fix(training): switch to PAGE XML format for kurrent recognition training

Kraken 7 removed support for the legacy `path` format (image + .gt.txt
pairs) in VGSLRecognitionDataModule despite the CLI still advertising it.
Switching to PAGE XML (-f page) format which is the supported standard.

- Java export now writes .xml alongside .png (PAGE XML with TextLine,
  Baseline at 75% height, and Unicode transcription)
- XML special characters in transcription text are escaped (& < >)
- Python trainer globs *.xml and passes -f page to ketos train
- Regenerated frontend API types to include cer/loss/accuracy/epochs on
  OcrTrainingRun (were missing, causing empty CER column in history)
- Updated and extended TrainingDataExportServiceTest

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

ocr-service/main.py

@@ -357,7 +357,7 @@ async def train_model(
 
             log.info("Extracted %d ZIP entries to %s", len(os.listdir(tmp_dir)), tmp_dir)
 
-            ground_truth = glob.glob(os.path.join(tmp_dir, "*.gt.txt"))
+            ground_truth = glob.glob(os.path.join(tmp_dir, "*.xml"))
             if not ground_truth:
                 raise HTTPException(status_code=422, detail="No ground-truth files found in ZIP")
 
@@ -368,7 +368,7 @@ async def train_model(
             cmd = [
                 "ketos", "--workers", "0", "--device", "cpu", "--threads", "2",
                 "train",
-                "-f", "path",
+                "-f", "page",
                 "-o", checkpoint_dir,
                 "-q", "fixed",
                 "-N", "10",