Merge pull request #163 from bertsky/add-gcv-alto

add PRImA converter for GCV→ALTO

README.md

@@ -156,6 +156,7 @@ ocr-transform [OPTIONS] -L|--list               List available from/to, and exit
         alto4.2 alto2.1
         alto page
         alto text
+        gcv alto
         gcv hocr
         gcv page
         hocr alto2.0
@@ -191,7 +192,7 @@ capable stylesheet transformer.
 | ALTO                | ✓    | ✓    | ✓       |
 | PAGEXML             | ✓    | ✓    | ✓       |
 | FineReader          | ✓    | -    | ✓       |
-| Google Cloud Vision | ✓    | -    | ✓       |
+| Google Cloud Vision | ✓    | ✓    | ✓       |
 | Amazon AWS Textract | -    | -    | ✓       |
 | TEI                 | ✓    | -    | -       |
 

script/transform/gcv__alto

@@ -0,0 +1,32 @@
+#!/bin/bash
+
+SCRIPTDIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+VENDORDIR="$(cd $SCRIPTDIR/../../vendor/; pwd)"
+JAR="$VENDORDIR/JPageConverter/PageConverter.jar"
+INFILE="$1"
+OUTFILE="$2"
+ARGUMENT="$3"
+
+if [[ "$1" = "-" ]]; then
+    INFILE="$(mktemp)"
+    cat >"$INFILE"
+fi
+
+if [[ "$2" = "-" ]]; then
+    OUTFILE="$(mktemp)"
+fi
+
+java -jar "$JAR" -neg-coords toZero -source-json "$INFILE" -target-xml "$OUTFILE" -convert-to ALTO 2>&1
+
+if [[ "$1" = "-" ]]; then
+    rm "$INFILE"
+fi
+
+if [[ "$2" = "-" ]]; then
+    if [[ -z "$ARGUMENT" ]]; then
+        cat "$OUTFILE"
+    else
+        java -cp "$VENDORDIR/saxon.jar" net.sf.saxon.Query -s:"$OUTFILE" -qs:/ "$ARGUMENT"
+    fi
+    rm "$OUTFILE"
+fi