Rename units package to physical_quantities and bump to 0.3.0rc4

README.md

@@ -17,13 +17,13 @@ tag. They depend on each other by version, so install the ones you use
 **and the packages they depend on** from the same tag:
 
 ```bash
-pip install "lambdafeedback-slr-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=slr_parsing"
-pip install "lambdafeedback-criteria @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=criteria"
-pip install "lambdafeedback-expression-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=expression_parsing"
-pip install "lambdafeedback-units @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=units"
+pip install "lambdafeedback-slr-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=slr_parsing"
+pip install "lambdafeedback-criteria @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=criteria"
+pip install "lambdafeedback-expression-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=expression_parsing"
+pip install "lambdafeedback-physical-quantities @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=physical_quantities"
 ```
 
-With Poetry: `lambdafeedback-units = { git = "https://github.com/lambda-feedback/lambda-feedback-packages.git", tag = "v0.3.0-rc.3", subdirectory = "units" }`, and likewise for each package.
+With Poetry: `lambdafeedback-physical-quantities = { git = "https://github.com/lambda-feedback/lambda-feedback-packages.git", tag = "v0.3.0-rc.4", subdirectory = "physical_quantities" }`, and likewise for each package.
 
 ## Packages
 
@@ -32,7 +32,7 @@ With Poetry: `lambdafeedback-units = { git = "https://github.com/lambda-feedback
 | `slr_parsing` | `lambdafeedback-slr-parsing` | `lambdafeedback.slr_parsing` | Generic SLR(1) parser engine | — |
 | `criteria` | `lambdafeedback-criteria` | `lambdafeedback.criteria` | Criteria DSL parser and evaluation graph | `slr_parsing` |
 | `expression_parsing` | `lambdafeedback-expression-parsing` | `lambdafeedback.expression_parsing` | Safe SymPy parsing, preprocessing and LaTeX input/output | `slr_parsing`, `sympy`, `latex2sympy2_extended` |
-| `units` | `lambdafeedback-units` | `lambdafeedback.units` | Unit-system data, physical-quantity parsing and dimensional analysis | `slr_parsing`, `expression_parsing`, `sympy` |
+| `physical_quantities` | `lambdafeedback-physical-quantities` | `lambdafeedback.physical_quantities` | Unit-system data, physical-quantity parsing and dimensional analysis | `slr_parsing`, `expression_parsing`, `sympy` |
 
 ```python
 from lambdafeedback.expression_parsing import (
@@ -41,7 +41,7 @@ from lambdafeedback.expression_parsing import (
     parse_expression,
     preprocess_expression,
 )
-from lambdafeedback.units import QuantityParams, parse_quantity
+from lambdafeedback.physical_quantities import QuantityParams, parse_quantity
 
 params = ExpressionParams.from_dict(evaluation_params)  # JSON parameters → typed
 pre = preprocess_expression(response, params, name="response")  # aliases, brackets, |x|
@@ -54,7 +54,7 @@ quantity = parse_quantity("9.81 m/s^2", QuantityParams.from_dict(evaluation_para
 from [toolkit-python](https://github.com/lambda-feedback/toolkit-python), with
 `CriteriaGraph.export_feedback` for criteria feedback.
 
-Each package uses a `src/` layout: `units/src/lambdafeedback/units/`.
+Each package uses a `src/` layout: `physical_quantities/src/lambdafeedback/physical_quantities/`.
 There is deliberately no `lambdafeedback/__init__.py`; the namespace is
 [PEP 420](https://peps.python.org/pep-0420/), so the separately installed
 packages merge under one import root.
@@ -69,7 +69,7 @@ that installs every package in develop mode, plus the tooling, into one venv:
 ```bash
 poetry install                      # creates the venv with all packages + dev tools
 poetry run scripts/test.sh -q       # every package's tests (one pytest process each)
-poetry run pytest units/tests       # a single package
+poetry run pytest physical_quantities/tests  # a single package
 poetry run ruff check . && poetry run ruff format --check .
 poetry run mypy
 poetry run python tools/parity/check.py   # diff behaviour against the v0.1 baseline
@@ -121,7 +121,7 @@ attached to a GitHub Release (marked as a pre-release for pre-release tags).
 ## Logging
 
 The packages log at DEBUG only, through loggers named after their modules
-(`lambdafeedback.expression_parsing.preprocessing`, `lambdafeedback.units.quantity`, ...):
+(`lambdafeedback.expression_parsing.preprocessing`, `lambdafeedback.physical_quantities.quantity`, ...):
 how a response was preprocessed and read, how a quantity was split into value
 and unit, which evaluations a criteria graph ran. Failures are exceptions, not
 log records. To see them:

criteria/CHANGELOG.md

@@ -1,6 +1,6 @@
 # Changelog: lambdafeedback-criteria
 
-## 0.3.0 (unreleased; pre-release 0.3.0rc3, tag v0.3.0-rc.3)
+## 0.3.0 (unreleased; pre-release 0.3.0rc4, tag v0.3.0-rc.4)
 
 The first public release, after the extraction refactor. The import path is now `lambdafeedback.criteria` (was `criteria`). There was no 0.2.0 release: all four `lambdafeedback-*` packages share one version, and the first release is 0.3.0. Supports Python 3.12 to 3.14 (3.11 is not supported); `__version__` gives the installed version.
 

criteria/README.md

@@ -14,7 +14,7 @@ evaluation function.
 From git (not on PyPI yet):
 
 ```bash
-pip install "lambdafeedback-criteria @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=criteria"
+pip install "lambdafeedback-criteria @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=criteria"
 ```
 
 Install the packages it depends on (`slr_parsing`) from the same tag; see the [repository README](../README.md#install).

criteria/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [project]
 name = "lambdafeedback-criteria"
-version = "0.3.0rc3"
+version = "0.3.0rc4"
 description = "Criteria parser and criteria graphs that turn evaluations into feedback, for Lambda Feedback evaluation functions."
 readme = "README.md"
 license = "MIT"
@@ -25,7 +25,7 @@ classifiers = [
 ]
 requires-python = ">=3.12"
 dependencies = [
-    "lambdafeedback-slr-parsing (>=0.3.0rc3,<0.4.0)",
+    "lambdafeedback-slr-parsing (>=0.3.0rc4,<0.4.0)",
 ]
 
 [project.urls]

expression_parsing/CHANGELOG.md

@@ -1,6 +1,6 @@
 # Changelog: lambdafeedback-expression-parsing
 
-## 0.3.0 (unreleased; pre-release 0.3.0rc3, tag v0.3.0-rc.3)
+## 0.3.0 (unreleased; pre-release 0.3.0rc4, tag v0.3.0-rc.4)
 
 The first public release. Further tidy-up after the 0.2.0 extraction: previewing a response and mapping failures to learner-facing feedback text are app-layer concerns, so they move out of this package. `__version__` gives the installed version.
 
@@ -59,9 +59,9 @@ The first public release. Further tidy-up after the 0.2.0 extraction: previewing
 - (since 0.3.0rc1) `parse_latex` leaves division between numbers unevaluated, as the lambda-feedback fork did: `\frac{10}{2}`, `10/2` and `10 \div 2` give `10/2`, and `\frac{3}{6}` gives `3/6`. In rc1 latex2sympy2_extended divided two integers as a `Rational` (`5` and `1/2`), so a "written as" criterion couldn't tell `\frac{3}{6}` from `\frac{1}{2}`. `2\frac{1}{2}` is the product `2*1/2` again (rc1 read it as the mixed number `5/2`).
 - `parse_expression` leaves every SymPy function unevaluated, not only the ones SymPy's own `evaluate=False` covers (trigonometric, `exp`, `log`, ...). `factorial`, `binomial` and the number-theory functions computed while parsing: `10000000!` took minutes, and `100000!` gave an integer too long to print. As a result `5!` now parses as `factorial(5)`, not `120`, like `sin(0)` stays `sin(0)`; with `simplify` they are still computed. `parse_latex` likewise leaves `\binom` unevaluated (`\binom{5}{2}` gives `binomial(5, 2)`, not `10`): latex2sympy2_extended computed it, and `\binom{100000000}{50000000}` didn't finish. Operator classes called by name (`Pow`, `Add`, `Eq`, ...) are left unevaluated too: `Pow(9,Pow(9,9))` never finished, in every mode.
 - **One LaTeX differential broke every later LaTeX parse in the process.** latex2sympy2 (and the lambda-feedback fork) overwrote a module-level variable when parsing `\mathrm{d}x`, after which every `parse_latex` call failed with "argument of type 'Symbol' is not iterable"; in a warm evaluation-function container, one learner's answer broke LaTeX for every later request. latex2sympy2_extended keeps no state between calls. (Upstream issue OrangeX4/latex2sympy#4, closed unfixed.)
-- `substitute` scales with the input's length. It tried every substitution at every position, and a `FunctionName` copied the rest of the input on each check, so a long response with many substitutions was slow: 100000 characters with the units package's ~1000 reserved names took over 10 s. Substitutions are now indexed by their first character (keeping their order, so results are unchanged).
+- `substitute` scales with the input's length. It tried every substitution at every position, and a `FunctionName` copied the rest of the input on each check, so a long response with many substitutions was slow: 100000 characters with the physical quantities package's ~1000 reserved names took over 10 s. Substitutions are now indexed by their first character (keeping their order, so results are unchanged).
 - `is_complex_number_on_exponential_form` took exponential time on inputs like `2*exp(` followed by a run of digits (26 digits took over 10 s), which a learner's response could trigger through "written as" criteria. `NUMBER_REGEX` and `NONNEGATIVE_NUMBER_REGEX` could match an empty string after a digit, and the check repeated them; they now never match empty (they accept the same numbers), and the complex-form checks match each number atomically. `is_complex_number_on_cartesian_form` no longer accepts an empty string or a lone `+`, and both checks no longer accept numbers run together (`5E7e7`, `exp04I`). `generate_arbitrary_number_pattern_matcher` matches each number atomically too: a pattern whose numbers are separated only by spaces (`1 2 3`) took polynomial time on a long run of digits (1000 digits: over 8 s), and matched a single number (`3`) through the empty-match bug.
-- `ExpressionParams` (and so `SympyParsingConfig`, and subclasses such as `lambdafeedback.units.QuantityParams`) can be pickled and deep-copied. `symbols` is a read-only `MappingProxyType`, which pickle can't handle; it is stored as a `dict` and wrapped again on unpickling, without validating again.
+- `ExpressionParams` (and so `SympyParsingConfig`, and subclasses such as `lambdafeedback.physical_quantities.QuantityParams`) can be pickled and deep-copied. `symbols` is a read-only `MappingProxyType`, which pickle can't handle; it is stored as a `dict` and wrapped again on unpickling, without validating again.
 - A symbol code that merely contains `lambda` (e.g. `lambda_1`) keeps its name. Every substitution's replacement had `lambda` rewritten to `lamda`, so `lambda_1` became `lamda_1`, which was then split as `lamda*_1` and lost its LaTeX.
 - A task symbol whose code is a Python keyword (`in` for inches, `del`, `as`, ...) parses as that symbol, keeping its LaTeX and assumptions; it used to raise (or, for `as`, become `a*s`).
 - `sympy_to_latex` prints symbols that have assumptions (from `symbol_assumptions`) with the task's LaTeX. The printer matched symbols by equality, and `Symbol('x', positive=True)` is not `Symbol('x')`.

expression_parsing/README.md

@@ -11,7 +11,7 @@ functions such as compareExpressions.
 From git (not on PyPI yet):
 
 ```bash
-pip install "lambdafeedback-expression-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.3#subdirectory=expression_parsing"
+pip install "lambdafeedback-expression-parsing @ git+https://github.com/lambda-feedback/lambda-feedback-packages.git@v0.3.0-rc.4#subdirectory=expression_parsing"
 ```
 
 Install the packages it depends on (`slr_parsing`) from the same tag; see the [repository README](../README.md#install).

expression_parsing/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "poetry.core.masonry.api"
 
 [project]
 name = "lambdafeedback-expression-parsing"
-version = "0.3.0rc3"
+version = "0.3.0rc4"
 description = "Safe SymPy parsing and preprocessing of learner-written mathematical expressions, for Lambda Feedback evaluation functions."
 readme = "README.md"
 license = "MIT"
@@ -27,7 +27,7 @@ classifiers = [
 ]
 requires-python = ">=3.12"
 dependencies = [
-    "lambdafeedback-slr-parsing (>=0.3.0rc3,<0.4.0)",
+    "lambdafeedback-slr-parsing (>=0.3.0rc4,<0.4.0)",
     "sympy (>=1.14,<2.0)",
     "latex2sympy2-extended[antlr4-13-2] (>=1.11.0,<2.0)",
 ]

units/CHANGELOG.md → physical_quantities/CHANGELOG.md

@@ -1,11 +1,12 @@
-# Changelog: lambdafeedback-units
+# Changelog: lambdafeedback-physical-quantities
 
-## 0.3.0 (unreleased; pre-release 0.3.0rc3, tag v0.3.0-rc.3)
+## 0.3.0 (unreleased; pre-release 0.3.0rc4, tag v0.3.0-rc.4)
 
 The first public release. This is a further tidy-up after the 0.2.0 extraction: the package now only does unit conversion. Package metadata (MIT licence, README, URLs) is in place for PyPI. Python 3.13 and later is supported, now that `lambdafeedback-expression-parsing` no longer depends on latex2sympy2. SymPy 1.14 or later is required (was 1.12): 1.12 predates Python 3.12, and a configuration with a declared function (`symbol_assumptions` `function`) can only be pickled from 1.14.
 
 ### Breaking
 
+- (since 0.3.0rc4) **Renamed from `lambdafeedback-units` to `lambdafeedback-physical-quantities`.** The import path is `lambdafeedback.physical_quantities` (was `lambdafeedback.units`), and the repository subdirectory is `physical_quantities` (was `units`). The API is otherwise unchanged.
 - Requires Python 3.12 or later (3.11 is not supported).
 - `PhysicalQuantity.standard_values` is a tuple of expressions (several for a value such as `{1, 2}`), replacing `standard_value`, which was one expression or a set.
 
@@ -20,7 +21,7 @@ The first public release. This is a further tidy-up after the 0.2.0 extraction:
 - An unknown `strictness` raises `QuantityParamsError` (with `.field`). It is still an `ExpressionParsingError`, so existing handlers keep working.
 - `UNIT_SETS` and `CONVERSION_TO_BASE_SI` are read-only mappings (`MappingProxyType`). Assigning to them raises `TypeError`, where before it would have silently changed every cached parser.
 - `parse_quantity` and `preprocess_quantity` now take `QuantityParams` only; the `Mapping[str, Any]` (raw dict) fallback is gone, along with the `as_quantity_params` helper. Convert evaluation-function parameters explicitly at the call site with `QuantityParams.from_dict(evaluation_params)`.
-- `preview_function` and `fix_exponents` have moved out of this package entirely, to compareExpressions (`app/context/physical_quantity_preview.py`): building a LaTeX/SymPy preview string is an app-layer display concern, not unit conversion. Import them from there instead of `lambdafeedback.units`.
+- `preview_function` and `fix_exponents` have moved out of this package entirely, to compareExpressions (`app/context/physical_quantity_preview.py`): building a LaTeX/SymPy preview string is an app-layer display concern, not unit conversion. Import them from there instead of `lambdafeedback.physical_quantities`.
 - `preprocess_quantity` now returns a plain `str` (`expression_parsing.preprocess_expression`, which it matches the shape of, no longer returns `Preprocessed`).
 - `FeedbackTag` is removed, matching `expression_parsing`'s move from feedback tags to typed exceptions: mapping a parsing fact to feedback text is an app-layer concern, not this package's. `PhysicalQuantity.messages` (`(message_id, FeedbackTag)` pairs) is replaced by `PhysicalQuantity.reverted_units: tuple[RevertedUnit, ...]` — plain `before`/`marked`/`after` facts about unit-like text found inside a value, with no tag name or synthetic id. This isn't a raised exception (unlike `expression_parsing`'s failures): it's non-fatal and can fire multiple times per response, so it stays a collection.
 - **`PhysicalQuantity` is a frozen result.** Its attributes can't be set (`dataclasses.FrozenInstanceError`); `value`, `unit` and `ast_root` are still ordinary parse trees. `parse_quantity` builds it through the new `PhysicalQuantity.from_tree(name, params, ast_root, parser)`, which runs each step (splitting value from unit, restoring the value's text, LaTeX, SI forms) as a separate function; the constructor now takes every attribute. Quantities still compare by identity, and pickle and deep-copy without a custom `__getstate__`/`__setstate__`.
@@ -29,7 +30,7 @@ The first public release. This is a further tidy-up after the 0.2.0 extraction:
 
 ### Added
 
-- DEBUG logging through module loggers (`lambdafeedback.units.*`), with a `NullHandler` on the package logger; nothing is logged at INFO or above. See the repository README. `parse_quantity` logs the normalised text and the value/unit split.
+- DEBUG logging through module loggers (`lambdafeedback.physical_quantities.*`), with a `NullHandler` on the package logger; nothing is logged at INFO or above. See the repository README. `parse_quantity` logs the normalised text and the value/unit split.
 - `DisallowedUnitError` (see Breaking).
 - `PhysicalQuantity.base_dimension`: the dimension without the unit's factor (`5 km/h` gives `length/time`, where `dimension` is `10**3*length/((3600*time))`), and `same_dimension(a, b)` to compare two quantities' dimensions.
 - `MAX_INPUT_LENGTH` (10,000 characters): `parse_quantity` raises `QuantityParseError` for longer input, and `preprocess_quantity` returns it unchanged. This bounds the running time on any input.
@@ -62,7 +63,7 @@ The first public release. This is a further tidy-up after the 0.2.0 extraction:
 
 ### Changed
 
-- `quantity` keeps the public `PhysicalQuantity`, `RevertedUnit`, `parse_quantity` and `same_dimension`; the steps of reading a quantity are in private modules (`_split`, `_latex`, `_conversion`). Import public names from `lambdafeedback.units`: `MAX_UNIT_EXPONENT` is no longer in `lambdafeedback.units.quantity`.
+- `quantity` keeps the public `PhysicalQuantity`, `RevertedUnit`, `parse_quantity` and `same_dimension`; the steps of reading a quantity are in private modules (`_split`, `_latex`, `_conversion`). Import public names from `lambdafeedback.physical_quantities`: `MAX_UNIT_EXPONENT` is no longer in `lambdafeedback.physical_quantities.quantity`.
 - The regular expressions in preprocessing and parsing are replaced by string scanners, with the same results. The quantity grammar's tokens are `slr_parsing.Lexeme`s (requires `lambdafeedback-slr-parsing` with literal token patterns).
 
 ### Deprecated