Upgrade StorageFeature request: Save clean publisher HTML full text as an attachment
Could Zotero Connector offer an option to save a clean HTML version of an article’s full text, alongside the PDF or as a preferred attachment when available?
Zotero 10’s PDF Reading Mode is useful, but in my experience, the reconstructed layout can still contain misplaced content and awkward positioning, particularly in scientific papers with figures, equations, and multiple columns.
Many publishers already provide a full-text HTML version. That version already contains the article’s text, headings, figures, captions, tables, and equations in a structure intended for reading in a browser. It seems worth using that existing structure directly, instead of relying primarily on reconstructing a reading layout from the PDF.
For example:
https://ieeexplore.ieee.org/document/7517726
The difficulty with the publisher’s webpage is that the article is surrounded by navigation, recommendations, sharing controls, download buttons, and other interface elements. What I would like to save is the article itself.
Ideally, a “Save clean HTML full text” option would:
Preserve the title, authors, complete abstract, article sections, figures and captions, tables, equations, and references.
Remove surrounding website navigation, recommendations, action buttons, and duplicate or truncated abstract sections.
Include images and the resources needed to display equations, so the attachment remains readable offline.
Capture the fully loaded article, including dynamically loaded content.
Save the result under the normal bibliographic item, with the original URL retained.
I understand Zotero already has Reading Mode for webpage snapshots. This request is specifically about saving a clean, self-contained HTML attachment, rather than only simplifying the display of a full-page snapshot inside Zotero.
Publisher-specific cleanup would still be needed, but the publisher has already supplied the article’s structure and browser layout. Preserving that HTML could provide a better starting point for reading, annotation, and accessibility than reconstructing the same content from PDF.
Would this be feasible within the Connector and translator architecture, perhaps starting with selected academic publishers?
Related discussion:
https://forums.zotero.org/discussion/123142/website-snapshots-in-simplified-reader-view
Zotero 10’s PDF Reading Mode is useful, but in my experience, the reconstructed layout can still contain misplaced content and awkward positioning, particularly in scientific papers with figures, equations, and multiple columns.
Many publishers already provide a full-text HTML version. That version already contains the article’s text, headings, figures, captions, tables, and equations in a structure intended for reading in a browser. It seems worth using that existing structure directly, instead of relying primarily on reconstructing a reading layout from the PDF.
For example:
https://ieeexplore.ieee.org/document/7517726
The difficulty with the publisher’s webpage is that the article is surrounded by navigation, recommendations, sharing controls, download buttons, and other interface elements. What I would like to save is the article itself.
Ideally, a “Save clean HTML full text” option would:
Preserve the title, authors, complete abstract, article sections, figures and captions, tables, equations, and references.
Remove surrounding website navigation, recommendations, action buttons, and duplicate or truncated abstract sections.
Include images and the resources needed to display equations, so the attachment remains readable offline.
Capture the fully loaded article, including dynamically loaded content.
Save the result under the normal bibliographic item, with the original URL retained.
I understand Zotero already has Reading Mode for webpage snapshots. This request is specifically about saving a clean, self-contained HTML attachment, rather than only simplifying the display of a full-page snapshot inside Zotero.
Publisher-specific cleanup would still be needed, but the publisher has already supplied the article’s structure and browser layout. Preserving that HTML could provide a better starting point for reading, annotation, and accessibility than reconstructing the same content from PDF.
Would this be feasible within the Connector and translator architecture, perhaps starting with selected academic publishers?
Related discussion:
https://forums.zotero.org/discussion/123142/website-snapshots-in-simplified-reader-view