terset 0.2.1__tar.gz

terset-0.2.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 TrueWear, LLC
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

terset-0.2.1/PKG-INFO

@@ -0,0 +1,132 @@
+Metadata-Version: 2.4
+Name: terset
+Version: 0.2.1
+Summary: Encode JSON records as compact, typed tables.
+Author: TrueWear, LLC
+License-Expression: MIT
+Project-URL: Homepage, https://terset.dev
+Project-URL: Source, https://github.com/TrueWear/terset
+Keywords: json,llm,prompt,tokens,table,encoding
+Classifier: Development Status :: 4 - Beta
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Dynamic: license-file
+
+# Terset for Python
+
+Terset writes JSON records as compact, typed tables using format v2. Python 3.10 or newer.
+Zero runtime dependencies.
+
+## Install
+
+Coming soon on PyPI. This package is not published yet.
+
+```sh
+pip install terset
+```
+
+## Use
+
+```python
+from terset import decode, encode, rules_for
+
+data = [{"id": "1", "name": "Ada"}, {"id": "2", "name": "Ben"}]
+text = encode(data)
+print(text)
+print(decode(text) == data)
+```
+
+Output (the gaps are tabs):
+
+```text
+[2]
+id	name
+1	Ada
+2	Ben
+True
+```
+
+`encode(value)` accepts a dictionary or list at the root. Values can be
+strings, finite numbers, booleans, `None`, dictionaries with string keys,
+and lists. It returns canonical Terset text. Put `rules_for(text)` before the data
+when a reader needs the format explained. It validates canonical text and returns
+the base rules, plus rules for keyed tables and dotted headers when present.
+The exported `RULES` keeps all rules for any text. `decode(text)` returns the same
+JSON value and rejects text the encoder would not produce.
+
+Lists of at least two dictionaries become tables when the header and rows
+save UTF-8 bytes. Named tables start with `name[N]:`. Root tables start
+with `[N]` on its own line. Compact JSON lists have no count.
+Missing fields, empty strings and `None` stay distinct.
+
+A map in a direct root field can use `name{N}:` when the complete named
+block is strictly shorter than `name: JSON`. Its first column is `@key`,
+which holds each entry's key and is not a field. Root maps stay fields.
+Uniform nested objects with identical ordered simple keys expand into
+dotted headers, recursively. Literal dotted field names are quoted.
+Other nested values stay compact JSON. The decoder rejects count mismatches
+and accepts only the current canonical format.
+See the full [format v2 specification](https://github.com/TrueWear/terset/blob/main/packages/terset/SPEC.md).
+
+## Python details
+
+- The format uses JavaScript binary64 numbers. `1.0` writes as `1`.
+  Decoded integer literals within `[-2**53, 2**53]` become Python `int`
+  values. Tokens with a fraction or exponent, and integers outside this
+  range, become `float` values.
+  Python integers must be exactly representable as binary64 numbers;
+  otherwise encoding raises `TypeError`. This avoids silently rounding
+  integers such as `9007199254740993`.
+- Numeric negative zero, infinities, NaN, cycles, non-JSON types and depth
+  above 500 raise `TypeError`. Use plain built-in dictionaries and lists.
+- Dictionary order is preserved, with JavaScript's array-index keys first
+  in numeric order. For example, `"2"` precedes `"10"`; `"01"` keeps its
+  insertion position among other keys.
+- Unicode stays unchanged, including normalization forms and lone
+  surrogates. Explicit UTF-16 surrogate pairs become Python astral code
+  points. Keys that become identical through this conversion are rejected.
+- Format validation errors raise `ValueError`, including non-finite numbers
+  and excessive depth. Passing a non-string to `decode` raises `TypeError`.
+
+`encode` uses the covariant `JsonInputRoot` and `JsonInputValue` aliases,
+based on read-only `Mapping` and `Sequence` interfaces. This lets typed
+lists and dictionaries pass without widening their element types. Runtime
+validation still requires plain built-in lists and dictionaries.
+`decode` returns concrete `JsonRoot`, `JsonObject` and `JsonValue` types.
+The package includes `py.typed`.
+
+## Limits
+
+The format rejects numeric negative zero and nesting deeper than 500.
+With deep data and an already deep caller stack,
+`encode`, `decode` and `rules_for` can raise `RecursionError`.
+This depends on the available stack depth, not the Python version.
+
+## Local tests
+
+From the repository root, without installing the package:
+
+```sh
+python3 -B -m unittest discover -s packages/terset-py/tests -v
+```
+
+Tests read the shared cases generated by the TypeScript encoder and check
+exact output, decoding, the rules string, and invalid inputs. To regenerate
+the cases from the repository root with Node 24:
+
+```sh
+node packages/terset/tests/generate-cases.ts
+```
+
+## License
+
+MIT. Copyright (c) 2026 TrueWear, LLC. The license text is included in `LICENSE`.

terset-0.2.1/README.md

@@ -0,0 +1,109 @@
+# Terset for Python
+
+Terset writes JSON records as compact, typed tables using format v2. Python 3.10 or newer.
+Zero runtime dependencies.
+
+## Install
+
+Coming soon on PyPI. This package is not published yet.
+
+```sh
+pip install terset
+```
+
+## Use
+
+```python
+from terset import decode, encode, rules_for
+
+data = [{"id": "1", "name": "Ada"}, {"id": "2", "name": "Ben"}]
+text = encode(data)
+print(text)
+print(decode(text) == data)
+```
+
+Output (the gaps are tabs):
+
+```text
+[2]
+id	name
+1	Ada
+2	Ben
+True
+```
+
+`encode(value)` accepts a dictionary or list at the root. Values can be
+strings, finite numbers, booleans, `None`, dictionaries with string keys,
+and lists. It returns canonical Terset text. Put `rules_for(text)` before the data
+when a reader needs the format explained. It validates canonical text and returns
+the base rules, plus rules for keyed tables and dotted headers when present.
+The exported `RULES` keeps all rules for any text. `decode(text)` returns the same
+JSON value and rejects text the encoder would not produce.
+
+Lists of at least two dictionaries become tables when the header and rows
+save UTF-8 bytes. Named tables start with `name[N]:`. Root tables start
+with `[N]` on its own line. Compact JSON lists have no count.
+Missing fields, empty strings and `None` stay distinct.
+
+A map in a direct root field can use `name{N}:` when the complete named
+block is strictly shorter than `name: JSON`. Its first column is `@key`,
+which holds each entry's key and is not a field. Root maps stay fields.
+Uniform nested objects with identical ordered simple keys expand into
+dotted headers, recursively. Literal dotted field names are quoted.
+Other nested values stay compact JSON. The decoder rejects count mismatches
+and accepts only the current canonical format.
+See the full [format v2 specification](https://github.com/TrueWear/terset/blob/main/packages/terset/SPEC.md).
+
+## Python details
+
+- The format uses JavaScript binary64 numbers. `1.0` writes as `1`.
+  Decoded integer literals within `[-2**53, 2**53]` become Python `int`
+  values. Tokens with a fraction or exponent, and integers outside this
+  range, become `float` values.
+  Python integers must be exactly representable as binary64 numbers;
+  otherwise encoding raises `TypeError`. This avoids silently rounding
+  integers such as `9007199254740993`.
+- Numeric negative zero, infinities, NaN, cycles, non-JSON types and depth
+  above 500 raise `TypeError`. Use plain built-in dictionaries and lists.
+- Dictionary order is preserved, with JavaScript's array-index keys first
+  in numeric order. For example, `"2"` precedes `"10"`; `"01"` keeps its
+  insertion position among other keys.
+- Unicode stays unchanged, including normalization forms and lone
+  surrogates. Explicit UTF-16 surrogate pairs become Python astral code
+  points. Keys that become identical through this conversion are rejected.
+- Format validation errors raise `ValueError`, including non-finite numbers
+  and excessive depth. Passing a non-string to `decode` raises `TypeError`.
+
+`encode` uses the covariant `JsonInputRoot` and `JsonInputValue` aliases,
+based on read-only `Mapping` and `Sequence` interfaces. This lets typed
+lists and dictionaries pass without widening their element types. Runtime
+validation still requires plain built-in lists and dictionaries.
+`decode` returns concrete `JsonRoot`, `JsonObject` and `JsonValue` types.
+The package includes `py.typed`.
+
+## Limits
+
+The format rejects numeric negative zero and nesting deeper than 500.
+With deep data and an already deep caller stack,
+`encode`, `decode` and `rules_for` can raise `RecursionError`.
+This depends on the available stack depth, not the Python version.
+
+## Local tests
+
+From the repository root, without installing the package:
+
+```sh
+python3 -B -m unittest discover -s packages/terset-py/tests -v
+```
+
+Tests read the shared cases generated by the TypeScript encoder and check
+exact output, decoding, the rules string, and invalid inputs. To regenerate
+the cases from the repository root with Node 24:
+
+```sh
+node packages/terset/tests/generate-cases.ts
+```
+
+## License
+
+MIT. Copyright (c) 2026 TrueWear, LLC. The license text is included in `LICENSE`.

terset-0.2.1/pyproject.toml

@@ -0,0 +1,36 @@
+[build-system]
+requires = ["setuptools>=77"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "terset"
+version = "0.2.1"
+description = "Encode JSON records as compact, typed tables."
+readme = "README.md"
+requires-python = ">=3.10"
+license = "MIT"
+license-files = ["LICENSE"]
+authors = [{ name = "TrueWear, LLC" }]
+dependencies = []
+keywords = ["json", "llm", "prompt", "tokens", "table", "encoding"]
+classifiers = [
+  "Development Status :: 4 - Beta",
+  "Programming Language :: Python :: 3",
+  "Programming Language :: Python :: 3 :: Only",
+  "Programming Language :: Python :: 3.10",
+  "Programming Language :: Python :: 3.11",
+  "Programming Language :: Python :: 3.12",
+  "Programming Language :: Python :: 3.13",
+  "Programming Language :: Python :: 3.14",
+  "Typing :: Typed",
+]
+
+[project.urls]
+Homepage = "https://terset.dev"
+Source = "https://github.com/TrueWear/terset"
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.setuptools.package-data]
+terset = ["py.typed"]

terset-0.2.1/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

terset-0.2.1/src/terset/__init__.py

@@ -0,0 +1,5 @@
+"""Encode and decode canonical Terset text without runtime dependencies."""
+
+from ._terset import RULES, JsonInputRoot, JsonInputValue, JsonObject, JsonRoot, JsonValue, decode, encode, rules_for
+
+__all__ = ["RULES", "JsonInputRoot", "JsonInputValue", "JsonObject", "JsonRoot", "JsonValue", "decode", "encode", "rules_for"]