structtype 0.12.0__tar.gz → 0.14.0__tar.gz

structtype-0.14.0/.opencode/opencode.json

@@ -0,0 +1,4 @@
+{
+  "$schema": "https://opencode.ai/config.json",
+  "plugin": []
+}

{structtype-0.12.0 → structtype-0.14.0}/AGENTS.md

@@ -3,7 +3,7 @@
 ## Project
 
 Fast struct validation + JSON serialization for Python.
-Core is a monolithic C extension (`src/structtype/_core.c`, ~21K lines).
+Core is a monolithic C extension (`src/structtype/_core.c`, ~22K lines).
 No runtime deps.
 
 ## Setup
@@ -25,16 +25,23 @@ Use `make` targets where available. Targeted tests can be run directly with
 | Coverage (Python + C) | `make test-cov-c` |
 | Coverage (Python + C, all Pythons, merged) | `make test-cov-c-all` |
 | Tests in all supported Pythons | `make test-all` |
+| Doctests | `make test-doc` |
+| Build sdist + wheel | `make build` |
 | Build docs | `make docs` |
 | Format | `make format` |
 | Lint | `make ruff-check` |
 | Type check | `make type-check` |
 | Static checks | `make check` |
 
+Benchmark targets: `make bench`, `make bench-validators`, `make bench-codecs`,
+`make bench-csv-1m`, `make bench-strings` (run sequentially; CPU-bound).
+
 ## Conventions
 
 - **88-char lines**, formatted with `ruff format`
-- `ruff check` with rules `E`, `F`, `I`, `W`
+- `ruff check` with Ruff's default rule set (the project config sets only
+  `target-version` and isort `combine-as-imports`; line length is enforced by
+  `ruff format`, not `E501`)
 - Private modules/functions prefixed with `_`
 - C code uses `ms_`/`MS_` prefix
 - Type stubs (`.pyi`) alongside public modules
@@ -59,6 +66,8 @@ Use `make` targets where available. Targeted tests can be run directly with
 - `obj.struct_check_types()` — validate field values against types + constraints (pure type-check, no conversion)
 - `cls.struct_validate_json(buf, *, strict=True)` — deserialize from JSON
 - `cls.struct_validate(obj, *, strict=True, from_attributes=False)` — convert built-in types to struct
+- `obj.struct_dump_csv()` — encode one row as `list[str]` for `csv.writer.writerow()`; flat scalar fields only, positional
+- `cls.struct_validate_csv(row, *, null_values=("",))` — decode one CSV row (sequence of cell strings); always lax
 
 ### Dict & Iteration Protocol
 
@@ -75,15 +84,3 @@ Struct instances support the mapping protocol:
 - Validation matches keys by the **alias** name only, except `struct_validate(obj, from_attributes=True)` on a **non-dict object**, which matches by both the python field name and the alias. Dict/JSON input (even with `from_attributes=True`) and all dump/serialization use only the alias name.
 - For `Annotated[MyType, Serializer(load=...)]`, an existing `MyType` instance bypasses `load`; other Python values and JSON representations are passed to `load`. Serializers remain rejected on `Any`, `bool`, `int`, `float`, `str`, `list`, `dict`, `tuple`, `TypedDict`, `NamedTuple`, `frozendict`, and `Literal`. `Serializer` and `Constraint` must attach to a concrete type inside `Annotated` — `Annotated[Union[A, B], ...]` and `Annotated[Optional[T], ...]` are rejected; use `Annotated[T, ...] | None` for optional fields.
 
-## graphify
-
-This project has a knowledge graph at graphify-out/ with god nodes, community structure, and cross-file relationships.
-
-When the user types `/graphify`, use the installed graphify skill or instructions before doing anything else.
-
-Rules:
-- For codebase questions, first run `graphify query "<question>"` when graphify-out/graph.json exists. Use `graphify path "<A>" "<B>"` for relationships and `graphify explain "<concept>"` for focused concepts. These return a scoped subgraph, usually much smaller than GRAPH_REPORT.md or raw grep output.
-- Dirty graphify-out/ files are expected after hooks or incremental updates; dirty graph files are not a reason to skip graphify. Only skip graphify if the task is about stale or incorrect graph output, or the user explicitly says not to use it.
-- If graphify-out/wiki/index.md exists, use it for broad navigation instead of raw source browsing.
-- Read graphify-out/GRAPH_REPORT.md only for broad architecture review or when query/path/explain do not surface enough context.
-- After modifying code, run `graphify update .` to keep the graph current (AST-only, no API cost).

{structtype-0.12.0 → structtype-0.14.0}/Makefile

@@ -143,6 +143,14 @@ bench-validators: ## run Serializer/Validator benchmarks
 bench-codecs: ## run Constraint + Serializer heavy benchmarks
 	uv run -p 3.15 benchmarks/bench_codecs.py
 
+.PHONY: bench-csv-1m
+bench-csv-1m: ## run 1M-row on-disk CSV benchmark
+	uv run -p 3.15 benchmarks/bench_csv_1m.py
+
+.PHONY: bench-strings
+bench-strings: ## run long-string JSON codec benchmark
+	uv run -p 3.15 benchmarks/bench_strings.py
+
 ##@ Utility
 .PHONY: clean
 clean: ## Delete all temporary files
@@ -157,6 +165,7 @@ clean: ## Delete all temporary files
 	rm -rf htmlcov-c
 	rm -f .coverage
 	rm -f coverage-c.info coverage-c.info.*
+	rm -rf .venv*
 
 .PHONY: install
 install: install-uv ## Install virtual environment

{structtype-0.12.0/src/structtype.egg-info → structtype-0.14.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: structtype
-Version: 0.12.0
+Version: 0.14.0
 Summary: Fast Struct type with validation and JSON serialization for Python.
 Maintainer-email: Wolfgang Langner <tds333@mailbox.org>
 License-Expression: BSD-3-Clause
@@ -77,6 +77,8 @@ User.struct_validate_json(
 
 Full documentation is available at **https://tds333.github.io/structtype/**.
 
+The library is in beta phase, interface and features may have breaking changes.
+
 ## Benchmarks
 
 structtype is as fast as msgspec and about 3-5x faster than pydantic. See the
@@ -93,4 +95,4 @@ structtype is as fast as msgspec and about 3-5x faster than pydantic. See the
 ## License
 
 New BSD. See the [License File](LICENSE).
-The core is based on the work of Jim Crist-Harif from msgspec.
+Parts of the core are based on the work of Jim Crist-Harif from msgspec.

{structtype-0.12.0 → structtype-0.14.0}/README.md

@@ -51,6 +51,8 @@ User.struct_validate_json(
 
 Full documentation is available at **https://tds333.github.io/structtype/**.
 
+The library is in beta phase, interface and features may have breaking changes.
+
 ## Benchmarks
 
 structtype is as fast as msgspec and about 3-5x faster than pydantic. See the
@@ -67,4 +69,4 @@ structtype is as fast as msgspec and about 3-5x faster than pydantic. See the
 ## License
 
 New BSD. See the [License File](LICENSE).
-The core is based on the work of Jim Crist-Harif from msgspec.
+Parts of the core are based on the work of Jim Crist-Harif from msgspec.

{structtype-0.12.0 → structtype-0.14.0}/src/structtype/__init__.pyi

@@ -1,6 +1,6 @@
 # ruff: noqa: PYI041, PYI015
 import enum
-from collections.abc import Callable, Iterable, Iterator, Mapping
+from collections.abc import Callable, Iterable, Iterator, Mapping, Sequence
 from inspect import Signature
 from typing import (
     Any,
@@ -85,6 +85,7 @@ class Struct(metaclass=StructMeta):
         uuid_as_hex: bool = False,
         sort_keys: bool = False,
     ) -> bytes: ...
+    def struct_dump_csv(self) -> list[str]: ...
     def struct_dump(
         self,
         *,
@@ -101,6 +102,13 @@ class Struct(metaclass=StructMeta):
         strict: bool = True,
     ) -> Self: ...
     @classmethod
+    def struct_validate_csv(
+        cls,
+        row: Sequence[str],
+        *,
+        null_values: Sequence[str] = ("",),
+    ) -> Self: ...
+    @classmethod
     def struct_validate(
         cls,
         obj: Any,