From 832c6edb971ab427318ad64cd49b6eb87d04bdbc Mon Sep 17 00:00:00 2001
From: Shree <shreeshrii@gmail.com>
Date: Fri, 14 Jun 2019 09:25:54 +0000
Subject: [PATCH 1/3] Allow saving of box/tiff pairs during base tesseract
 training

---
 src/training/tesstrain_utils.sh | 36 ++++++++++++++++++++++-----------
 1 file changed, 24 insertions(+), 12 deletions(-)

diff --git a/src/training/tesstrain_utils.sh b/src/training/tesstrain_utils.sh
index 9b0efb6b..e053b902 100644
--- a/src/training/tesstrain_utils.sh
+++ b/src/training/tesstrain_utils.sh
@@ -259,7 +259,7 @@ generate_font_image() {
     common_args+=" --char_spacing=${CHAR_SPACING} --exposure=${EXPOSURE}"
     common_args+=" --outputbase=${outbase} --max_pages=${MAX_PAGES}"
     if $DISTORT_IMAGE; then
-        common_args+=" --distort_image "
+        common_args+=" --distort_image --invert=false"
     fi
 
     # add --writing_mode=vertical-upright to common_args if the font is
@@ -326,6 +326,17 @@ phase_I_generate_image() {
             check_file_readable ${outbase}.box ${outbase}.tif
         done
     done
+    if $SAVE_BOX_TIFF && ( ! $LINEDATA ) ; then
+    tlog "\n=== Saving box/tiff pairs for training data ==="
+        for f in "${TRAINING_DIR}/${LANG_CODE}".*.box; do
+            tlog "Moving ${f} to ${OUTPUT_DIR}"
+            cp "${f}" "${OUTPUT_DIR}"
+        done
+        for f in "${TRAINING_DIR}/${LANG_CODE}".*.tif; do
+            tlog "Moving ${f} to ${OUTPUT_DIR}"
+            cp "${f}" "${OUTPUT_DIR}"
+        done
+    fi
 }
 
 # Phase UP : Generate (U)nicharset and (P)roperties file.
@@ -386,7 +397,7 @@ phase_D_generate_dawg() {
 
     # Punctuation DAWG
     # -r arguments to wordlist2dawg denote RTL reverse policy
-    # (see Trie::RTLReversePolicy enum in tesseract/src/dict/trie.h).
+    # (see Trie::RTLReversePolicy enum in third_party/tesseract/dict/trie.h).
     # We specify 0/RRP_DO_NO_REVERSE when generating number DAWG,
     # 1/RRP_REVERSE_IF_HAS_RTL for freq and word DAWGS,
     # 2/RRP_FORCE_REVERSE for the punctuation DAWG.
@@ -562,17 +573,18 @@ make__lstmdata() {
     --output_dir "${OUTPUT_DIR}" --lang "${LANG_CODE}" \
     "${pass_through}" "${lang_is_rtl}"
 
-  if $SAVE_BOX_TIFF; then
+    if $SAVE_BOX_TIFF ; then
     tlog "\n=== Saving box/tiff pairs for training data ==="
-  for f in "${TRAINING_DIR}/${LANG_CODE}".*.box; do
-    tlog "Moving ${f} to ${OUTPUT_DIR}"
-    mv "${f}" "${OUTPUT_DIR}"
-  done
-  for f in "${TRAINING_DIR}/${LANG_CODE}".*.tif; do
-    tlog "Moving ${f} to ${OUTPUT_DIR}"
-    mv "${f}" "${OUTPUT_DIR}"
-  done
-  fi
+        for f in "${TRAINING_DIR}/${LANG_CODE}".*.box; do
+            tlog "Moving ${f} to ${OUTPUT_DIR}"
+            mv "${f}" "${OUTPUT_DIR}"
+        done
+        for f in "${TRAINING_DIR}/${LANG_CODE}".*.tif; do
+            tlog "Moving ${f} to ${OUTPUT_DIR}"
+            mv "${f}" "${OUTPUT_DIR}"
+        done
+    fi
+
   tlog "\n=== Moving lstmf files for training data ==="
   for f in "${TRAINING_DIR}/${LANG_CODE}".*.lstmf; do
     tlog "Moving ${f} to ${OUTPUT_DIR}"

From 45cdf741ae9347f1b345b55419780b3705dd5be5 Mon Sep 17 00:00:00 2001
From: Shree <shreeshrii@gmail.com>
Date: Fri, 14 Jun 2019 09:32:41 +0000
Subject: [PATCH 2/3] Allow saving of box/tiff pairs during base tesseract
 training

---
 src/training/tesstrain_utils.sh | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/src/training/tesstrain_utils.sh b/src/training/tesstrain_utils.sh
index e053b902..51c370e0 100644
--- a/src/training/tesstrain_utils.sh
+++ b/src/training/tesstrain_utils.sh
@@ -397,7 +397,7 @@ phase_D_generate_dawg() {
 
     # Punctuation DAWG
     # -r arguments to wordlist2dawg denote RTL reverse policy
-    # (see Trie::RTLReversePolicy enum in third_party/tesseract/dict/trie.h).
+    # (see Trie::RTLReversePolicy enum in tesseract/src/dict/trie.h).
     # We specify 0/RRP_DO_NO_REVERSE when generating number DAWG,
     # 1/RRP_REVERSE_IF_HAS_RTL for freq and word DAWGS,
     # 2/RRP_FORCE_REVERSE for the punctuation DAWG.
@@ -573,17 +573,18 @@ make__lstmdata() {
     --output_dir "${OUTPUT_DIR}" --lang "${LANG_CODE}" \
     "${pass_through}" "${lang_is_rtl}"
 
-    if $SAVE_BOX_TIFF ; then
+  if $SAVE_BOX_TIFF; then
     tlog "\n=== Saving box/tiff pairs for training data ==="
-        for f in "${TRAINING_DIR}/${LANG_CODE}".*.box; do
-            tlog "Moving ${f} to ${OUTPUT_DIR}"
-            mv "${f}" "${OUTPUT_DIR}"
-        done
-        for f in "${TRAINING_DIR}/${LANG_CODE}".*.tif; do
-            tlog "Moving ${f} to ${OUTPUT_DIR}"
-            mv "${f}" "${OUTPUT_DIR}"
-        done
-    fi
+  for f in "${TRAINING_DIR}/${LANG_CODE}".*.box; do
+    tlog "Moving ${f} to ${OUTPUT_DIR}"
+    mv "${f}" "${OUTPUT_DIR}"
+  done
+  for f in "${TRAINING_DIR}/${LANG_CODE}".*.tif; do
+    tlog "Moving ${f} to ${OUTPUT_DIR}"
+    mv "${f}" "${OUTPUT_DIR}"
+  done
+  fi
+  done
 
   tlog "\n=== Moving lstmf files for training data ==="
   for f in "${TRAINING_DIR}/${LANG_CODE}".*.lstmf; do

From 6fa458794979d6b170642852e78a62844ec6169e Mon Sep 17 00:00:00 2001
From: Shree <shreeshrii@gmail.com>
Date: Fri, 14 Jun 2019 09:35:39 +0000
Subject: [PATCH 3/3] Allow saving of box/tiff pairs during base tesseract
 training

---
 src/training/tesstrain_utils.sh | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/training/tesstrain_utils.sh b/src/training/tesstrain_utils.sh
index 51c370e0..ef8457e2 100644
--- a/src/training/tesstrain_utils.sh
+++ b/src/training/tesstrain_utils.sh
@@ -584,7 +584,6 @@ make__lstmdata() {
     mv "${f}" "${OUTPUT_DIR}"
   done
   fi
-  done
 
   tlog "\n=== Moving lstmf files for training data ==="
   for f in "${TRAINING_DIR}/${LANG_CODE}".*.lstmf; do