Add support for output formats.

pixelocr/gui/ocrengine.py

@@ -14,6 +14,7 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 
+import itertools
 from glob import glob
 from os import path
 from queue import Queue
@@ -27,9 +28,10 @@ from PyQt4.QtGui import (
     qApp
 )
 
+from .. import formatting
 from ..image import Image
 from ..page import Page, Glyph, Space
-from ..glyphdb import GlyphDB
+from ..glyphdb import GlyphDB, SPACE, NEWLINE
 
 
 class OCREngine(QThread):
@@ -53,7 +55,6 @@ class OCREngine(QThread):
 
     def run(self):
         for page_text in self.recognize():
-            print()
             print(page_text)
         if self.quit:
             qApp.quit()
@@ -62,27 +63,30 @@ class OCREngine(QThread):
         for filename in self.filenames:
             page = self.load_page(filename)
             self.pageChanged.emit(page)
-            yield '\n'.join(self.recognize_page(page))
+            yield self.recognize_page(page)
 
     def recognize_page(self, page):
-        for line in page.lines:
-            yield ''.join(self.recognize_line(line))
+        glyph_data_seq = itertools.chain(*(self.recognize_line(line) for line in page.lines))
+        output_format = formatting.TextFormat()
+        return ''.join(output_format.format(glyph_data_seq))
 
     def recognize_line(self, line):
-        yield from ' ' * int(line.indent / self.SPACE_WIDTH)
+        yield from [SPACE] * int(line.indent / self.SPACE_WIDTH)
         for glyph in line.glyphs:
             yield self.recognize_glyph(glyph)
+        yield NEWLINE
 
     def recognize_glyph(self, glyph):
         qApp.processEvents()
+
         if isinstance(glyph, Space):
-            return ' '
+            return SPACE
         try:
-            return self.glyphdb[glyph].text
+            glyph_data = self.glyphdb[glyph]
         except KeyError:
             text, bold, italic = self.ask_for_help(glyph)
-            self.glyphdb.add_glyph(glyph, text, bold, italic)
-            return text
+            glyph_data = self.glyphdb.add_glyph(glyph, text, bold, italic)
+        return glyph_data
 
     def ask_for_help(self, unknown_glyph):
         self.unknownGlyph.emit(unknown_glyph)