vesseract: added image_to_boxes

Makefile

@@ -5,7 +5,7 @@ all: format test
 
 test:
 	@echo "Running tests..."
-	$(v_compiler) test .
+	$(v_compiler) -g test .
 
 fmt:
 	$(v_compiler) fmt -w .

alternatives.v

@@ -10,4 +10,10 @@ pub fn image_to_alto_xml_path(image_path string) ?string {
 [inline]
 pub fn image_to_string_path(filepath string) ?string {
 	return image_to_string(image: filepath, lang: 'eng', args: '')
-}
+}
+
+// Variant of image_to_boxes, only a file path is required
+[inline]
+pub fn image_to_boxes_path(filepath string) ?[]Tesseract_box {
+	return image_to_boxes(image: filepath, lang: 'eng', args: '')
+}

vesseract.v

@@ -6,10 +6,11 @@ import os
 pub struct Tesseract_box {
 pub:
 	letter string
-	x1 int
-	y1 int
-	x2 int
-	y2 int
+	x1     u32
+	y1     u32
+	x2     u32
+	y2     u32
+	page   u32
 }
 
 // Used as a parameter
@@ -137,3 +138,47 @@ pub fn image_to_alto_xml(t Tesseract) ?string {
 	// Get XML
 	return xml
 }
+
+// Get bounding boxes from Tesseract
+// Return an array of Tesseract boxes
+pub fn image_to_boxes(t Tesseract) ?[]Tesseract_box {
+	// Run tesseract with bounding box detection
+	result := extract_text_tesseract(
+		image: t.image
+		lang: t.lang
+		args: t.args + ' batch.nochop makebox'
+	) or { return err }
+
+	// Load box file
+	box_file := os.read_file(result.id + '.box') or { return err }
+	lines := box_file.split('\n')
+
+	// Delete "box" file and txt
+	os.rm(result.id + '.box') ?
+
+	// Hold results
+	mut boxes := []Tesseract_box{}
+
+	// Parse
+	for line in lines {
+		// Letter, x1, y1, x2, y1, page
+		// Example: H 68 206 91 235 0
+		content := line.split(' ')
+
+		// Skip malformed lines
+		if content.len != 6 {
+			continue
+		}
+
+		boxes << Tesseract_box{
+			letter: content[0]
+			x1: content[1].u32()
+			y1: content[2].u32()
+			x2: content[3].u32()
+			y2: content[4].u32()
+			page: content[5].u32()
+		}
+	}
+
+	return boxes
+}

vesseract_test.v

@@ -58,3 +58,13 @@ fn test_image_to_alto_xml_path() {
 	xml := image_to_alto_xml_path('sample/demo.png') or { panic(err) }
 	assert xml.contains('http://www.loc.gov/standards/alto/ns-v3#')
 }
+
+fn test_image_to_boxes() {
+	boxes := image_to_boxes(image: 'sample/demo.png', lang: 'eng', args: '') or { panic(err) }
+
+	assert boxes[0].x1 == 68
+	assert boxes[0].y1 == 206
+	assert boxes[0].x2 == 91
+	assert boxes[0].y2 == 235
+	assert boxes.len == 16
+}

wrapper.v

@@ -15,6 +15,8 @@ pub:
 	output_filename string
 	// Tesseract output stdout
 	stdout_result string
+	// ID generated
+	id string
 }
 
 // Generate a id for a document to be processed
@@ -55,9 +57,9 @@ fn extract_text_tesseract(t Tesseract) ?Tesseract_output {
 	args << t.image
 
 	// Output tmp - Random ID
-	id := generate_id()
+	doc_id := generate_id()
 	// Output file (tesseract append .txt)
-	args << id
+	args << doc_id
 
 	if t.lang.len > 0 {
 		args << '-l ' + t.lang
@@ -74,7 +76,8 @@ fn extract_text_tesseract(t Tesseract) ?Tesseract_output {
 	return Tesseract_output{
 		image_path: t.image
 		arguments: args
-		output_filename: id + '.txt'
+		output_filename: doc_id + '.txt'
 		stdout_result: stdout
+		id: doc_id
 	}
 }