PureBasic Requires Chilkat v11.7.0+

Extract Text from a Two-Column PDF

See more PDF Text Extraction Examples

Demonstrates how multi-column pages are read. By default the columns are detected and each column is read top to bottom before the next ("columns": "auto"); "columns": "none" reads the full page width line by line, which interleaves the columns. The quality report tells how many columns were found. PureBasic

IncludeFile "CkPdf.pb"
IncludeFile "CkJsonObject.pb"

Procedure ChilkatExample()

    success.i = 0

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

    ;  The two-column PDF used in this example is available at two_column_article.pdf

    pdf.i = CkPdf::ckCreate()
    If pdf.i = 0
        Debug "Failed to create object."
        ProcedureReturn
    EndIf

    success = CkPdf::ckLoadFile(pdf,"qa_data/pdf/two_column_article.pdf")
    If success = 0
        Debug CkPdf::ckLastErrorText(pdf)
        CkPdf::ckDispose(pdf)
        ProcedureReturn
    EndIf

    json.i = CkJsonObject::ckCreate()
    If json.i = 0
        Debug "Failed to create object."
        ProcedureReturn
    EndIf

    jsonOut.i = CkJsonObject::ckCreate()
    If jsonOut.i = 0
        Debug "Failed to create object."
        ProcedureReturn
    EndIf

    ;  With the default options the columns are detected, and the text of the left column
    ;  comes before the text of the right column.
    success = CkPdf::ckExtractText(pdf,"text",json,jsonOut)
    If success = 0
        Debug CkPdf::ckLastErrorText(pdf)
        CkPdf::ckDispose(pdf)
        CkJsonObject::ckDispose(json)
        CkJsonObject::ckDispose(jsonOut)
        ProcedureReturn
    EndIf

    Debug "Columns detected: " + Str(CkJsonObject::ckIntOf(jsonOut,"quality.pages[0].columns"))
    Debug "Reading order: " + CkJsonObject::ckStringOf(jsonOut,"quality.pages[0].readingOrderSource")
    Debug CkJsonObject::ckStringOf(jsonOut,"text")

    ;  "none" switches column detection off: the page is read across its full width, line by line,
    ;  so each output line holds a line of the left column followed by a line of the right column.
    ;  This is rarely wanted, but it shows what the detection does.
    CkJsonObject::ckUpdateString(json,"columns","none")
    CkJsonObject::ckUpdateString(json,"layout","lines")
    success = CkPdf::ckExtractText(pdf,"text",json,jsonOut)
    If success = 0
        Debug CkPdf::ckLastErrorText(pdf)
        CkPdf::ckDispose(pdf)
        CkJsonObject::ckDispose(json)
        CkJsonObject::ckDispose(jsonOut)
        ProcedureReturn
    EndIf

    Debug "---- columns: none ----"
    Debug CkJsonObject::ckStringOf(jsonOut,"text")

    ;  Sample output (abbreviated):

    ;  Columns detected: 2
    ;  Reading order: visual
    ;  The Harbor Lighthouse, Season Two
    ;  
    ;  The harbor lighthouse was built in 1874 to mark the entrance to the inner channel. It operated continuously until 1968, ...
    ;  
    ;  The restoration society took ownership in 2021 with three goals: ...
    ;  ...
    ;  The lantern room frame was stripped, primed and repainted in the original dark green. ...
    ;  ...
    ;  ---- columns: none ----
    ;  The Harbor Lighthouse, Season Two
    ;  The harbor lighthouse was built in 1874 to The lantern room frame was stripped, primed
    ;  mark the entrance to the inner channel. It and repainted in the original dark green. New
    ;  ...


    CkPdf::ckDispose(pdf)
    CkJsonObject::ckDispose(json)
    CkJsonObject::ckDispose(jsonOut)


    ProcedureReturn
EndProcedure