Merge pull request #3338 from stweil/vector

Replace more GenericVector by std::vector

Makefile.am

@@ -387,7 +387,6 @@ endif
 noinst_HEADERS += src/classify/classify.h
 if !DISABLED_LEGACY_ENGINE
 noinst_HEADERS += src/classify/adaptive.h
-noinst_HEADERS += src/classify/blobclass.h
 noinst_HEADERS += src/classify/cluster.h
 noinst_HEADERS += src/classify/clusttool.h
 noinst_HEADERS += src/classify/featdefs.h

src/api/baseapi.cpp

@@ -23,9 +23,6 @@
 #  include "config_auto.h"
 #endif
 
-#ifndef DISABLED_LEGACY_ENGINE
-#  include "blobclass.h" // for ExtractFontName
-#endif
 #include "boxword.h"    // for BoxWord
 #include "coutln.h"     // for C_OUTLINE_IT, C_OUTLINE_LIST
 #include "dawg_cache.h" // for DawgCache
@@ -125,6 +122,34 @@ static const char *kOldVarsFile = "failed_vars.txt";
 /** Max string length of an int.  */
 const int kMaxIntSize = 22;
 
+#ifndef DISABLED_LEGACY_ENGINE
+static const char kUnknownFontName[] = "UnknownFont";
+
+static STRING_VAR(classify_font_name, kUnknownFontName,
+                  "Default font name to be used in training");
+
+// Finds the name of the training font and returns it in fontname, by cutting
+// it out based on the expectation that the filename is of the form:
+// /path/to/dir/[lang].[fontname].exp[num]
+// The [lang], [fontname] and [num] fields should not have '.' characters.
+// If the global parameter classify_font_name is set, its value is used instead.
+static void ExtractFontName(const char* filename, std::string* fontname) {
+  *fontname = classify_font_name;
+  if (*fontname == kUnknownFontName) {
+    // filename is expected to be of the form [lang].[fontname].exp[num]
+    // The [lang], [fontname] and [num] fields should not have '.' characters.
+    const char *basename = strrchr(filename, '/');
+    const char *firstdot = strchr(basename ? basename : filename, '.');
+    const char *lastdot  = strrchr(filename, '.');
+    if (firstdot != lastdot && firstdot != nullptr && lastdot != nullptr) {
+      ++firstdot;
+      *fontname = firstdot;
+      fontname->resize(lastdot - firstdot);
+    }
+  }
+}
+#endif
+
 /* Add all available languages recursively.
  */
 static void addAvailableLanguages(const std::string &datadir, const std::string &base,

src/ccstruct/normalis.cpp

@@ -356,6 +356,7 @@ void DENORM::LocalDenormTransform(const TPOINT &pt, TPOINT *original) const {
   original->x = IntCastRounded(float_result.x());
   original->y = IntCastRounded(float_result.y());
 }
+
 void DENORM::LocalDenormTransform(const FCOORD &pt, FCOORD *original) const {
   FCOORD rotated(pt.x() - final_xshift_, pt.y() - final_yshift_);
   if (x_map_ != nullptr && y_map_ != nullptr) {

src/ccstruct/normalis.h

@@ -21,6 +21,7 @@
 
 #include <vector>
 #include <tesseract/export.h>
+#include <vector>
 
 struct Pix;
 

src/ccutil/indexmapbidi.cpp

@@ -29,7 +29,11 @@ IndexMap::~IndexMap() = default;
 // Uses a binary search to find the result. For faster speed use
 // IndexMapBiDi, but that takes more memory.
 int IndexMap::SparseToCompact(int sparse_index) const {
-  int result = compact_map_.binary_search(sparse_index);
+  auto pos = std::upper_bound(compact_map_.begin(), compact_map_.end(), sparse_index);
+  if (pos > compact_map_.begin()) {
+    --pos;
+  }
+  auto result = pos - compact_map_.begin();
   return compact_map_[result] == sparse_index ? result : -1;
 }
 
@@ -45,7 +49,7 @@ void IndexMap::CopyFrom(const IndexMapBiDi &src) {
 
 // Writes to the given file. Returns false in case of error.
 bool IndexMap::Serialize(FILE *fp) const {
-  return tesseract::Serialize(fp, &sparse_size_) && compact_map_.Serialize(fp);
+  return tesseract::Serialize(fp, &sparse_size_) && tesseract::Serialize(fp, compact_map_);
 }
 
 // Reads from the given file. Returns false in case of error.
@@ -60,7 +64,7 @@ bool IndexMap::DeSerialize(bool swap, FILE *fp) {
   if (sparse_size > UINT16_MAX)
     return false;
   sparse_size_ = sparse_size;
-  return compact_map_.DeSerialize(swap, fp);
+  return tesseract::DeSerialize(swap, fp, compact_map_);
 }
 
 // Destructor.
@@ -85,7 +89,7 @@ void IndexMapBiDi::InitAndSetupRange(int sparse_size, int start, int end) {
 // Call Setup immediately after, or make calls to SetMap first to adjust the
 // mapping and then call Setup before using the map.
 void IndexMapBiDi::Init(int size, bool all_mapped) {
-  sparse_map_.init_to_size(size, -1);
+  sparse_map_.resize(size, -1);
   if (all_mapped) {
     for (int i = 0; i < size; ++i)
       sparse_map_[i] = i;
@@ -107,7 +111,7 @@ void IndexMapBiDi::Setup() {
       sparse_map_[i] = compact_size++;
     }
   }
-  compact_map_.init_to_size(compact_size, -1);
+  compact_map_.resize(compact_size, -1);
   for (int i = 0; i < sparse_map_.size(); ++i) {
     if (sparse_map_[i] >= 0) {
       compact_map_[sparse_map_[i]] = i;
@@ -168,7 +172,7 @@ void IndexMapBiDi::CompleteMerges() {
       compact_size = compact_index + 1;
   }
   // Re-generate the compact_map leaving holes for unused indices.
-  compact_map_.init_to_size(compact_size, -1);
+  compact_map_.resize(compact_size, -1);
   for (int i = 0; i < sparse_map_.size(); ++i) {
     if (sparse_map_[i] >= 0) {
       if (compact_map_[sparse_map_[i]] == -1)
@@ -177,8 +181,8 @@ void IndexMapBiDi::CompleteMerges() {
   }
   // Compact the compact_map, leaving tmp_compact_map saying where each
   // index went to in the compacted map.
-  GenericVector<int32_t> tmp_compact_map;
-  tmp_compact_map.init_to_size(compact_size, -1);
+  std::vector<int32_t> tmp_compact_map;
+  tmp_compact_map.resize(compact_size, -1);
   compact_size = 0;
   for (int i = 0; i < compact_map_.size(); ++i) {
     if (compact_map_[i] >= 0) {
@@ -186,7 +190,7 @@ void IndexMapBiDi::CompleteMerges() {
       compact_map_[compact_size++] = compact_map_[i];
     }
   }
-  compact_map_.truncate(compact_size);
+  compact_map_.resize(compact_size);
   // Now modify the entries in the sparse map to point to the new locations.
   for (int i = 0; i < sparse_map_.size(); ++i) {
     if (sparse_map_[i] >= 0) {
@@ -202,14 +206,14 @@ bool IndexMapBiDi::Serialize(FILE *fp) const {
   // Make a vector containing the rest of the map. If the map is many-to-one
   // then each additional sparse entry needs to be stored.
   // Normally we store only the compact map to save space.
-  GenericVector<int32_t> remaining_pairs;
+  std::vector<int32_t> remaining_pairs;
   for (int i = 0; i < sparse_map_.size(); ++i) {
     if (sparse_map_[i] >= 0 && compact_map_[sparse_map_[i]] != i) {
       remaining_pairs.push_back(i);
       remaining_pairs.push_back(sparse_map_[i]);
     }
   }
-  if (!remaining_pairs.Serialize(fp))
+  if (!tesseract::Serialize(fp, remaining_pairs))
     return false;
   return true;
 }
@@ -219,10 +223,10 @@ bool IndexMapBiDi::Serialize(FILE *fp) const {
 bool IndexMapBiDi::DeSerialize(bool swap, FILE *fp) {
   if (!IndexMap::DeSerialize(swap, fp))
     return false;
-  GenericVector<int32_t> remaining_pairs;
-  if (!remaining_pairs.DeSerialize(swap, fp))
+  std::vector<int32_t> remaining_pairs;
+  if (!tesseract::DeSerialize(swap, fp, remaining_pairs))
     return false;
-  sparse_map_.init_to_size(sparse_size_, -1);
+  sparse_map_.resize(sparse_size_, -1);
   for (int i = 0; i < compact_map_.size(); ++i) {
     sparse_map_[compact_map_[i]] = i;
   }

src/ccutil/indexmapbidi.h

@@ -2,7 +2,6 @@
 // File:        indexmapbidi.h
 // Description: Bi-directional mapping between a sparse and compact space.
 // Author:      rays@google.com (Ray Smith)
-// Created:     Tue Apr 06 11:33:59 PDT 2010
 //
 // (C) Copyright 2010, Google Inc.
 // Licensed under the Apache License, Version 2.0 (the "License");
@@ -77,7 +76,7 @@ protected:
   int32_t sparse_size_;
   // The compact space covers integers in the range [0, compact_map_.size()-1].
   // Each element contains the corresponding sparse index.
-  GenericVector<int32_t> compact_map_;
+  std::vector<int32_t> compact_map_;
 };
 
 // Bidirectional many-to-one mapping between a sparse and a compact discrete
@@ -170,7 +169,7 @@ private:
   }
 
   // Direct look-up of the compact index for each element in sparse space.
-  GenericVector<int32_t> sparse_map_;
+  std::vector<int32_t> sparse_map_;
 };
 
 } // namespace tesseract.

src/classify/blobclass.cpp

@@ -15,8 +15,6 @@
  ** limitations under the License.
  ******************************************************************************/
 
-#include "blobclass.h"
-
 #include <cstdio>
 
 #include "classify.h"
@@ -26,35 +24,6 @@
 
 namespace tesseract {
 
-static const char kUnknownFontName[] = "UnknownFont";
-
-static STRING_VAR(classify_font_name, kUnknownFontName, "Default font name to be used in training");
-
-/**----------------------------------------------------------------------------
-            Public Code
-----------------------------------------------------------------------------**/
-
-// Finds the name of the training font and returns it in fontname, by cutting
-// it out based on the expectation that the filename is of the form:
-// /path/to/dir/[lang].[fontname].exp[num]
-// The [lang], [fontname] and [num] fields should not have '.' characters.
-// If the global parameter classify_font_name is set, its value is used instead.
-void ExtractFontName(const char *filename, std::string *fontname) {
-  *fontname = classify_font_name;
-  if (*fontname == kUnknownFontName) {
-    // filename is expected to be of the form [lang].[fontname].exp[num]
-    // The [lang], [fontname] and [num] fields should not have '.' characters.
-    const char *basename = strrchr(filename, '/');
-    const char *firstdot = strchr(basename ? basename : filename, '.');
-    const char *lastdot = strrchr(filename, '.');
-    if (firstdot != lastdot && firstdot != nullptr && lastdot != nullptr) {
-      ++firstdot;
-      *fontname = firstdot;
-      fontname->resize(lastdot - firstdot);
-    }
-  }
-}
-
 /*---------------------------------------------------------------------------*/
 
 // Extracts features from the given blob and saves them in the tr_file_data_

src/classify/blobclass.h

@@ -1,33 +0,0 @@
-/******************************************************************************
- ** Filename: blobclass.h
- ** Purpose: Interface to high level classification and training.
- ** Author:  Dan Johnson
- **
- ** (c) Copyright Hewlett-Packard Company, 1988.
- ** Licensed under the Apache License, Version 2.0 (the "License");
- ** you may not use this file except in compliance with the License.
- ** You may obtain a copy of the License at
- ** http://www.apache.org/licenses/LICENSE-2.0
- ** Unless required by applicable law or agreed to in writing, software
- ** distributed under the License is distributed on an "AS IS" BASIS,
- ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- ** See the License for the specific language governing permissions and
- ** limitations under the License.
- ******************************************************************************/
-
-#ifndef BLOBCLASS_H
-#define BLOBCLASS_H
-
-#include <string>
-
-namespace tesseract {
-// Finds the name of the training font and returns it in fontname, by cutting
-// it out based on the expectation that the filename is of the form:
-// /path/to/dir/[lang].[fontname].exp[num]
-// The [lang], [fontname] and [num] fields should not have '.' characters.
-// If the global parameter classify_font_name is set, its value is used instead.
-void ExtractFontName(const char *filename, std::string *fontname);
-
-} // namespace tesseract.
-
-#endif

src/textord/makerow.cpp

@@ -126,13 +126,8 @@ static int row_y_order(       // sort function
  * Qsort style function to compare 2 TO_ROWS based on their spacing value.
  */
 static int row_spacing_order( // sort function
-    const void *item1, // items to compare
-    const void *item2) {
-  // converted ptr
-  const TO_ROW *row1 = *reinterpret_cast<const TO_ROW *const *>(item1);
-  // converted ptr
-  const TO_ROW *row2 = *reinterpret_cast<const TO_ROW *const *>(item2);
-
+    const TO_ROW *row1, // items to compare
+    const TO_ROW *row2) {
   return row1->spacing < row2->spacing;
 }
 

src/training/cntraining.cpp

@@ -136,7 +136,7 @@ int main(int argc, char *argv[]) {
   pCharList = CharList;
   // The norm protos will count the source protos, so we keep them here in
   // freeable_protos, so they can be freed later.
-  GenericVector<LIST> freeable_protos;
+  std::vector<LIST> freeable_protos;
   iterate(pCharList) {
     // Cluster
     CharSample = reinterpret_cast<LABELEDLIST> first_node(pCharList);

src/training/common/ctc.cpp

@@ -2,7 +2,6 @@
 // File:        ctc.cpp
 // Description: Slightly improved standard CTC to compute the targets.
 // Author:      Ray Smith
-// Created:     Wed Jul 13 15:50:06 PDT 2016
 //
 // (C) Copyright 2016, Google Inc.
 // Licensed under the Apache License, Version 2.0 (the "License");
@@ -18,7 +17,6 @@
 
 #include "ctc.h"
 
-#include "genericvector.h"
 #include "matrix.h"
 #include "network.h"
 #include "networkio.h"
@@ -217,7 +215,7 @@ static int BestLabel(const GENERIC_2D_ARRAY<float> &outputs, int t) {
 // to the network outputs.
 float CTC::CalculateBiasFraction() {
   // Compute output labels via basic decoding.
-  GenericVector<int> output_labels;
+  std::vector<int> output_labels;
   for (int t = 0; t < num_timesteps_; ++t) {
     int label = BestLabel(outputs_, t);
     while (t + 1 < num_timesteps_ && BestLabel(outputs_, t + 1) == label)
@@ -226,8 +224,8 @@ float CTC::CalculateBiasFraction() {
       output_labels.push_back(label);
   }
   // Simple bag of labels error calculation.
-  GenericVector<int> truth_counts(num_classes_, 0);
-  GenericVector<int> output_counts(num_classes_, 0);
+  std::vector<int> truth_counts(num_classes_, 0);
+  std::vector<int> output_counts(num_classes_, 0);
   for (int l = 0; l < num_labels_; ++l) {
     ++truth_counts[labels_[l]];
   }
@@ -353,10 +351,10 @@ void CTC::NormalizeSequence(GENERIC_2D_ARRAY<double> *probs) const {
 void CTC::LabelsToClasses(const GENERIC_2D_ARRAY<double> &probs, NetworkIO *targets) const {
   // For each timestep compute the max prob for each class over all
   // instances of the class in the labels_.
-  GenericVector<double> class_probs;
+  std::vector<double> class_probs;
   for (int t = 0; t < num_timesteps_; ++t) {
     float *targets_t = targets->f(t);
-    class_probs.init_to_size(num_classes_, 0.0);
+    class_probs.resize(num_classes_);
     for (int u = 0; u < num_labels_; ++u) {
       double prob = probs(t, u);
       // Note that although Graves specifies sum over all labels of the same

src/training/common/intfeaturedist.cpp

@@ -97,7 +97,7 @@ double IntFeatureDist::FeatureDistance(const std::vector<int> &features) const {
 
 // Compute the distance between the given feature vector and the last
 // Set feature vector.
-double IntFeatureDist::DebugFeatureDistance(const GenericVector<int> &features) const {
+double IntFeatureDist::DebugFeatureDistance(const std::vector<int> &features) const {
   const int num_test_features = features.size();
   const double denominator = total_feature_weight_ + num_test_features;
   double misses = denominator;

src/training/common/intfeaturedist.h

@@ -19,7 +19,7 @@
 #ifndef TESSERACT_CLASSIFY_INTFEATUREDIST_H_
 #define TESSERACT_CLASSIFY_INTFEATUREDIST_H_
 
-#include "genericvector.h"
+#include <vector>
 
 namespace tesseract {
 
@@ -53,7 +53,7 @@ public:
   // Compute the distance between the given feature vector and the last
   // Set feature vector.
   double FeatureDistance(const std::vector<int> &features) const;
-  double DebugFeatureDistance(const GenericVector<int> &features) const;
+  double DebugFeatureDistance(const std::vector<int> &features) const;
 
 private:
   // Clear all data.

src/training/common/intfeaturemap.cpp

@@ -4,7 +4,6 @@
 // File:        intfeaturemap.cpp
 // Description: Encapsulation of IntFeatureSpace with IndexMapBiDi
 //              to provide a subspace mapping and fast feature lookup.
-// Created:     Tue Oct 26 08:58:30 PDT 2010
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -133,7 +132,7 @@ int IntFeatureMap::FindNZFeatureMapping(SampleIterator *it) {
   int total_samples = 0;
   for (it->Begin(); !it->AtEnd(); it->Next()) {
     const TrainingSample &sample = it->GetSample();
-    GenericVector<int> features;
+    std::vector<int> features;
     feature_space_.IndexAndSortFeatures(sample.features(), sample.num_features(), &features);
     int num_features = features.size();
     for (int f = 0; f < num_features; ++f)
@@ -164,7 +163,7 @@ int IntFeatureMap::FinalizeMapping(SampleIterator *it) {
 }
 
 // Prints the map features from the set in human-readable form.
-void IntFeatureMap::DebugMapFeatures(const GenericVector<int> &map_features) const {
+void IntFeatureMap::DebugMapFeatures(const std::vector<int> &map_features) const {
   for (int i = 0; i < map_features.size(); ++i) {
     INT_FEATURE_STRUCT f = InverseMapFeature(map_features[i]);
     f.print();

src/training/common/intfeaturemap.h

@@ -116,7 +116,7 @@ public:
   }
 
   // Prints the map features from the set in human-readable form.
-  void DebugMapFeatures(const GenericVector<int> &map_features) const;
+  void DebugMapFeatures(const std::vector<int> &map_features) const;
 
 private:
   void Clear();

src/training/common/mastertrainer.cpp

@@ -89,7 +89,7 @@ bool MasterTrainer::Serialize(FILE *fp) const {
     return false;
   if (!fontinfo_table_.Serialize(fp))
     return false;
-  if (!xheights_.Serialize(fp))
+  if (!tesseract::Serialize(fp, xheights_))
     return false;
   return true;
 }
@@ -507,7 +507,7 @@ void MasterTrainer::SetupFlatShapeTable(ShapeTable *shape_table) {
   // must be clustered in order the fonts arrived, and reverse order of the
   // characters within each font.
   // Get a list of the fonts in the order they appeared.
-  GenericVector<int> active_fonts;
+  std::vector<int> active_fonts;
   int num_shapes = flat_shapes_.NumShapes();
   for (int s = 0; s < num_shapes; ++s) {
     int font = flat_shapes_.GetShape(s)[0].font_ids[0];
@@ -547,7 +547,7 @@ CLUSTERER *MasterTrainer::SetupForClustering(const ShapeTable &shape_table,
   shape_map.SetMap(shape_id, true);
   shape_map.Setup();
   // Reverse the order of the samples to match the previous behavior.
-  GenericVector<const TrainingSample *> sample_ptrs;
+  std::vector<const TrainingSample *> sample_ptrs;
   SampleIterator it;
   it.Init(&shape_map, &shape_table, false, &samples_);
   for (it.Begin(); !it.AtEnd(); it.Next()) {
@@ -588,10 +588,10 @@ void MasterTrainer::WriteInttempAndPFFMTable(const UNICHARSET &unicharset,
   // Now write pffmtable. This is complicated by the fact that the adaptive
   // classifier still wants one indexed by unichar-id, but the static
   // classifier needs one indexed by its shape class id.
-  // We put the shapetable_cutoffs in a GenericVector, and compute the
+  // We put the shapetable_cutoffs in a vector, and compute the
   // unicharset cutoffs along the way.
-  GenericVector<uint16_t> shapetable_cutoffs;
-  GenericVector<uint16_t> unichar_cutoffs;
+  std::vector<uint16_t> shapetable_cutoffs;
+  std::vector<uint16_t> unichar_cutoffs;
   for (int c = 0; c < unicharset.size(); ++c)
     unichar_cutoffs.push_back(0);
   /* then write out each class */
@@ -620,7 +620,7 @@ void MasterTrainer::WriteInttempAndPFFMTable(const UNICHARSET &unicharset,
   if (fp == nullptr) {
     tprintf("Error, failed to open file \"%s\"\n", pffmtable_file);
   } else {
-    shapetable_cutoffs.Serialize(fp);
+    tesseract::Serialize(fp, shapetable_cutoffs);
     for (int c = 0; c < unicharset.size(); ++c) {
       const char *unichar = unicharset.id_to_unichar(c);
       if (strcmp(unichar, " ") == 0) {
@@ -894,7 +894,8 @@ void MasterTrainer::ClusterShapes(int min_shapes, int max_shape_unichars, float
                                   ShapeTable *shapes) {
   int num_shapes = shapes->NumShapes();
   int max_merges = num_shapes - min_shapes;
-  auto *shape_dists = new GenericVector<ShapeDist>[num_shapes];
+  // TODO: avoid new / delete.
+  auto *shape_dists = new std::vector<ShapeDist>[num_shapes];
   float min_dist = kInfiniteDist;
   int min_s1 = 0;
   int min_s2 = 0;

src/training/common/mastertrainer.h

@@ -261,7 +261,7 @@ private:
   // Font metrics gathered from multiple files.
   FontInfoTable fontinfo_table_;
   // Array of xheights indexed by font ids in fontinfo_table_;
-  GenericVector<int32_t> xheights_;
+  std::vector<int32_t> xheights_;
 
   // Non-serialized data initialized by other means or used temporarily
   // during loading of training samples.

src/training/common/trainingsampleset.cpp

@@ -22,6 +22,7 @@
 #include <allheaders.h>
 #include "boxread.h"
 #include "fontinfo.h"
+//#include "helpers.h"
 #include "indexmapbidi.h"
 #include "intfeaturedist.h"
 #include "intfeaturemap.h"

src/training/degradeimage.cpp

@@ -3,7 +3,6 @@
  * Description: Function to degrade an image (usually of text) as if it
  *              has been printed and then scanned.
  * Authors:     Ray Smith
- * Created:     Tue Nov 19 2013
  *
  * (C) Copyright 2013, Google Inc.
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -22,7 +21,6 @@
 
 #include <allheaders.h> // from leptonica
 #include <cstdlib>
-#include "genericvector.h"
 #include "helpers.h" // For TRand.
 #include "rect.h"
 
@@ -175,7 +173,7 @@ Pix *DegradeImage(Pix *input, int exposure, TRand *randomizer, float *rotation)
 // Returns nullptr on error. The returned Pix must be pixDestroyed.
 Pix *PrepareDistortedPix(const Pix *pix, bool perspective, bool invert, bool white_noise,
                          bool smooth_noise, bool blur, int box_reduction, TRand *randomizer,
-                         GenericVector<TBOX> *boxes) {
+                         std::vector<TBOX> *boxes) {
   Pix *distorted = pixCopy(nullptr, const_cast<Pix *>(pix));
   // Things to do to synthetic training data.
   if ((white_noise || smooth_noise) && randomizer->SignedRand(1.0) > 0.0) {
@@ -214,7 +212,7 @@ Pix *PrepareDistortedPix(const Pix *pix, bool perspective, bool invert, bool whi
 // perspective distortion. Width and height only need to be set if there
 // is no pix. If there is a pix, then they will be taken from there.
 void GeneratePerspectiveDistortion(int width, int height, TRand *randomizer, Pix **pix,
-                                   GenericVector<TBOX> *boxes) {
+                                   std::vector<TBOX> *boxes) {
   if (pix != nullptr && *pix != nullptr) {
     width = pixGetWidth(*pix);
     height = pixGetHeight(*pix);

src/training/degradeimage.h

@@ -3,7 +3,6 @@
  * Description: Function to degrade an image (usually of text) as if it
  *              has been printed and then scanned.
  * Authors:     Ray Smith
- * Created:     Tue Nov 19 2013
  *
  * (C) Copyright 2013, Google Inc.
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -21,7 +20,6 @@
 #define TESSERACT_TRAINING_DEGRADEIMAGE_H_
 
 #include <allheaders.h>
-#include "genericvector.h"
 #include "helpers.h" // For TRand.
 #include "rect.h"
 
@@ -41,12 +39,12 @@ struct Pix *DegradeImage(struct Pix *input, int exposure, TRand *randomizer, flo
 // Returns nullptr on error. The returned Pix must be pixDestroyed.
 Pix *PrepareDistortedPix(const Pix *pix, bool perspective, bool invert, bool white_noise,
                          bool smooth_noise, bool blur, int box_reduction, TRand *randomizer,
-                         GenericVector<TBOX> *boxes);
+                         std::vector<TBOX> *boxes);
 // Distorts anything that has a non-null pointer with the same pseudo-random
 // perspective distortion. Width and height only need to be set if there
 // is no pix. If there is a pix, then they will be taken from there.
 void GeneratePerspectiveDistortion(int width, int height, TRand *randomizer, Pix **pix,
-                                   GenericVector<TBOX> *boxes);
+                                   std::vector<TBOX> *boxes);
 // Computes the coefficients of a randomized projective transformation.
 // The image transform requires backward transformation coefficient, and the
 // box transform the forward coefficients.

src/training/lstmeval.cpp

@@ -16,7 +16,6 @@
 ///////////////////////////////////////////////////////////////////////
 
 #include "commontraining.h"
-#include "genericvector.h"
 #include "lstmtester.h"
 #include "tprintf.h"
 
@@ -52,7 +51,7 @@ int main(int argc, char **argv) {
       tprintf("Failed to load language model from %s!\n", FLAGS_traineddata.c_str());
       return 1;
     }
-    GenericVector<char> model_data;
+    std::vector<char> model_data;
     if (!tesseract::LoadDataFromFile(FLAGS_model.c_str(), &model_data)) {
       tprintf("Failed to load model from: %s\n", FLAGS_model.c_str());
       return 1;