feat(format): B2 — PPTX placeholder filter, SmartArt diagrams, chart XPath matching

tests/test_translators/test_b2_pptx_fixes.py

@@ -0,0 +1,623 @@
+"""
+Tests for Track B2 — PowerPoint format preservation fixes.
+
+Covers:
+  B2.1 — Placeholder filter (DATE=16, SLIDE_NUMBER=13, FOOTER=15)
+         PowerPoint auto-fills these at render time, so translating them
+         is pointless and just wastes API calls.
+  B2.2 — SmartArt diagram text collection (ppt/diagrams/data*.xml)
+  B2.3 — SmartArt diagram translation applied via ZIP rewrite
+  B2.4 — Chart text matched by element path (not string equality)
+  B2.5 — Whitespace preservation in chart apply
+
+These tests are intentionally narrow: they validate that a specific
+bug is fixed. Existing tests in test_pptx_translator.py handle the
+broader surface.
+"""
+import os
+import tempfile
+import zipfile
+from pathlib import Path
+from typing import List, Dict
+from unittest.mock import MagicMock, patch
+
+import pytest
+from pptx import Presentation
+from pptx.util import Inches, Pt
+from pptx.enum.shapes import MSO_SHAPE_TYPE
+from lxml import etree
+
+from translators.pptx_translator import PowerPointTranslator, pptx_translator
+
+
+# ---------------- Mock provider ----------------
+
+class MockProvider:
+    """Mock translation provider: returns 'TR_<text>' unless a mapping
+    is provided."""
+
+    def __init__(self, translations: Dict[str, str] = None):
+        self._translations = translations or {}
+        self.calls: List[str] = []
+
+    def translate_batch(
+        self,
+        texts: List[str],
+        target_language: str = "fr",
+        source_language: str = "auto",
+    ) -> List[str]:
+        out = []
+        for t in texts:
+            self.calls.append(t)
+            if t in self._translations:
+                out.append(self._translations[t])
+            elif t.startswith("TR_"):
+                out.append(t)
+            else:
+                out.append(t)
+        return out
+
+    def get_name(self):
+        return "mock"
+
+    def is_available(self):
+        return True
+
+
+# ---------------- Helpers ----------------
+
+_NS_A = "http://schemas.openxmlformats.org/drawingml/2006/main"
+_NS_C = "http://schemas.openxmlformats.org/drawingml/2006/chart"
+_NS_P = "http://schemas.openxmlformats.org/presentationml/2006/main"
+
+
+def _add_textbox_with_text(slide, text: str):
+    tb = slide.shapes.add_textbox(Inches(1), Inches(1), Inches(3), Inches(1))
+    tf = tb.text_frame
+    tf.text = text
+    return tb
+
+
+def _inject_diagram_part(pptx_path: Path, diagram_filename: str, diagram_xml: bytes):
+    """Inject a SmartArt diagram XML part into a .pptx file."""
+    tmp_out = pptx_path.with_suffix(".tmp")
+    with zipfile.ZipFile(pptx_path, "r") as zin, \
+         zipfile.ZipFile(tmp_out, "w", zipfile.ZIP_DEFLATED) as zout:
+        existing = {item: zin.read(item) for item in zin.namelist()}
+
+        # Add the diagram data part
+        existing[diagram_filename] = diagram_xml
+
+        # Register in [Content_Types].xml
+        ct_path = "[Content_Types].xml"
+        if ct_path in existing:
+            ct_content = existing[ct_path].decode("utf-8")
+            override = (
+                f'<Override PartName="/{diagram_filename}" '
+                f'ContentType="application/vnd.openxmlformats-officedocument.'
+                f'drawingml+xml+diagram"/>'
+            )
+            if override not in ct_content:
+                ct_content = ct_content.replace("</Types>", f"{override}</Types>")
+            existing[ct_path] = ct_content.encode("utf-8")
+
+        for item, data in existing.items():
+            zout.writestr(item, data)
+
+    tmp_out.replace(pptx_path)
+
+
+def _inject_chart_part(pptx_path: Path, chart_filename: str, chart_xml: bytes):
+    """Inject a chart XML part into a .pptx file."""
+    tmp_out = pptx_path.with_suffix(".tmp")
+    with zipfile.ZipFile(pptx_path, "r") as zin, \
+         zipfile.ZipFile(tmp_out, "w", zipfile.ZIP_DEFLATED) as zout:
+        existing = {item: zin.read(item) for item in zin.namelist()}
+
+        existing[chart_filename] = chart_xml
+
+        ct_path = "[Content_Types].xml"
+        if ct_path in existing:
+            ct_content = existing[ct_path].decode("utf-8")
+            override = (
+                f'<Override PartName="/{chart_filename}" '
+                f'ContentType="application/vnd.openxmlformats-officedocument.'
+                f'drawingml+xml+chart"/>'
+            )
+            if override not in ct_content:
+                ct_content = ct_content.replace("</Types>", f"{override}</Types>")
+            existing[ct_path] = ct_content.encode("utf-8")
+
+        for item, data in existing.items():
+            zout.writestr(item, data)
+
+    tmp_out.replace(pptx_path)
+
+
+# ============================================================================
+# B2.1 — Placeholder filter (DATE / SLIDE_NUMBER / FOOTER)
+# ============================================================================
+
+class TestPptxPlaceholderFilter:
+    """Bug fix: PowerPoint auto-generates DATE, SLIDE_NUMBER, and FOOTER
+    placeholders at render time. Translating them wastes API calls
+    because the translated value gets overwritten on next render."""
+
+    def _build_pptx_with_placeholders(self, tmp_path: Path) -> Path:
+        """Build a PPTX with a date placeholder, a slide-number placeholder,
+        a footer placeholder, plus one real text box. Returns the path."""
+        from pptx import Presentation
+
+        prs = Presentation()
+        slide_layout = prs.slide_layouts[6]  # blank
+        slide = prs.slides.add_slide(slide_layout)
+
+        # Add real text — should be translated
+        _add_textbox_with_text(slide, "Real text")
+
+        # python-pptx doesn't expose a direct "add footer/date/sldNum
+        # placeholder" API, so we inject the XML directly. The placeholder
+        # types that should be filtered:
+        #   13 (sldNum), 15 (ftr), 16 (dt)
+        from pptx.oxml.ns import qn
+        from lxml import etree as _etree
+
+        spTree = slide.shapes._spTree
+        nvSpPr_template = """
+        <p:sp xmlns:p="http://schemas.openxmlformats.org/presentationml/2006/main"
+              xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+          <p:nvSpPr>
+            <p:nvPr>
+              <p:ph type="{ph_type}" idx="1"/>
+            </p:nvPr>
+            <p:cNvPr id="{shape_id}" name="PH{shape_id}"/>
+            <p:cNvSpPr/>
+          </p:nvSpPr>
+          <p:spPr>
+            <a:xfrm>
+              <a:off x="0" y="0"/>
+              <a:ext cx="1000000" cy="500000"/>
+            </a:xfrm>
+          </p:spPr>
+          <p:txBody>
+            <a:bodyPr/>
+            <a:lstStyle/>
+            <a:p>
+              <a:r>
+                <a:rPr lang="en-US"/>
+                <a:t>{text}</a:t>
+              </a:r>
+            </a:p>
+          </p:txBody>
+        </p:sp>
+        """
+
+        # Footer placeholder (type 15)
+        sp_xml = nvSpPr_template.format(
+            ph_type="ftr",
+            shape_id=100,
+            text="Footer auto",
+        )
+        sp_elem = _etree.fromstring(sp_xml)
+        spTree.append(sp_elem)
+
+        # Date placeholder (type 16)
+        sp_xml = nvSpPr_template.format(
+            ph_type="dt",
+            shape_id=101,
+            text="Date auto",
+        )
+        sp_elem = _etree.fromstring(sp_xml)
+        spTree.append(sp_elem)
+
+        # Slide number placeholder (type 13)
+        sp_xml = nvSpPr_template.format(
+            ph_type="sldNum",
+            shape_id=102,
+            text="Slide number auto",
+        )
+        sp_elem = _etree.fromstring(sp_xml)
+        spTree.append(sp_elem)
+
+        path = tmp_path / "placeholders.pptx"
+        prs.save(path)
+        return path
+
+    def test_placeholders_skipped(self, tmp_path):
+        """The three auto-generated placeholder types should NOT be sent
+        to the translation provider."""
+        provider = MockProvider({"Real text": "TR_Real text"})
+        translator = PowerPointTranslator(provider=provider)
+
+        input_file = self._build_pptx_with_placeholders(tmp_path)
+        output_file = tmp_path / "output.pptx"
+
+        translator.translate_file(input_file, output_file, "fr")
+
+        # Real text should be translated
+        assert "Real text" in provider.calls, (
+            f"Real text missing from provider calls: {provider.calls}"
+        )
+
+        # Auto-generated placeholders should NOT be in the calls
+        assert "Footer auto" not in provider.calls, (
+            f"Footer placeholder was sent to provider: {provider.calls}"
+        )
+        assert "Date auto" not in provider.calls, (
+            f"Date placeholder was sent to provider: {provider.calls}"
+        )
+        assert "Slide number auto" not in provider.calls, (
+            f"Slide number placeholder was sent to provider: {provider.calls}"
+        )
+
+
+# ============================================================================
+# B2.2 — SmartArt diagram text collection
+# ============================================================================
+
+class TestPptxDiagramTextCollection:
+    """New feature: SmartArt text (in ppt/diagrams/data*.xml) is now
+    collected and translated."""
+
+    DIAGRAM_XML = f"""<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<dgm:dataModel xmlns:dgm="http://schemas.openxmlformats.org/drawingml/2006/diagram"
+               xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main"
+               xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships">
+  <dgm:ptLst>
+    <dgm:pt modelId="0" type="doc">
+      <dgm:prSet loTypeId="urn:microsoft.com/office/officeart/2005/8/layout/hierarchy1" loCatId="hierarchy"/>
+      <dgm:t>
+        <a:bodyPr/>
+        <a:lstStyle/>
+        <a:p>
+          <a:r>
+            <a:rPr lang="en-US"/>
+            <a:t>Root node</a:t>
+          </a:r>
+        </a:p>
+      </dgm:t>
+    </dgm:pt>
+    <dgm:pt modelId="1">
+      <dgm:t>
+        <a:bodyPr/>
+        <a:lstStyle/>
+        <a:p>
+          <a:r>
+            <a:rPr lang="en-US"/>
+            <a:t>Child one</a:t>
+          </a:r>
+        </a:p>
+      </dgm:t>
+    </dgm:pt>
+    <dgm:pt modelId="2">
+      <dgm:t>
+        <a:bodyPr/>
+        <a:lstStyle/>
+        <a:p>
+          <a:r>
+            <a:rPr lang="en-US"/>
+            <a:t>Child two</a:t>
+          </a:r>
+        </a:p>
+      </dgm:t>
+    </dgm:pt>
+  </dgm:ptLst>
+</dgm:dataModel>
+"""
+
+    def _build_pptx_with_diagram(self, tmp_path: Path) -> Path:
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[6])
+        _add_textbox_with_text(slide, "Hello")
+        path = tmp_path / "with_diagram.pptx"
+        prs.save(path)
+        _inject_diagram_part(
+            path, "ppt/diagrams/data1.xml",
+            self.DIAGRAM_XML.encode("utf-8"),
+        )
+        return path
+
+    def test_diagram_text_collected(self, tmp_path):
+        """Diagram text should appear in provider calls."""
+        provider = MockProvider()
+        translator = PowerPointTranslator(provider=provider)
+
+        input_file = self._build_pptx_with_diagram(tmp_path)
+        output_file = tmp_path / "output.pptx"
+
+        translator.translate_file(input_file, output_file, "fr")
+
+        # All three diagram nodes should be collected
+        assert "Root node" in provider.calls, (
+            f"Diagram 'Root node' not collected. Calls: {provider.calls}"
+        )
+        assert "Child one" in provider.calls, (
+            f"Diagram 'Child one' not collected. Calls: {provider.calls}"
+        )
+        assert "Child two" in provider.calls, (
+            f"Diagram 'Child two' not collected. Calls: {provider.calls}"
+        )
+
+    def test_diagram_text_applied(self, tmp_path):
+        """Translated diagram text should be written back to the .pptx
+        ZIP (in ppt/diagrams/data*.xml)."""
+        provider = MockProvider({
+            "Root node": "TR_Root node",
+            "Child one": "TR_Child one",
+            "Child two": "TR_Child two",
+        })
+        translator = PowerPointTranslator(provider=provider)
+
+        input_file = self._build_pptx_with_diagram(tmp_path)
+        output_file = tmp_path / "output.pptx"
+
+        translator.translate_file(input_file, output_file, "fr")
+
+        # The diagram part in the output ZIP should have the translated text
+        with zipfile.ZipFile(output_file, "r") as zf:
+            assert "ppt/diagrams/data1.xml" in zf.namelist(), (
+                f"Diagram part missing in output. Files: {zf.namelist()}"
+            )
+            data = zf.read("ppt/diagrams/data1.xml").decode("utf-8")
+        assert "TR_Root node" in data, (
+            f"Diagram 'Root node' not translated in output:\n{data[:1000]}"
+        )
+        assert "TR_Child one" in data, (
+            f"Diagram 'Child one' not translated in output"
+        )
+        assert "TR_Child two" in data, (
+            f"Diagram 'Child two' not translated in output"
+        )
+
+    def test_no_diagram_no_error(self, tmp_path):
+        """A PPTX without a diagram part should not raise."""
+        provider = MockProvider({"Hello": "TR_Hello"})
+        translator = PowerPointTranslator(provider=provider)
+
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[6])
+        _add_textbox_with_text(slide, "Hello")
+        input_file = tmp_path / "no_diagram.pptx"
+        output_file = tmp_path / "output.pptx"
+        prs.save(input_file)
+
+        # Should not raise
+        translator.translate_file(input_file, output_file, "fr")
+
+
+# ============================================================================
+# B2.3 — Diagram whitespace preservation
+# ============================================================================
+
+class TestPptxDiagramWhitespace:
+    """Whitespace around translated diagram text should be preserved."""
+
+    DIAGRAM_PADDED_XML = f"""<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<dgm:dataModel xmlns:dgm="http://schemas.openxmlformats.org/drawingml/2006/diagram"
+               xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+  <dgm:ptLst>
+    <dgm:pt modelId="0">
+      <dgm:t>
+        <a:bodyPr/>
+        <a:lstStyle/>
+        <a:p>
+          <a:r>
+            <a:rPr lang="en-US"/>
+            <a:t>  Padded text  </a:t>
+          </a:r>
+        </a:p>
+      </dgm:t>
+    </dgm:pt>
+  </dgm:ptLst>
+</dgm:dataModel>
+"""
+
+    def test_padded_diagram_text_preserved(self, tmp_path):
+        provider = MockProvider({"Padded text": "TR_Padded text"})
+        translator = PowerPointTranslator(provider=provider)
+
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[6])
+        input_file = tmp_path / "padded.pptx"
+        prs.save(input_file)
+        _inject_diagram_part(
+            input_file, "ppt/diagrams/data1.xml",
+            self.DIAGRAM_PADDED_XML.encode("utf-8"),
+        )
+        output_file = tmp_path / "output.pptx"
+
+        translator.translate_file(input_file, output_file, "fr")
+
+        with zipfile.ZipFile(output_file, "r") as zf:
+            data = zf.read("ppt/diagrams/data1.xml").decode("utf-8")
+        # Leading/trailing spaces should be preserved
+        assert "  TR_Padded text  " in data, (
+            f"Whitespace not preserved in diagram translation:\n{data[:1000]}"
+        )
+
+
+# ============================================================================
+# B2.4 — Element path navigation
+# ============================================================================
+
+class TestPptxElementPathNavigation:
+    """Bug fix: chart/diagram elements should be matched by stored
+    element_path, not by string equality. This means identical text
+    in multiple elements (e.g. two 'Revenue' series) is handled correctly."""
+
+    SIMPLE_DIAGRAM = f"""<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<root xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+  <a:t>Alpha</a:t>
+  <a:t>Beta</a:t>
+  <a:t>Gamma</a:t>
+</root>
+"""
+
+    def test_path_navigation_basic(self):
+        """The `_get_element_path` / `_find_element_by_path` pair should
+        round-trip correctly through a small XML tree."""
+        root = etree.fromstring(self.SIMPLE_DIAGRAM.encode("utf-8"))
+        translator = PowerPointTranslator()
+
+        # Get paths for all three <a:t> elements
+        elements = list(root.iter(f"{{{_NS_A}}}t"))
+        assert len(elements) == 3
+
+        paths = [translator._get_element_path(e) for e in elements]
+
+        # All three paths should be unique
+        assert len(set(paths)) == 3, f"Paths not unique: {paths}"
+
+        # Each path should resolve back to the same element
+        for elem, path in zip(elements, paths):
+            found = translator._find_element_by_path(root, path)
+            assert found is not None, f"Path didn't resolve: {path}"
+            assert found.text == elem.text, (
+                f"Path {path} resolved to wrong element: "
+                f"{found.text!r} vs {elem.text!r}"
+            )
+
+    def test_path_fallback_to_string_match(self):
+        """If the stored path is empty, the apply step should fall back
+        to string equality on `entry['original']`."""
+        root = etree.fromstring(self.SIMPLE_DIAGRAM.encode("utf-8"))
+        translator = PowerPointTranslator()
+
+        # Simulate a legacy entry with no element_path
+        legacy_entry = {
+            "original": "Beta",
+            "translated": "TR_Beta",
+            "element_path": None,  # legacy
+        }
+
+        # Simulate the apply logic
+        target = None
+        for candidate in root.iter(f"{{{_NS_A}}}t"):
+            if (candidate.text or "").strip() == legacy_entry["original"]:
+                target = candidate
+                break
+        assert target is not None, "Fallback string match didn't find Beta"
+        assert target.text == "Beta"
+
+    def test_diagram_with_duplicate_text_paths_differ(self, tmp_path):
+        """Two diagram nodes with the same text should have different
+        element_path values, so the apply step can target them correctly."""
+        dup_xml = f"""<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<dgm:dataModel xmlns:dgm="http://schemas.openxmlformats.org/drawingml/2006/diagram"
+               xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+  <dgm:ptLst>
+    <dgm:pt modelId="0">
+      <dgm:t><a:p><a:r><a:t>Revenue</a:t></a:r></a:p></dgm:t>
+    </dgm:pt>
+    <dgm:pt modelId="1">
+      <dgm:t><a:p><a:r><a:t>Revenue</a:t></a:r></a:p></dgm:t>
+    </dgm:pt>
+  </dgm:ptLst>
+</dgm:dataModel>
+"""
+        provider = MockProvider()
+        translator = PowerPointTranslator(provider=provider)
+
+        prs = Presentation()
+        slide = prs.slides.add_slide(prs.slide_layouts[6])
+        input_file = tmp_path / "dup.pptx"
+        prs.save(input_file)
+        _inject_diagram_part(
+            input_file, "ppt/diagrams/data1.xml",
+            dup_xml.encode("utf-8"),
+        )
+        output_file = tmp_path / "output.pptx"
+
+        # Should not raise; both should be translated independently
+        translator.translate_file(input_file, output_file, "fr")
+
+        # Both 'Revenue' texts should be in the calls
+        assert provider.calls.count("Revenue") == 2, (
+            f"Expected 2 'Revenue' calls, got {provider.calls.count('Revenue')}. "
+            f"All calls: {provider.calls}"
+        )
+
+
+# ============================================================================
+# B2.5 — Chart whitespace preservation
+# ============================================================================
+
+class TestPptxChartWhitespace:
+    """Whitespace around translated chart text should be preserved."""
+
+    CHART_PADDED_XML = f"""<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<c:chartSpace xmlns:c="http://schemas.openxmlformats.org/drawingml/2006/chart"
+              xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+  <c:chart>
+    <c:title>
+      <c:tx>
+        <c:rich>
+          <a:p>
+            <a:r>
+              <a:t>  Padded chart title  </a:t>
+            </a:r>
+          </a:p>
+        </c:rich>
+      </c:tx>
+    </c:title>
+  </c:chart>
+</c:chartSpace>
+"""
+
+    def test_padded_chart_text_via_internal_method(self):
+        """The internal chart apply logic should preserve whitespace."""
+        provider = MockProvider({"Padded chart title": "Titre avec espaces"})
+        translator = PowerPointTranslator(provider=provider)
+
+        # Build a chart entry by hand (simulating collect time)
+        chart_xml = etree.fromstring(self.CHART_PADDED_XML.encode("utf-8"))
+        entries = []
+        for t_elem in chart_xml.iter(f"{{{_NS_A}}}t"):
+            text_raw = t_elem.text or ""
+            text = text_raw.strip()
+            if not text:
+                continue
+            entry = {
+                "element": t_elem,
+                "original": text,
+                "original_raw": text_raw,
+                "translated": "Titre avec espaces",
+                "tag": "a:t",
+                "element_path": translator._get_element_path(t_elem),
+            }
+            entries.append(entry)
+
+        if not hasattr(translator, "_chart_entries"):
+            translator._chart_entries = []
+
+        class _FakePart:
+            def __init__(self, blob):
+                self._blob = blob
+            @property
+            def blob(self):
+                return self._blob
+            @blob.setter
+            def blob(self, value):
+                self._blob = value
+
+        fake_part = _FakePart(
+            etree.tostring(
+                chart_xml, xml_declaration=True, encoding="UTF-8", standalone=True
+            )
+        )
+        translator._chart_entries.append({
+            "chart_part": fake_part,
+            "entries": entries,
+        })
+
+        # Apply
+        translator._apply_chart_translations(Path("dummy"))
+
+        # Re-parse and check whitespace preserved
+        updated_xml = etree.fromstring(fake_part._blob)
+        all_t = list(updated_xml.iter(f"{{{_NS_A}}}t"))
+        # Find the title text
+        title_text = all_t[0].text or ""
+        assert "  Titre avec espaces  " in title_text, (
+            f"Chart whitespace not preserved: {title_text!r}"
+        )