Copy-paste-able/searchable PDF files

PDF files generated from TeX (and friends), will by default hold their text in the encoding of the original TeX font used by the document.

When PDF readers, etc., offer copy-paste or searching functions, the operations take place on the glyph codes used for the fonts selected by the document. This is fine, for the simplest documents (in English, at least); the problem comes when you're using an inflected language (with accented letters, or composite glyphs such as 'æ') — TeX will typically use a non-standard encoding, and there are likely be problems, since PDF readers assume the text is presented in Unicode.

For PDF generated from LaTeX (the DVI being converted, by whatever means), or from pdfLaTeX, the character codes used in the PDF file are in fact those of the document's font encoding; if you're using OT1 or T1, your document will be OK for almost all ASCII characters, but it's likely that anything “out of the ordinary” will not be represented properly. (Of course, PDF generated from XeTeX- or LuaTeX-based formats is going to be OK, since those engines work in Unicode throughout.)

The solution comes from the character-mapping facilities in the PDF specification: the file may specify a table of translations of characters present in the coding used in the file, to a Unicode version of the characters.

Packages cmap and mmap both offer means of generating such tables (mmap has wider coverage, including the various maths encodings); both work with pdfTeX and no other engine. Thus your document becomes something like:

\usepackage{mmap} % (or cmap)
... % your other packages
... % your actual text

Unfortunately, the packages only work with fonts that are directly encoded, such as the default (Computer Modern, i.e., cm fonts, and things such as cm-super or the Latin Modern sets. Fonts like Adobe Times Roman (which are encoded for (La)TeX use via virtual fonts) are not amenable to this treatment.

Source: Copy-paste-able/searchable PDF files

5_fichiers/pdf/produire_des_pdf_cherchables.txt · Dernière modification : 2018/12/04 00:22 de jejust
CC Attribution-Share Alike 4.0 International
Driven by DokuWiki Recent changes RSS feed Valid CSS Valid XHTML 1.0