OCTO-11541 RP 2052-10 alignment preservation + codebase readability refactoring

ai_artifacts/project_understanding.md

@@ -60,7 +60,7 @@ pycaption/
 │   ├── srt.py                    # SRT format (reader + writer in single file)
 │   ├── microdvd.py               # MicroDVD format (reader + writer)
 │   └── transcript.py             # Plain text transcript (writer only, uses nltk)
-├── tests/                        # Test suite (504 tests)
+├── tests/                        # Test suite (505 tests)
 │   ├── conftest.py               # Fixture imports (re-exports from fixtures/)
 │   ├── fixtures/                 # Pytest fixture modules (inline caption strings)
 │   │   ├── scc.py, translated_scc.py, dfxp.py, webvtt.py, srt.py, sami.py, microdvd.py
@@ -217,6 +217,10 @@ The SAMI reader uses a two-phase approach:
    overflow (not 100% as might be expected)
 6. **Style filtering**: Writers filter internal keys (e.g. `classes`, `webvtt_positioning`) from
    output via format-specific exclusion sets
+7. **RP 2052-10 implicit defaults**: When source format's visual default differs from
+   target format's default, writers explicitly emit the source's alignment. VTT/SRT/SCC
+   default to center; DFXP/SAMI default to left/start. The `is_positional_anchor` flag
+   on Layout distinguishes SCC coordinate-system positioning from visual alignment intent.
 
 ---
 
@@ -286,7 +290,7 @@ Additional workflows for format spec compliance:
 ### Running Tests Locally
 
 ```bash
-# Run all tests (504 tests, ~0.5s)
+# Run all tests (505 tests, ~0.5s)
 python -m pytest tests/ -q
 
 # Run specific format tests
@@ -299,7 +303,7 @@ python -m pytest tests/ -v
 python -m pytest tests/test_webvtt_conversion.py -q
 ```
 
-- Tests cover all formats (504 tests total)
+- Tests cover all formats (505 tests total)
 - Fixtures defined in `tests/fixtures/` as pytest session-scoped fixtures
 - Each fixture is an inline string containing a complete caption file
 - Conversion tests verify round-trip and cross-format fidelity

docs/changelog.rst

@@ -2,15 +2,38 @@ Changelog
 ---------
 2.3.2
 ^^^^^^
-  - DFXP writer: output ``textAlign="center"`` (not ``"start"``) as the
-    fallback alignment for cues that carry no explicit alignment.
-    Separates the writer's fallback (CENTER) from the reader's default
-    region (START) via a new ``DFXP_WRITER_FALLBACK_ALIGNMENT`` constant.
-
-  - SAMI reader: apply the SAMI spec default ``text-align: left`` only at
-    the root layout level (no parent to inherit from). Child layouts
-    without an explicit text-align now correctly inherit from their parent
-    stylesheet rather than being forced to left.
+  - RP 2052-10 compliance for all 25 conversion paths: when source format's
+    visual default is CENTER (VTT/SRT/SCC) and target default is LEFT/START
+    (DFXP/SAMI), explicitly emit center alignment. DFXP/SAMI sources retain
+    their original alignment for round-trip fidelity.
+
+  - DFXP writer: suppress ``tts:origin`` from SCC positional layouts
+    (row/column anchors), emit ``tts:textAlign="center"`` instead.
+
+  - WebVTT writer: suppress SCC positional cue settings (``align:left
+    position:N% line:N% size:N%``). SCC coordinate anchors are not visual
+    alignment — omitting lets VTT default to center.
+
+  - SAMI reader: apply ``text-align: left`` only at the root layout level;
+    child layouts now inherit from parent rather than being forced to left.
+
+  - SAMI writer: emit semantic tags (``<i>``, ``<b>``, ``<u>``) instead of
+    ``<span style="...">``. Add trailing ``&nbsp;`` sync to clear the final
+    caption at its end time.
+
+  - Geometry: add ``is_positional_anchor`` flag to ``Layout`` to distinguish
+    SCC coordinate-system positioning from visual text alignment.
+
+  - Alignment preservation now uses an explicit ``visual_alignment_default``
+    attribute on ``CaptionSet`` (set by each reader) rather than inferring
+    intent from missing layout objects. Writers compare source vs target
+    defaults via ``BaseWriter._get_visual_alignment_default()``.
+
+  - Codebase-wide readability pass: dict lookups replace if/elif chains,
+    ``@total_ordering`` on ``Size``, ``__eq__`` returns ``NotImplemented``,
+    writer signatures unified to ``**kwargs``, dead code removed
+    (``NodeCreatorFactory``, ``Caption.is_empty``, ``Layout.__ne__``).
+    Net −105 lines across 17 files, no behavioral changes.
 
 2.3.1
 ^^^^^^

docs/conf.py

@@ -53,9 +53,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = "2.3.1"
+version = "2.3.2.dev1"
 # The full version, including alpha/beta/rc tags.
-release = "2.3.1"
+release = "2.3.2.dev1"
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.