Fix some typos (most found by codespell)

Signed-off-by: Stefan Weil <sw@weilnetz.de>

include/tesseract/baseapi.h

@@ -498,7 +498,7 @@ public:
    * metadata used by side-effect processes, such as reading a box
    * file or formatting as hOCR.
    *
-   * See ProcessPages for desciptions of other parameters.
+   * See ProcessPages for descriptions of other parameters.
    */
   bool ProcessPage(Pix *pix, int page_index, const char *filename,
                    const char *retry_config, int timeout_millisec,

java/com/google/scrollview/ScrollView.java

@@ -85,7 +85,7 @@ public class ScrollView {
   }
 
   /**
-   * The main program loop. Basically loops trough receiving messages and
+   * The main program loop. Basically loops through receiving messages and
    * processing them and then sending messages (if there are any).
    */
   private static void IOLoop() {

src/ccmain/control.cpp

@@ -235,7 +235,7 @@ bool Tesseract::RecogAllWordsPassN(int pass_n, ETEXT_DESC *monitor, PAGE_RES_IT
         continue;
       }
     }
-    // Sync pr_it with the wth WordData.
+    // Sync pr_it with the WordData.
     while (pr_it->word() != nullptr && pr_it->word() != word->word) {
       pr_it->forward();
     }

src/ccmain/fixspace.cpp

@@ -25,7 +25,7 @@
 #include "errcode.h"        // for ASSERT_HOST
 #include "normalis.h"       // for kBlnXHeight, kBlnBaselineOffset
 #include "pageres.h"        // for WERD_RES_IT, WERD_RES, WERD_RES_LIST
-#include "params.h"         // for IntParam, StringParam, BoolParam, Doub...
+#include "params.h"         // for IntParam, StringParam, BoolParam, DoubleParam, ...
 #include "ratngs.h"         // for WERD_CHOICE, FREQ_DAWG_PERM, NUMBER_PERM
 #include "rect.h"           // for TBOX
 #include "stepblob.h"       // for C_BLOB_IT, C_BLOB_LIST, C_BLOB

src/ccmain/tesseractclass.cpp

@@ -256,7 +256,7 @@ Tesseract::Tesseract()
     , INT_MEMBER(fixsp_non_noise_limit, 1, "How many non-noise blbs either side?", this->params())
     , double_MEMBER(fixsp_small_outlines_size, 0.28, "Small if lt xht x this", this->params())
     , BOOL_MEMBER(tessedit_prefer_joined_punct, false, "Reward punctuation joins", this->params())
-    , INT_MEMBER(fixsp_done_mode, 1, "What constitues done for spacing", this->params())
+    , INT_MEMBER(fixsp_done_mode, 1, "What constitutes done for spacing", this->params())
     , INT_MEMBER(debug_fix_space_level, 0, "Contextual fixspace debug", this->params())
     , STRING_MEMBER(numeric_punctuation, ".,", "Punct. chs expected WITHIN numbers", this->params())
     , INT_MEMBER(x_ht_acceptance_tolerance, 8,

src/ccmain/tesseractclass.h

@@ -902,7 +902,7 @@ public:
   INT_VAR_H(fixsp_non_noise_limit, 1, "How many non-noise blbs either side?");
   double_VAR_H(fixsp_small_outlines_size, 0.28, "Small if lt xht x this");
   BOOL_VAR_H(tessedit_prefer_joined_punct, false, "Reward punctuation joins");
-  INT_VAR_H(fixsp_done_mode, 1, "What constitues done for spacing");
+  INT_VAR_H(fixsp_done_mode, 1, "What constitutes done for spacing");
   INT_VAR_H(debug_fix_space_level, 0, "Contextual fixspace debug");
   STRING_VAR_H(numeric_punctuation, ".,", "Punct. chs expected WITHIN numbers");
   INT_VAR_H(x_ht_acceptance_tolerance, 8, "Max allowed deviation of blob top outside of font data");

src/ccstruct/blobbox.h

@@ -93,7 +93,7 @@ enum BlobSpecialTextType {
   BSTT_NONE,    // No special.
   BSTT_ITALIC,  // Italic style.
   BSTT_DIGIT,   // Digit symbols.
-  BSTT_MATH,    // Mathmatical symobls (not including digit).
+  BSTT_MATH,    // Mathematical symbols (not including digit).
   BSTT_UNCLEAR, // Characters with low recognition rate.
   BSTT_SKIP,    // Characters that we skip labeling (usually too small).
   BSTT_COUNT

src/ccstruct/pageres.h

@@ -217,7 +217,7 @@ public:
   std::vector<std::vector<std::pair<const char *, float>>> timesteps;
   // Stores the lstm choices of every timestep segmented by character
   std::vector<std::vector<std::vector<std::pair<const char *, float>>>> segmented_timesteps;
-  // Symbolchoices aquired during CTC
+  // Symbolchoices acquired during CTC
   std::vector<std::vector<std::pair<const char *, float>>> CTC_symbol_choices;
   // Stores if the timestep vector starts with a space
   bool leading_space = false;

src/ccstruct/polyaprx.cpp

@@ -296,7 +296,7 @@ void fix2(         // polygonal approx
     /*single fixed step */
     if (edgept->flags[FLAGS] & FIXED &&
         edgept->flags[RUNLENGTH] == 1
-        /*and neighours free */
+        /*and neighbours free */
         && edgept->next->flags[FLAGS] & FIXED &&
         (edgept->prev->flags[FLAGS] & FIXED) == 0
         /*same pair of dirs */

src/ccutil/ambigs.cpp

@@ -333,7 +333,7 @@ bool UnicharAmbigs::ParseAmbiguityLine(int line_num, int version, int debug_leve
     return false;
   }
   if (version > 0) {
-    // The next field being true indicates that the abiguity should
+    // The next field being true indicates that the ambiguity should
     // always be substituted (e.g. '' should always be changed to ").
     // For such "certain" n -> m ambigs tesseract will insert character
     // fragments for the n pieces in the unicharset. AmbigsFound()

src/classify/adaptmatch.cpp

@@ -190,7 +190,7 @@ void SetAdaptiveThreshold(float Threshold);
  * @param Blob    blob to be classified
  * @param[out] Choices    List of choices found by adaptive matcher.
  * filled on return with the choices found by the
- * class pruner and the ratings therefrom. Also
+ * class pruner and the ratings there from. Also
  * contains the detailed results of the integer matcher.
  *
  */

src/dict/dict.h

@@ -322,7 +322,7 @@ public:
    * initialized to NO_EDGE.  Since the punctuation dawg includes the empty
    * pattern " " (meaning anything without surrounding punctuation), having a
    * single entry for the punctuation dawg will cover all dawgs reachable
-   * therefrom -- that includes all number and word dawgs. The only dawg
+   * there from -- that includes all number and word dawgs. The only dawg
    * non-reachable from the punctuation_dawg is the pattern dawg.
    * If hyphen state needs to be applied, initial dawg_args->active_dawgs can
    * be copied from the saved hyphen state (maintained by Dict).

src/lstm/recodebeam.h

@@ -335,7 +335,7 @@ private:
                               int xCoord);
 
   // Calculates more accurate character boundaries which can be used to
-  // provide more acurate alternative symbol choices.
+  // provide more accurate alternative symbol choices.
   static void calculateCharBoundaries(std::vector<int> *starts, std::vector<int> *ends,
                                       std::vector<int> *character_boundaries_, int maxWidth);
 

src/textord/tablerecog.cpp

@@ -733,7 +733,7 @@ int StructuredTable::CountHorizontalIntersections(int y) {
 }
 
 // Counts how many text partitions are in this box.
-// This is used to count partitons in cells, as that can indicate
+// This is used to count partitions in cells, as that can indicate
 // how "strong" a potential table row/column (or even full table) actually is.
 int StructuredTable::CountPartitions(const TBOX &box) {
   ColPartitionGridSearch gsearch(text_grid_);

src/textord/tabvector.cpp

@@ -776,7 +776,7 @@ void TabVector::Evaluate(const ICOORD &vertical, TabFind *finder) {
 }
 
 // (Re)Fit a line to the stored points. Returns false if the line
-// is degenerate. Althougth the TabVector code mostly doesn't care about the
+// is degenerate. Although the TabVector code mostly doesn't care about the
 // direction of lines, XAtY would give silly results for a horizontal line.
 // The class is mostly aimed at use for vertical lines representing
 // horizontal tab stops.

src/textord/tabvector.h

@@ -363,7 +363,7 @@ public:
   void Evaluate(const ICOORD &vertical, TabFind *finder);
 
   // (Re)Fit a line to the stored points. Returns false if the line
-  // is degenerate. Althougth the TabVector code mostly doesn't care about the
+  // is degenerate. Although the TabVector code mostly doesn't care about the
   // direction of lines, XAtY would give silly results for a horizontal line.
   // The class is mostly aimed at use for vertical lines representing
   // horizontal tab stops.

src/textord/textlineprojection.cpp

@@ -35,7 +35,7 @@ const int kWrongWayPenalty = 4;
 // Ratio between parallel gap and perpendicular gap used to measure total
 // distance of a box from a target box in curved textline space.
 // parallel-gap is treated more favorably by this factor to allow catching
-// quotes and elipsis at the end of textlines.
+// quotes and ellipsis at the end of textlines.
 const int kParaPerpDistRatio = 4;
 // Multiple of scale_factor_ that the inter-line gap must be before we start
 // padding the increment box perpendicular to the text line.

src/textord/tordmain.cpp

@@ -502,14 +502,14 @@ bool Textord::clean_noise_from_row( // remove empties
           blob_box = outline->bounding_box();
           blob_size = blob_box.width() > blob_box.height() ? blob_box.width() : blob_box.height();
           if (blob_size < textord_noise_sizelimit * row->x_height()) {
-            dot_count++; // count smal outlines
+            dot_count++; // count small outlines
           }
           if (!outline->child()->empty() &&
               blob_box.height() < (1 + textord_noise_syfract) * row->x_height() &&
               blob_box.height() > (1 - textord_noise_syfract) * row->x_height() &&
               blob_box.width() < (1 + textord_noise_sxfract) * row->x_height() &&
               blob_box.width() > (1 - textord_noise_sxfract) * row->x_height()) {
-            super_norm_count++; // count smal outlines
+            super_norm_count++; // count small outlines
           }
         }
       } else {
@@ -598,14 +598,14 @@ void Textord::clean_noise_from_words( // remove empties
           blob_box = outline->bounding_box();
           blob_size = blob_box.width() > blob_box.height() ? blob_box.width() : blob_box.height();
           if (blob_size < textord_noise_sizelimit * row->x_height()) {
-            dot_count++; // count smal outlines
+            dot_count++; // count small outlines
           }
           if (!outline->child()->empty() &&
               blob_box.height() < (1 + textord_noise_syfract) * row->x_height() &&
               blob_box.height() > (1 - textord_noise_syfract) * row->x_height() &&
               blob_box.width() < (1 + textord_noise_sxfract) * row->x_height() &&
               blob_box.width() > (1 - textord_noise_sxfract) * row->x_height()) {
-            norm_count++; // count smal outlines
+            norm_count++; // count small outlines
           }
         }
       } else {

src/textord/tospace.cpp

@@ -64,7 +64,7 @@ void Textord::to_spacing(ICOORD page_tr,       // topright of page
                         block_non_space_gap_width);
     // Make sure relative values of block-level space and non-space gap
     // widths are reasonable. The ratio of 1:3 is also used in
-    // block_spacing_stats, to corrrect the block_space_gap_width
+    // block_spacing_stats, to correct the block_space_gap_width
     // Useful for arabic and hindi, when the non-space gap width is
     // often over-estimated and should not be trusted. A similar ratio
     // is found in block_spacing_stats.
@@ -1695,7 +1695,7 @@ TBOX Textord::reduced_box_next(TO_ROW *row,    // current row
  * the xheight.
  *
  *
- * !!!!!!! WONT WORK WITH LARGE UPPER CASE CHARS - T F V W - look at examples on
+ * !!!!!!! WON'T WORK WITH LARGE UPPER CASE CHARS - T F V W - look at examples on
  *         "home".  Perhaps we need something which say if the width ABOVE the
  *         xht alone includes the whole of the reduced width, then use the full
  *         blob box - Might still fail on italic F

src/training/degradeimage.cpp

@@ -78,7 +78,7 @@ const int kMinRampSize = 1000;
 // With no dilation, after covolution, the images are so light that a heavy
 // constant offset is required to make the 0 image look reasonable. A simple
 // constant offset multiple of exposure to undo this value is enough to achieve
-// all the required lightening. This gives the advantage that exposure level 1
+// all the required lighting. This gives the advantage that exposure level 1
 // with a single dilation gives a good impression of the broken-yet-too-dark
 // problem that is often seen in scans.
 // A small random rotation gives some varying greyscale values on the edges,

unittest/README.md

@@ -69,7 +69,7 @@
 
 ### Fonts
 
-* Microsoft fonts: arialbi.ttf, times.ttf, verdana.ttf - [instalation guide](https://www.makeuseof.com/tag/how-to-install-microsoft-core-fonts-in-ubuntu-linux/)
+* Microsoft fonts: arialbi.ttf, times.ttf, verdana.ttf - [installation guide](https://www.makeuseof.com/tag/how-to-install-microsoft-core-fonts-in-ubuntu-linux/)
 * [ae_Arab.ttf](https://www.wfonts.com/download/data/2014/12/03/ae-arab/ae-arab.zip)
 * dejavu-fonts: [DejaVuSans-ExtraLight.ttf](https://dejavu-fonts.github.io/Download.html)
 * [Lohit-Hindi.ttf](https://raw.githubusercontent.com/pratul/packageofpractices/master/assets/fonts/Lohit-Hindi.ttf)

unittest/paragraphs_test.cc

@@ -182,7 +182,7 @@ void EvaluateParagraphDetection(const TextAndModel *correct, int n,
       }
       dbg_lines.push_back(absl::StrCat(correct[i].ascii, annotation));
     }
-    LOG(INFO) << "Discrepency!\n" << absl::StrJoin(dbg_lines, "\n");
+    LOG(INFO) << "Discrepancy!\n" << absl::StrJoin(dbg_lines, "\n");
   }
 }