Skip to contents

Glyph outline for a single glyph in a text page-object's font

Source: R/glyph_paths.R
pdf_glyph_path.Rd

Returns the path segments of the glyph rendered at font_size in PDF user-space points. Useful for:

Usage

pdf_glyph_path(obj, glyph_code, font_size = NA_real_)

Arguments

obj

A pdfium_obj of type "text".

glyph_code

Single non-negative integer; see the section above.

font_size

Numeric font size in PDF points. When NA (default), uses the text object's own font size — the most common choice when matching what is drawn on the page.

Value

A tibble with one row per glyph-path segment:

  • segment_index integer - 1-based.

  • segment_type character - "moveto", "lineto", "bezierto", or "unknown".

  • x, y numeric - point coordinates in PDF user-space points (the glyph's local coordinate system, scaled to the requested font_size).

  • close_figure logical - TRUE if this segment closes the current sub-path. Returns an empty tibble when PDFium reports no glyph outline.

Details

  • Reconstructing challenging character mappings — render the glyph at the character's reported unicode code point and compare to a reference rendering of that code point to see whether the font actually draws what its ToUnicode CMap claims.

  • Visualising the glyphs PDFium picked when extracting text.

  • Computing exact glyph silhouettes for layout / collision detection beyond what bounding boxes give you.

Wraps FPDFTextObj_GetFont -> FPDFFont_GetGlyphPath -> FPDFGlyphPath_CountGlyphSegments / FPDFGlyphPath_GetGlyphPathSegment.

Glyph code interpretation

glyph_code is the font's glyph identifier, not the unicode code point — though for many fonts they coincide:

  • TrueType fonts with /Identity-H encoding (most modern embedded CID-keyed fonts): glyph code equals unicode code point. Pass chars$codepoint from pdf_text_chars().

  • TrueType fonts with a cmap (e.g. WinAnsi or MacRoman encoding): glyph code is the encoded character code in the PDF stream, not the unicode value. The unicode <-> glyph map is opaque through the public PDFium API.

  • Type 1 fonts: glyph code is the encoding-specific character code (1-byte for almost all PDF Type 1 fonts).

If the path comes back empty, the glyph code likely doesn't map to a glyph in this font's encoding — try the character code from the source content stream (visible in tools like pdfinfo -text) instead.

See also

pdf_glyph_width(), pdf_text_font_metrics(), pdf_text_chars() for the per-character readout that drives most "investigate this glyph" workflows, pdf_text_obj_rendered_bitmap() when you want the rendered pixels instead of the outline.

Examples

if (FALSE) { # \dontrun{
doc <- pdf_doc_open("weird-font.pdf")
page <- pdf_page_load(doc, 1)
text_obj <- Filter(\(o) o$type == "text", pdf_page_objects(page))[[1]]
# First visible character on the page:
chars <- pdf_text_chars(page)
first <- chars[!chars$is_generated, ][1, ]
pdf_glyph_path(text_obj, first$codepoint)
} # }