Skip to content

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

fix(import-sources): decode .txt sources instead of forcing UTF-8

hugohe339b470fe6e80
diff --git a/skills/ppt-master/scripts/project_management/cli.py b/skills/ppt-master/scripts/project_management/cli.py
index 9f2ac918..f2682141 100644
--- a/skills/ppt-master/scripts/project_management/cli.py
+++ b/skills/ppt-master/scripts/project_management/cli.py
@@ -553,8 +553,25 @@ class ProjectManager:
return file_path
def _normalize_text_source(self, source_path: Path, sources_dir: Path) -> Path:
+ from bs4 import UnicodeDammit
+
target = self._ensure_unique_path(sources_dir / f"{source_path.stem}.md")
- content = source_path.read_text(encoding="utf-8", errors="replace")
+ # UTF-8 first; a GB18030 / Shift-JIS note used to become U+FFFD silently.
+ raw = source_path.read_bytes()
+ dammit = UnicodeDammit(raw, user_encodings=["utf-8"])
+ content = dammit.unicode_markup
+ if content is None:
+ content = raw.decode("utf-8", errors="replace")
+ if (
+ dammit.unicode_markup is None
+ or dammit.contains_replacement_characters
+ or dammit.original_encoding not in ("utf-8", "ascii")
+ ):
+ print(
+ f"note: {source_path.name} decoded as {dammit.original_encoding or 'utf-8 (lossy)'}; "
+ "check the converted text",
+ file=sys.stderr,
+ )
target.write_text(content, encoding="utf-8")
return target
diff --git a/skills/ppt-master/scripts/tests/test_audit_r3_sources.py b/skills/ppt-master/scripts/tests/test_audit_r3_sources.py
index 1c02319c..1daf8eb5 100644
--- a/skills/ppt-master/scripts/tests/test_audit_r3_sources.py
+++ b/skills/ppt-master/scripts/tests/test_audit_r3_sources.py
@@ -633,6 +633,21 @@ class SameStemProjectImportTests(unittest.TestCase):
self.assertIn("22500", Path(summary["markdown"][0]).read_text(encoding="utf-8"))
+class TextSourceEncodingTests(unittest.TestCase):
+ def test_gb18030_text_source_is_decoded_not_replaced(self) -> None:
+ with tempfile.TemporaryDirectory() as tmp:
+ root = Path(tmp)
+ manager = project_cli.ProjectManager(base_dir=root / "projects")
+ with contextlib.redirect_stdout(io.StringIO()):
+ project = Path(manager.init_project("target", quick_generate=True))
+ source = root / "report.txt"
+ source.write_bytes("收入增长和利润说明".encode("gb18030"))
+ with contextlib.redirect_stdout(io.StringIO()), contextlib.redirect_stderr(io.StringIO()):
+ result = manager.import_sources(str(project), [str(source)], copy=True)
+ text = Path(result["markdown"][0]).read_text(encoding="utf-8")
+ self.assertIn("收入增长和利润说明", text)
+
+
class _CurlHeaders(dict):
"""Represent repeated curl headers without importing optional curl_cffi."""