Tcl Requires Chilkat v11.7.0+

Expand Ligatures and Normalize Unicode in Extracted PDF Text

See more PDF Text Extraction Examples

Demonstrates the "expandLigatures" and "normalize" clean-up options. Typeset PDFs use single glyphs for "fi" and "fl"; by default they are expanded to their letters. Unicode normalization is NFC by default; NFKC also folds compatibility characters such as circled digits, the trademark sign and vulgar fractions into plain characters, and "none" leaves the text as the PDF has it. Tcl

load ./chilkat.dll

set success 0

#  This example requires the Chilkat API to have been previously unlocked.
#  See Global Unlock Sample for sample code.

#  The PDF used in this example is available at ligatures.pdf

set pdf [new_CkPdf]

set success [CkPdf_LoadFile $pdf "qa_data/pdf/ligatures.pdf"]
if {$success == 0} then {
    puts [CkPdf_lastErrorText $pdf]
    delete_CkPdf $pdf
    exit
}

set json [new_CkJsonObject]

set jsonOut [new_CkJsonObject]

#  The first line is set with the fi and fl ligature glyphs; the second has circled digits,
#  a trademark sign and a vulgar fraction.
set success [CkPdf_ExtractText $pdf "text" $json $jsonOut]
if {$success == 0} then {
    puts [CkPdf_lastErrorText $pdf]
    delete_CkPdf $pdf
    delete_CkJsonObject $json
    delete_CkJsonObject $jsonOut
    exit
}

puts "---- default (expandLigatures: true, normalize: NFC) ----"
puts [CkJsonObject_stringOf $jsonOut text]

#  Keep the ligature characters as they are in the PDF.
CkJsonObject_UpdateBool $json "expandLigatures" 0
CkJsonObject_UpdateString $json "normalize" "none"
set success [CkPdf_ExtractText $pdf "text" $json $jsonOut]
if {$success == 0} then {
    puts [CkPdf_lastErrorText $pdf]
    delete_CkPdf $pdf
    delete_CkJsonObject $json
    delete_CkJsonObject $jsonOut
    exit
}

puts "---- expandLigatures: false, normalize: none ----"
puts [CkJsonObject_stringOf $jsonOut text]

#  NFKC folds the compatibility characters too (and expands ligatures on its own).
CkJsonObject_UpdateString $json "normalize" "NFKC"
set success [CkPdf_ExtractText $pdf "text" $json $jsonOut]
if {$success == 0} then {
    puts [CkPdf_lastErrorText $pdf]
    delete_CkPdf $pdf
    delete_CkJsonObject $json
    delete_CkJsonObject $jsonOut
    exit
}

puts "---- normalize: NFKC ----"
puts [CkJsonObject_stringOf $jsonOut text]

#  Sample output (shown here in ASCII: (fi) and (fl) stand for the single ligature characters, (1) and (2) for the circled digits,
#  (TM) for the trademark sign and 1/2 for the vulgar fraction; the NFKC output has plain "1", "2", "TM" and "1", fraction slash, "2"):

#  ---- default (expandLigatures: true, normalize: NFC) ----
#  Typeset Text
#  
#  The final flight of the season was officially confirmed on the fifth. Rooms (1) and (2) are reserved; the Harbor Cafe(TM) serves 1/2 portions.
#  ---- expandLigatures: false, normalize: none ----
#  Typeset Text
#  
#  The (fi)nal (fl)ight of the season was of(fi)cially con(fi)rmed on the (fi)fth. Rooms (1) and (2) are reserved; the Harbor Cafe(TM) serves 1/2 portions.
#  ---- normalize: NFKC ----
#  Typeset Text
#  
#  The final flight of the season was officially confirmed on the fifth. Rooms 1 and 2 are reserved; the Harbor CafeTM serves 1/2 portions.

delete_CkPdf $pdf
delete_CkJsonObject $json
delete_CkJsonObject $jsonOut