Output text matrix for diagnosis. This should not be merged to main.

stephen-byrne · Nov 1, 2023 · f6f5748 · f6f5748
1 parent e7d4549
commit f6f5748
Show file tree

Hide file tree

Showing 3 changed files with 13 additions and 4 deletions.
diff --git a/src/main/java/io/mfj/textricator/extractor/itext7/Itext7TextExtractor.kt b/src/main/java/io/mfj/textricator/extractor/itext7/Itext7TextExtractor.kt
@@ -21,6 +21,7 @@ import io.mfj.textricator.text.Text
 import java.io.InputStream
 
 import com.itextpdf.kernel.colors.*
+import com.itextpdf.kernel.geom.Matrix
 import com.itextpdf.kernel.geom.Vector
 import com.itextpdf.kernel.pdf.*
 import com.itextpdf.kernel.pdf.canvas.parser.EventType
@@ -123,16 +124,20 @@ class Itext7TextExtractor(input:InputStream):TextExtractor {
         val effMatrix = ri.textMatrix.multiply( ri.graphicsState.ctm )
         val fontSize = Vector(0f,ri.fontSize,0f).cross(effMatrix)[1]
 
+        val debug = "tm:${matrix.str()} ctm:${matrix.str()}"
+
         val text = Text(content = content, backgroundColor = null, pageNumber = pageNumber,
             fontSize = fontSize, font = font, color = color,
             ulx = ulx, uly = uly, lrx = lrx, lry = lry,
-            link = link )
+            link = link, debug = debug )
 
         texts.add( text )
       }
       super.eventOccurred(data, type)
     }
 
+    private fun Matrix.str():String = "[[${this[0]},${this[1]},${this[2]}],[${this[3]},${this[4]},${this[5]}],[${this[6]},${this[7]},${this[8]}]]"
+
     private fun Color.getHexColor(): String? =
         when ( this ) {
           is DeviceRgb -> getRgb( this )

diff --git a/src/main/java/io/mfj/textricator/text/Text.kt b/src/main/java/io/mfj/textricator/text/Text.kt
@@ -51,7 +51,9 @@ data class Text(
     val fontSize:Float,
     val color:String? = null,
     val backgroundColor:String? = null,
-    val link:String? = null ) {
+    val link:String? = null,
+    val debug:String? = null
+) {
 
   val width:Float
     @Transient

diff --git a/src/main/java/io/mfj/textricator/text/output/CsvTextOutput.kt b/src/main/java/io/mfj/textricator/text/output/CsvTextOutput.kt
@@ -64,7 +64,8 @@ class CsvTextOutput(output:OutputStream):TextOutput {
         "fontSize",
         "fontColor",
         "bgcolor",
-        "link"
+        "link",
+        "debug",
     )
   }
 
@@ -82,7 +83,8 @@ class CsvTextOutput(output:OutputStream):TextOutput {
         text.fontSize,
         text.color,
         text.backgroundColor,
-        text.link
+        text.link,
+        text.debug
     )
   }