Disclaimer: I am not a programmer. This issue report was generated by AI based on automated analysis of conversion output. All technical details below come from AI inspection of the calabash pipeline, not from my own expertise. Please bear with any inaccuracies — I'm a math teacher trying to convert exam papers, not a developer. 🙏
Description
When a Word .docx contains diagrams that combine vector graphics (curves, shaded regions, coordinate axes) with text boxes (axis labels like x, y, O), the calabash/docx2tex pipeline only preserves the text box content while dropping the graphical elements entirely.
Actual Output
Only the text box labels survive, appearing as orphaned {\centering ...\par} blocks without any \includegraphics:
\item[\raisebox{109pt}{(b)}]{\centering \textbf{\textit{x}}\par}
{\centering \textbf{\textit{O}}\par}
{\centering \textbf{\textit{y}}\par}
{\centering \textbf{\textit{y}}\textbf{ = }\textbf{\textit{xe}}\textsuperscript{\textit{x}}\par}
% ... curve and shaded region are entirely missingExpected Output
The entire diagram should be converted, either as a single \includegraphics{...} referencing an extracted image file, or the full graphical content preserved in some renderable form.
Impact
Affects math exam papers where coordinate diagrams (function graphs, area-under-curve illustrations) are integral to the questions. Without manual correction, these questions are unusable. About 14 occurrences found across 2 source documents.
Environment
- calabash/docx2tex: latest from transpect/docx2tex (https://github.com/transpect/docx2tex)
- Java: 8 (Temurin), also reproduced with Java 13
- Word documents: MathType equations + native Word diagrams
Description
When a Word
.docxcontains diagrams that combine vector graphics (curves, shaded regions, coordinate axes) with text boxes (axis labels likex,y,O), the calabash/docx2tex pipeline only preserves the text box content while dropping the graphical elements entirely.Actual Output
Only the text box labels survive, appearing as orphaned
{\centering ...\par}blocks without any\includegraphics:Expected Output
The entire diagram should be converted, either as a single \includegraphics{...} referencing an extracted image file, or the full graphical content preserved in some renderable form.
Impact
Affects math exam papers where coordinate diagrams (function graphs, area-under-curve illustrations) are integral to the questions. Without manual correction, these questions are unusable. About 14 occurrences found across 2 source documents.
Environment