=32,
-        embed=True,
-    ),
-    j2v.Category("returned", target=True, max_vocab_size=2),
+import json2vec as jv
+
+model = jv.Model.from_tree(
     name="order",
     d_model=64,
     n_layers=2,
     n_heads=4,
     embed=True,
+    customer_tier=jv.Category(size=16),
+    line_items=jv.Branch(
+        length=32,
+        embed=True,
+        sku=jv.Category(size=2048),
+        quantity=jv.Number,
+        price=jv.Number,
+    ),
+    returned=jv.Category(target=True, size=2),
 )
 ```
 
@@ -77,8 +77,8 @@ to emit embeddings at configured addresses.
 
 ## Train With Lightning
 
-`j2v.Model` is a LightningModule. `j2v.PolarsDataModule` and
-`j2v.StreamingDataModule` are LightningDataModule implementations. The schema
+`jv.Model` is a LightningModule. `jv.PolarsDataModule` and
+`jv.StreamingDataModule` are LightningDataModule implementations. The schema
 defines the model tree, typed losses, prediction outputs, and embeddings;
 Lightning runs `fit`, `validate`, `test`, and `predict`.
 
@@ -87,23 +87,23 @@ import lightning.pytorch as lit
 import polars as pl
 import torch
 
-import json2vec as j2v
+import json2vec as jv
 
 records = pl.read_ndjson("docs/data/iris.jsonl").head(36)
 
-model = j2v.Model.from_schema(
-    j2v.Number("sepal_length"),
-    j2v.Number("petal_length"),
-    j2v.Category("species", target=True, max_vocab_size=4, topk=[2]),
+model = jv.Model.from_tree(
     d_model=16,
     n_layers=1,
     n_heads=4,
     batch_size=8,
     embed=True,
     optimizer=lambda module: torch.optim.AdamW(module.parameters(), lr=1e-2),
+    sepal_length=jv.Number,
+    petal_length=jv.Number,
+    species=jv.Category(target=True, size=4, topk=[2]),
 )
 
-datamodule = j2v.PolarsDataModule(
+datamodule = jv.PolarsDataModule(
     model=model,
     train=records,
     validate=records,
@@ -131,7 +131,7 @@ trainer.fit(model=model, datamodule=datamodule)
 For larger jobs, the same model can run through normal Lightning callbacks,
 checkpointing, precision settings, device placement, and distributed
 strategies. See
-[Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning/).
+[Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning.html).
 
 ## Predict And Embed
 
@@ -140,8 +140,8 @@ For small interactive batches, call `model.predict(...)` with raw dictionaries.
 ```python
 predictions = model.predict(records.to_dicts()[:3])
 
-species = predictions[j2v.Address("record", "species")]
-record = predictions[j2v.Address("record")]
+species = predictions[jv.Address("record", "species")]
+record = predictions[jv.Address("record")]
 
 print(species["content"]["value"])
 print(species["content"]["probability"])
@@ -149,10 +149,10 @@ print(record["embedding"])
 ```
 
 For larger offline jobs, configure a `predict` split on a data module and attach
-`j2v.Writer` to Lightning's prediction loop.
+`jv.Writer` to Lightning's prediction loop.
 
 ```python
-writer = j2v.Writer("predictions")
+writer = jv.Writer("predictions")
 
 trainer = lit.Trainer(
     accelerator="cpu",
@@ -160,7 +160,7 @@ trainer = lit.Trainer(
     logger=False,
 )
 
-predict_datamodule = j2v.PolarsDataModule(
+predict_datamodule = jv.PolarsDataModule(
     model=model,
     predict=records.drop("species"),
     num_workers=0,
@@ -174,8 +174,8 @@ trainer.predict(model=model, datamodule=predict_datamodule)
 `Writer` creates rank-partitioned Parquet files such as
 `predictions/rank-0.parquet`. Use a postprocessor when downstream systems need
 flat columns, renamed addresses, redacted payloads, or fewer fields. See
-[Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference/)
-and [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors/).
+[Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference.html)
+and [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors.html).
 
 ## Learning Modes
 
@@ -215,7 +215,7 @@ Choose the data module by where the records live:
 arguments are compiled regular expressions matched against discovered file
 paths.
 
-See [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules/)
+See [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules.html)
 for split configuration, sharding, sampling, buffers, and preprocessors.
 
 ## What Makes This Different
@@ -226,7 +226,7 @@ for split configuration, sharding, sampling, buffers, and preprocessors.
   missing-state handling, masking, decoding, loss, metrics, and output writing.
 - **Unified training roles:** `target=True`, `p_prune`, and `p_mask` all use the
   same reconstruction path.
-- **Embedding trees:** embeddings can come from the root, arrays, or selected
+- **Embedding trees:** embeddings can come from the root, branches, or selected
   leaves.
 - **Schema evolution:** fields can be added, removed, updated, reset, or
   temporarily overridden after construction.
@@ -279,55 +279,53 @@ uv sync --extra docs
 
 The `text` extra installs Hugging Face `transformers`. The `serving` extra
 installs FastAPI-backed deployment dependencies. The `docs` extra installs the
-MkDocs toolchain.
+Python packages used by the Quarto docs.
 
 ## Documentation Map
 
 Start with:
 
-- [Getting Started](https://json2vec.github.io/json2vec/getting-started/)
-- [AI / Expert Quickstart](https://json2vec.github.io/json2vec/ai-quickstart/)
-- [Model Tree](https://json2vec.github.io/json2vec/core-concepts/model-tree/)
-- [Query Paths](https://json2vec.github.io/json2vec/core-concepts/querypaths/)
-- [Built-In Data Types](https://json2vec.github.io/json2vec/core-concepts/data-types/)
-- [Learning Modes & Embeddings](https://json2vec.github.io/json2vec/core-concepts/embeddings/)
-- [Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning/)
-- [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules/)
-- [Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference/)
-- [API Reference](https://json2vec.github.io/json2vec/reference/api/)
+- [Getting Started](https://json2vec.github.io/json2vec/getting-started.html)
+- [AI / Expert Quickstart](https://json2vec.github.io/json2vec/ai-quickstart.html)
+- [Model Tree](https://json2vec.github.io/json2vec/core-concepts/model-tree.html)
+- [Query Paths](https://json2vec.github.io/json2vec/core-concepts/querypaths.html)
+- [Built-In Data Types](https://json2vec.github.io/json2vec/core-concepts/data-types.html)
+- [Learning Modes & Embeddings](https://json2vec.github.io/json2vec/core-concepts/embeddings.html)
+- [Training With Lightning](https://json2vec.github.io/json2vec/guides/lightning.html)
+- [Data Modules](https://json2vec.github.io/json2vec/guides/data-modules.html)
+- [Batch Inference](https://json2vec.github.io/json2vec/guides/batch-inference.html)
 
 Tutorials and guides:
 
-- [Hello World](https://json2vec.github.io/json2vec/tutorials/hello-world/)
-- [Supervised Tabular Training](https://json2vec.github.io/json2vec/tutorials/supervised-tabular-training/)
-- [Masked Pretraining](https://json2vec.github.io/json2vec/tutorials/pretraining/)
-- [Nested Supervised Training](https://json2vec.github.io/json2vec/tutorials/nested-supervised-training/)
-- [Serving](https://json2vec.github.io/json2vec/tutorials/serving/)
-- [Preprocessors](https://json2vec.github.io/json2vec/guides/preprocessors/)
-- [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors/)
-- [Field Importance](https://json2vec.github.io/json2vec/guides/field-importance/)
-- [Field Stacking](https://json2vec.github.io/json2vec/guides/field-stacking/)
-- [Custom Data Types](https://json2vec.github.io/json2vec/data-types/tensorfields/)
-- [Device Tenure Case Study](https://json2vec.github.io/json2vec/case-studies/device-tenure/)
-- [Whitepaper](https://json2vec.github.io/json2vec/whitepaper.pdf)
+- [Postprocessors](https://json2vec.github.io/json2vec/guides/postprocessors.html)
+- [Field Importance](https://json2vec.github.io/json2vec/guides/field-importance.html)
+- [Field Stacking](https://json2vec.github.io/json2vec/guides/field-stacking.html)
+- [Branch](https://json2vec.github.io/json2vec/data-types/branch.html)
+- [Number](https://json2vec.github.io/json2vec/data-types/number.html)
+- [Category](https://json2vec.github.io/json2vec/data-types/category.html)
+- [Set](https://json2vec.github.io/json2vec/data-types/set.html)
+- [Entity](https://json2vec.github.io/json2vec/data-types/entity.html)
+- [DateParts](https://json2vec.github.io/json2vec/data-types/dateparts.html)
+- [Vector](https://json2vec.github.io/json2vec/data-types/vector.html)
+- [Text](https://json2vec.github.io/json2vec/data-types/text.html)
+- [Device Tenure Case Study](https://json2vec.github.io/json2vec/case-studies/device-tenure.html)
 
 Build the docs locally with:
 
 ```bash
-uv run --extra docs mkdocs build --strict
+make render
 ```
 
 ## Repository Layout
 
 - `src/json2vec/architecture`: model assembly, attention, pooling, and routing
-- `src/json2vec/data`: dataset fetch/read/process/batch/encode pipeline
+- `src/json2vec/data`: dataset fetch/read/process/batch/encode pipeline and preprocessor exports
 - `src/json2vec/inference`: serving and prediction callbacks
 - `src/json2vec/logging`: runtime logging callbacks
-- `src/json2vec/preprocessors`: preprocessor registry
 - `src/json2vec/structs`: pydantic config models, enums, and tree nodes
 - `src/json2vec/tensorfields`: tensorfield plugin system and built-in fields
 - `tests/`: package test suite
-- `docs/`: tutorials, guides, diagrams, and whitepaper source
+- `docs/`: Quarto project, pages, guides, stylesheets, and sample data
 
 ## Development
 

{json2vec-0.4.8 → json2vec-0.4.10}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "json2vec"
-version = "0.4.8"
+version = "0.4.10"
 description = "Schema-first PyTorch models for hierarchical / nested / sequence data structures"
 readme = "README.md"
 license = "Apache-2.0"
@@ -32,13 +32,7 @@ text = [
     "transformers>=4.55.0",
 ]
 docs = [
-    "fastapi>=0.124.0",
-    "mkdocs-material>=9.6",
-    "mkdocs-jupyter>=0.26.3",
-    "mkdocstrings[python]>=0.27",
-    "orjson>=3.10.0",
-    "pydantic-settings>=2.10.1",
-    "uvicorn>=0.38.0",
+    "marimo>=0.23.8",
 ]
 
 [dependency-groups]

{json2vec-0.4.8 → json2vec-0.4.10}/src/json2vec/__init__.py

@@ -1,24 +1,38 @@
 """Public `json2vec` SDK surface.
 
 The top-level package exports the constructors and helpers used by most
-applications: `Model.from_schema(...)` for model construction, tensorfield
+applications: `Model.from_tree(...)` for model construction, tensorfield
 request constructors such as `Category` and `Number`, data modules, schema
 mutation predicates, and the `@preprocess` decorator.
 """
 
 from typing import TYPE_CHECKING, Any
 
+from json2vec import helpers as helpers
+from json2vec.architecture.checkpoint import RollbackCheckpoint
+from json2vec.architecture.mutations import MutationLockCallback, RuntimePlacementCallback
 from json2vec.architecture.root import (
     Model,
-    MutationLockCallback,
     OptimizerConfig,
-    RollbackCheckpoint,
-    RuntimePlacementCallback,
     SchedulerConfig,
 )
 from json2vec.data.datasets import CustomDataModule, PolarsDataModule, StreamingDataModule
-from json2vec.inference.callback import Postprocessor, Writer
-from json2vec.preprocessors import PREPROCESSORS, Preprocessor, PreprocessorMode, preprocess
+from json2vec.data.nested import MASK_LITERAL, MaskLiteral
+from json2vec.data.processors import (
+    Metadata,
+    Observation,
+    Postprocessor,
+    PostprocessorProvider,
+    PostprocessorResult,
+    Predictions,
+    Preprocessor,
+    PreprocessorProvider,
+    RawBatch,
+    RawObservation,
+    postprocess,
+    preprocess,
+)
+from json2vec.inference.callback import Writer
 from json2vec.structs.enums import (
     AttentionMode,
     Component,
@@ -31,14 +45,14 @@ from json2vec.structs.enums import (
     Tokens,
 )
 from json2vec.structs.experiment import (
-    Hyperparameters,
     NodeAttribute,
     NodePredicate,
+    Schema,
     SchemaField,
     predicate,
     where,
 )
-from json2vec.structs.structure import Array
+from json2vec.structs.structure import Branch, Mask
 from json2vec.structs.tree import Address, Leaf
 from json2vec.tensorfields import TENSORFIELDS, DecoderBase, EmbedderBase, Plugin, RequestBase, TensorFieldBase
 from json2vec.tensorfields.extensions.category import Request as Category
@@ -95,7 +109,7 @@ def __dir__() -> list[str]:
 __all__ = [
     "Address",
     "Accelerator",
-    "Array",
+    "Branch",
     "AttentionMode",
     "Category",
     "Component",
@@ -105,25 +119,35 @@ __all__ = [
     "Deployment",
     "EmbedderBase",
     "Entity",
-    "Hyperparameters",
+    "helpers",
+    "Schema",
     "Input",
     "JSONBackend",
     "Leaf",
     "Metric",
+    "MASK_LITERAL",
+    "Mask",
+    "MaskLiteral",
+    "Metadata",
     "Model",
     "ModelSource",
     "MutationLockCallback",
     "NodeAttribute",
     "NodePredicate",
     "Number",
+    "Observation",
     "OptimizerConfig",
     "Overflow",
-    "PREPROCESSORS",
     "Plugin",
     "PolarsDataModule",
     "Postprocessor",
+    "PostprocessorProvider",
+    "PostprocessorResult",
+    "Predictions",
     "Preprocessor",
-    "PreprocessorMode",
+    "PreprocessorProvider",
+    "RawBatch",
+    "RawObservation",
     "RequestBase",
     "RollbackCheckpoint",
     "RuntimePlacementCallback",
@@ -144,6 +168,7 @@ __all__ = [
     "VocabularySyncCallback",
     "Writer",
     "predicate",
+    "postprocess",
     "preprocess",
     "where",
 ]