nixpkgs/pkgs/applications/misc/k2pdfopt/tesseract.patch

From 39aa8502eee7bb669a29d1a9b3bfe5c9595ad960 Mon Sep 17 00:00:00 2001
From: Daniel Fullmer <danielrf12@gmail.com>
Date: Fri, 13 Sep 2019 13:45:05 -0400
Subject: [PATCH] Willus mod changes from k2pdfopt

---
 src/api/Makefile.am        |   1 +
 src/api/baseapi.cpp        |  87 +++++++++++
 src/api/baseapi.h          |   3 +
 src/api/tesscapi.cpp       | 311 +++++++++++++++++++++++++++++++++++++
 src/api/tesseract.h        |  29 ++++
 src/ccmain/tessedit.cpp    |   5 +-
 src/ccutil/ccutil.h        |   7 +
 src/ccutil/genericvector.h |  21 ++-
 src/ccutil/mainblk.cpp     |  17 +-
 src/ccutil/params.cpp      |   3 +-
 src/ccutil/serialis.cpp    |   3 +
 src/ccutil/serialis.h      |   2 +
 src/lstm/input.cpp         |   3 +
 13 files changed, 488 insertions(+), 4 deletions(-)
 create mode 100644 src/api/tesscapi.cpp
 create mode 100644 src/api/tesseract.h

diff --git a/src/api/Makefile.am b/src/api/Makefile.am
index d9b76eb6..cd2dc30f 100644
--- a/src/api/Makefile.am
+++ b/src/api/Makefile.am
@@ -39,6 +39,7 @@ libtesseract_api_la_SOURCES += lstmboxrenderer.cpp
 libtesseract_api_la_SOURCES += pdfrenderer.cpp
 libtesseract_api_la_SOURCES += wordstrboxrenderer.cpp
 libtesseract_api_la_SOURCES += renderer.cpp
+libtesseract_api_la_SOURCES += tesscapi.cpp
 
 lib_LTLIBRARIES += libtesseract.la
 libtesseract_la_LDFLAGS = $(LEPTONICA_LIBS) $(OPENCL_LDFLAGS) $(libarchive_LIBS)
diff --git a/src/api/baseapi.cpp b/src/api/baseapi.cpp
index 9245d07c..ea964ee6 100644
--- a/src/api/baseapi.cpp
+++ b/src/api/baseapi.cpp
@@ -215,6 +215,14 @@ TessBaseAPI::TessBaseAPI()
   // Use the current locale if building debug code.
   std::locale::global(std::locale(""));
 #endif
+  const char *locale;
+  locale = std::setlocale(LC_ALL, nullptr);
+/* willus mod Remove assertions--taken care of in tesscapi.cpp */
+//  ASSERT_HOST(!strcmp(locale, "C"));
+  locale = std::setlocale(LC_CTYPE, nullptr);
+//  ASSERT_HOST(!strcmp(locale, "C"));
+  locale = std::setlocale(LC_NUMERIC, nullptr);
+//  ASSERT_HOST(!strcmp(locale, "C"));
 }
 
 TessBaseAPI::~TessBaseAPI() {
@@ -1333,6 +1341,85 @@ static void AddBoxToTSV(const PageIterator* it, PageIteratorLevel level,
   text->add_str_int("\t", bottom - top);
 }
 
+/* willus mod */
+int TessBaseAPI::GetOCRWords(int **x00,int **y00,int **x11,int **y11,int **ybaseline0,
+                             char **utf8words)
+
+    {
+    int iword,nwords,totlen,it8;
+    int *x0,*y0,*x1,*y1,*ybaseline;
+    char *tutf8;
+
+    ResultIterator *res_it = GetIterator();
+    /* Count words */
+    iword=0;
+    totlen=0;
+    while (!res_it->Empty(RIL_BLOCK))
+        {
+        if (res_it->Empty(RIL_WORD))
+            {
+            res_it->Next(RIL_WORD);
+            continue;
+            }
+        iword++;
+        STRING textstr=std::unique_ptr<const char[]>(res_it->GetUTF8Text(RIL_WORD)).get();
+        totlen+=strlen(textstr.string())+1;
+        res_it->Next(RIL_WORD);
+        }
+    nwords=iword;
+/*
+printf("\nnwords=%d, totlen=%d\n",nwords,totlen);
+*/
+    x0=(*x00)=(int *)malloc(sizeof(int)*5*nwords);
+    y0=(*y00)=&x0[nwords];
+    x1=(*x11)=&y0[nwords];
+    y1=(*y11)=&x1[nwords];
+    ybaseline=(*ybaseline0)=&y1[nwords];
+    tutf8=(*utf8words)=(char *)malloc(totlen);
+    iword=0;
+    it8=0;
+    res_it->Begin();
+    while (!res_it->Empty(RIL_BLOCK))
+        {
+        if (res_it->Empty(RIL_WORD))
+            {
+            res_it->Next(RIL_WORD);
+            continue;
+            }
+        STRING textstr=std::unique_ptr<const char[]>(res_it->GetUTF8Text(RIL_WORD)).get();
+        strcpy(&tutf8[it8],textstr.string());
+        it8 += strlen(&tutf8[it8])+1;
+        /*
+        STRING textstr("");
+        textstr += std::unique_ptr<const char[]>(res_it->GetUTF8Text(RIL_WORD)).get();
+        */
+/*
+printf("Word %d: '%s'\n",iword,textstr.string());
+*/
+        int left, top, right, bottom;
+        int u1,v1,u2,v2;
+        res_it->BoundingBox(RIL_WORD, &left, &top, &right, &bottom);
+        res_it->Baseline(RIL_WORD, &u1, &v1, &u2, &v2);
+        x0[iword]=left;
+        x1[iword]=right;
+        y0[iword]=top;
+        y1[iword]=bottom;
+        ybaseline[iword]=(v1+v2)/2;
+        iword++;
+/*
+printf("BB: (%d,%d)-(%d,%d)  BL: (%d,%d)-(%d,%d)\n",left,bottom,right,top,x1,y1,x2,y2);
+*/
+        res_it->Next(RIL_WORD);
+        }
+/*
+printf("iword=%d\n",iword);
+*/
+    return(iword);
+    }
+
+/* willus mod */
+int GetOCRWords(int **x0,int **y0,int **x1,int **y1,int **ybaseline,char **utf8words);
+
 /**
  * Make a TSV-formatted string from the internal data structures.
  * page_number is 0-based but will appear in the output as 1-based.
diff --git a/src/api/baseapi.h b/src/api/baseapi.h
index 3724dd92..23be5920 100644
--- a/src/api/baseapi.h
+++ b/src/api/baseapi.h
@@ -575,6 +575,9 @@ class TESS_API TessBaseAPI {
    */
   char* GetHOCRText(ETEXT_DESC* monitor, int page_number);
 
+/* willus mod */
+int GetOCRWords(int **x0,int **y0,int **x1,int **y1,int **ybaseline,char **utf8words);
+
   /**
    * Make a HTML-formatted string with hOCR markup from the internal
    * data structures.
diff --git a/src/api/tesscapi.cpp b/src/api/tesscapi.cpp
new file mode 100644
index 00000000..1752fafe
--- /dev/null
+++ b/src/api/tesscapi.cpp
@@ -0,0 +1,311 @@
+/*
+** tesscapi.cpp    willus.com attempt at C wrapper for tesseract.
+**                 (Butchered from tesseractmain.cpp)
+**                 Last udpated 9-1-12
+**
+** Copyright (C) 2012  http://willus.com
+**
+** This program is free software: you can redistribute it and/or modify
+** it under the terms of the GNU Affero General Public License as
+** published by the Free Software Foundation, either version 3 of the
+** License, or (at your option) any later version.
+**
+** This program is distributed in the hope that it will be useful,
+** but WITHOUT ANY WARRANTY; without even the implied warranty of
+** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+** GNU Affero General Public License for more details.
+**
+** You should have received a copy of the GNU Affero General Public License
+** along with this program.  If not, see <http://www.gnu.org/licenses/>.
+**
+*/
+
+/*
+#include "mfcpch.h"
+*/
+// #define USE_VLD //Uncomment for Visual Leak Detector.
+#if (defined _MSC_VER && defined USE_VLD)
+#include <vld.h>
+#endif
+
+// Include automatically generated configuration file if running autoconf
+#ifdef HAVE_CONFIG_H
+#include "config_auto.h"
+#endif
+#include <locale.h>
+#ifdef USING_GETTEXT
+#include <libintl.h>
+#define _(x) gettext(x)
+#else
+#define _(x) (x)
+#endif
+
+#include "allheaders.h"
+#include "baseapi.h"
+#include "strngs.h"
+#include "params.h"
+#include "blobs.h"
+#include "simddetect.h"
+#include "tesseractclass.h"
+/*
+#include "notdll.h"
+*/
+
+/* C Wrappers */
+#include "tesseract.h"
+
+// static tesseract::TessBaseAPI api[4];
+
+/*
+** ocr_type=0:  OEM_DEFAULT
+** ocr_type=1:  OEM_TESSERACT_ONLY
+** ocr_type=2:  OEM_LSTM_ONLY
+** ocr_type=3:  OEM_TESSERACT_LSTM_COMBINED
+*/
+void *tess_capi_init(char *datapath,char *language,int ocr_type,FILE *out,
+                     char *initstr,int maxlen,int *status)
+
+    {
+    char original_locale[256];
+    tesseract::TessBaseAPI *api = new tesseract::TessBaseAPI;
+/*
+printf("@tess_capi_init\n");
+printf("    datapath='%s'\n",datapath);
+printf("    language='%s'\n",language);
+printf("    ocr_type=%d\n",ocr_type);
+*/
+#ifdef USE_NLS
+    setlocale (LC_ALL, "");
+    bindtextdomain (PACKAGE, LOCALEDIR);
+    textdomain (PACKAGE);
+#endif
+    /* willus mod, 11-24-16 */
+    /* Tesseract needs "C" locale to correctly parse all data .traineddata files. */
+/*
+printf("locale='%s'\n",setlocale(LC_ALL,NULL));
+printf("ctype='%s'\n",setlocale(LC_CTYPE,NULL));
+printf("numeric='%s'\n",setlocale(LC_NUMERIC,NULL));
+*/
+    strncpy(original_locale,setlocale(LC_ALL,NULL),255);
+    original_locale[255]='\0';
+/*
+printf("original_locale='%s'\n",original_locale);
+*/
+    setlocale(LC_ALL,"C");
+/*
+printf("new locale='%s'\n",setlocale(LC_ALL,NULL));
+printf("new ctype='%s'\n",setlocale(LC_CTYPE,NULL));
+printf("new numeric='%s'\n",setlocale(LC_NUMERIC,NULL));
+*/
+    // fprintf(stderr, "tesseract %s\n", tesseract::TessBaseAPI::Version());
+    // Make the order of args a bit more forgiving than it used to be.
+    const char* lang = "eng";
+    tesseract::PageSegMode pagesegmode = tesseract::PSM_SINGLE_BLOCK;
+    if (language!=NULL && language[0]!='\0')
+        lang = language;
+    /*
+    if (output == NULL)
+        {
+        fprintf(stderr, _("Usage:%s imagename outputbase [-l lang] "
+                      "[-psm pagesegmode] [configfile...]\n"), argv[0]);
+        fprintf(stderr,
+            _("pagesegmode values are:\n"
+              "0 = Orientation and script detection (OSD) only.\n"
+              "1 = Automatic page segmentation with OSD.\n"
+              "2 = Automatic page segmentation, but no OSD, or OCR\n"
+              "3 = Fully automatic page segmentation, but no OSD. (Default)\n"
+              "4 = Assume a single column of text of variable sizes.\n"
+              "5 = Assume a single uniform block of vertically aligned text.\n"
+              "6 = Assume a single uniform block of text.\n"
+              "7 = Treat the image as a single text line.\n"
+              "8 = Treat the image as a single word.\n"
+              "9 = Treat the image as a single word in a circle.\n"
+              "10 = Treat the image as a single character.\n"));
+        fprintf(stderr, _("-l lang and/or -psm pagesegmode must occur before any"
+                      "configfile.\n"));
+        exit(1);
+        }
+    */
+/*
+printf("SSE = %s\n",SIMDDetect::IsSSEAvailable() ? "AVAILABLE" : "NOT AVAILABLE");
+printf("AVX = %s\n",SIMDDetect::IsAVXAvailable() ? "AVAILABLE" : "NOT AVAILABLE");
+*/
+/*
+v4.00 loads either TESSERACT enginer, LSTM engine, or both.  No CUBE.
+*/
+    ocr_type=0; /* Ignore specified and use default */
+    api->SetOutputName(NULL);
+    (*status)=api->Init(datapath,lang,
+              ocr_type==0 ? tesseract::OEM_DEFAULT :
+                (ocr_type==1 ? tesseract::OEM_TESSERACT_ONLY :
+                   (ocr_type==2 ? tesseract::OEM_LSTM_ONLY :
+                                  (tesseract::OEM_TESSERACT_LSTM_COMBINED))));
+    if ((*status)!=0)
+        {
+        /* willus mod, 11-24-16 */
+        setlocale(LC_ALL,original_locale);
+        api->End();
+        delete api;
+        return(NULL);
+        }
+    /*
+    api.Init("tesscapi",lang,tesseract::OEM_DEFAULT,
+           &(argv[arg]), argc - arg, NULL, NULL, false);
+    */
+    // We have 2 possible sources of pagesegmode: a config file and
+    // the command line. For backwards compatability reasons, the
+    // default in tesseract is tesseract::PSM_SINGLE_BLOCK, but the
+    // default for this program is tesseract::PSM_AUTO. We will let
+    // the config file take priority, so the command-line default
+    // can take priority over the tesseract default, so we use the
+    // value from the command line only if the retrieved mode
+    // is still tesseract::PSM_SINGLE_BLOCK, indicating no change
+    // in any config file. Therefore the only way to force
+    // tesseract::PSM_SINGLE_BLOCK is from the command line.
+    // It would be simpler if we could set the value before Init,
+    // but that doesn't work.
+    if (api->GetPageSegMode() == tesseract::PSM_SINGLE_BLOCK)
+        api->SetPageSegMode(pagesegmode);
+
+    /*
+    ** Initialization message
+    */
+    {
+    char istr[1024];
+    int sse,avx;
+
+// printf("tessedit_ocr_engine_mode = %d\n",tessedit_ocr_engine_mode);
+    sprintf(istr,"%s",api->Version());
+    sse=tesseract::SIMDDetect::IsSSEAvailable();
+    avx=tesseract::SIMDDetect::IsAVXAvailable();
+    if (sse || avx)
+        sprintf(&istr[strlen(istr)]," [%s]",sse&&avx?"SSE+AVX":(sse?"SSE":"AVX"));
+    sprintf(&istr[strlen(istr)],"\n    Tesseract data folder = '%s'",datapath==NULL?getenv("TESSDATA_PREFIX"):datapath);
+    strcat(istr,"\n    Tesseract languages: ");
+    GenericVector<STRING> languages;
+    api->GetLoadedLanguagesAsVector(&languages);
+/*
+printf("OEM=%d\n",api->oem());
+printf("Langs='%s'\n",api->GetInitLanguagesAsString());
+printf("AnyTessLang()=%d\n",(int)api->tesseract()->AnyTessLang());
+printf("AnyLSTMLang()=%d\n",(int)api->tesseract()->AnyLSTMLang());
+printf("num_sub_langs()=%d\n",api->tesseract()->num_sub_langs());
+printf("languages.size()=%d\n",(int)languages.size());
+*/
+
+    for (int i=0;i<=api->tesseract()->num_sub_langs();i++)
+        {
+        tesseract::Tesseract *lang1;
+        int eng;
+        lang1 = i==0 ? api->tesseract() : api->tesseract()->get_sub_lang(i-1);
+        eng=(int)lang1->tessedit_ocr_engine_mode;
+        sprintf(&istr[strlen(istr)],"%s%s [%s]",i==0?"":", ",lang1->lang.string(),
+                 eng==2?"LSTM+Tess":(eng==1?"LSTM":"Tess"));
+        }
+/*
+printf("%d. '%s'\n",i+1,languages[i].string());
+printf("    sublang[%d].oem_engine = %d\n",i+1,(int)api->tesseract()->get_sub_lang(i)->tessedit_ocr_engine_mode);
+*/
+
+    /*
+    if (ocr_type==0 || ocr_type==3)
+        sprintf(&istr[strlen(istr)],"[LSTM+] (lang=");
+    else if (ocr_type==2)
+        sprintf(&istr[strlen(istr)],"[LSTM] (lang=");
+    strncpy(&istr[strlen(istr)],language,253-strlen(istr));
+    istr[253]='\0';
+    strcat(istr,")");
+    */
+    if (out!=NULL)
+        fprintf(out,"%s\n",istr);
+    if (initstr!=NULL)
+        {
+        strncpy(initstr,istr,maxlen-1);
+        initstr[maxlen-1]='\0';
+        }
+    }
+
+
+    /* Turn off LSTM debugging output */
+    api->SetVariable("lstm_debug_level","0");
+#if (WILLUSDEBUG & 1)
+    api->SetVariable("lstm_debug_level","9");
+    api->SetVariable("paragraph_debug_level","9");
+    api->SetVariable("tessdata_manager_debug_level","9");
+    api->SetVariable("tosp_debug_level","9");
+    api->SetVariable("wordrec_debug_level","9");
+    api->SetVariable("segsearch_debug_level","9");
+#endif
+    /* willus mod, 11-24-16 */
+    setlocale(LC_ALL,original_locale);
+    return((void *)api);
+    }
+
+
+int tess_capi_get_ocr(void *vapi,PIX *pix,char *outstr,int maxlen,int segmode,FILE *out)
+
+    {
+    tesseract::TessBaseAPI *api;
+    static int old_segmode=-1;
+
+    api=(tesseract::TessBaseAPI *)vapi;
+    if (old_segmode != segmode)
+        {
+        old_segmode=segmode;
+        api->SetPageSegMode((tesseract::PageSegMode)segmode);
+        }
+    if (!api->ProcessPage(pix,0,NULL,NULL,0,NULL))
+        {
+        /* pixDestroy(&pix); */
+        if (out!=NULL)
+            fprintf(out,"tesscapi:  Error during bitmap processing.\n");
+        api->Clear();
+        return(-1);
+        }
+    strncpy(outstr,api->GetUTF8Text(),maxlen-1);
+    outstr[maxlen-1]='\0';
+    api->Clear();
+    return(0);
+    }
+
+
+int tess_capi_get_ocr_multiword(void *vapi,PIX *pix,int segmode,
+                                int **left,int **top,int **right,int **bottom,
+                                int **ybase,char **text,int *nw,
+                                FILE *out)
+
+    {
+    tesseract::TessBaseAPI *api;
+    static int old_segmode=-1;
+
+    api=(tesseract::TessBaseAPI *)vapi;
+    if (old_segmode != segmode)
+        {
+        old_segmode=segmode;
+        api->SetPageSegMode((tesseract::PageSegMode)segmode);
+        }
+    if (!api->ProcessPage(pix,0,NULL,NULL,0,NULL))
+        {
+        if (out!=NULL)
+            fprintf(out,"tesscapi:  Error during bitmap processing.\n");
+        api->Clear();
+        (*nw)=0;
+        return(-1);
+        }
+    (*nw)=api->GetOCRWords(left,top,right,bottom,ybase,text);
+    api->Clear();
+    return(0);
+    }
+
+
+void tess_capi_end(void *vapi)
+
+    {
+    tesseract::TessBaseAPI *api;
+
+    if (vapi==NULL)
+        return;
+    api=(tesseract::TessBaseAPI *)vapi;
+    api->End();
+    delete api;
+    }
diff --git a/src/api/tesseract.h b/src/api/tesseract.h
new file mode 100644
index 00000000..575948cc
--- /dev/null
+++ b/src/api/tesseract.h
@@ -0,0 +1,29 @@
+/*
+** Willus.com's Tesseract C Wrappers
+**
+** 6-8-12
+**
+*/
+
+#ifndef           _TESSERACT_H_
+#define           _TESSERACT_H_
+
+//#include <leptonica.h>
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+void *tess_capi_init(char *datapath,char *language,int ocr_type,FILE *out,
+                    char *initstr,int maxlen,int *status);
+int tess_capi_get_ocr(void *api,PIX *pix,char *outstr,int maxlen,int segmode,FILE *out);
+int tess_capi_get_ocr_multiword(void *vapi,PIX *pix,int segmode,
+                                int **left,int **top,int **right,int **bottom,
+                                int **ybase,char **text,int *nw,
+                                FILE *out);
+void tess_capi_end(void *api);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
diff --git a/src/ccmain/tessedit.cpp b/src/ccmain/tessedit.cpp
index 17f0951b..7af94ee2 100644
--- a/src/ccmain/tessedit.cpp
+++ b/src/ccmain/tessedit.cpp
@@ -101,6 +101,10 @@ bool Tesseract::init_tesseract_lang_data(
         " to your \"tessdata\" directory.\n");
     return false;
   }
+  /* willus mod */
+  TFile fp;
+  strncpy(fp.tfile_filename,tessdata_path.string(),511);
+  fp.tfile_filename[511]='\0';
 #ifndef DISABLED_LEGACY_ENGINE
   if (oem == OEM_DEFAULT) {
     // Set the engine mode from availability, which can then be overridden by
@@ -116,7 +120,6 @@ bool Tesseract::init_tesseract_lang_data(
 #endif  // ndef DISABLED_LEGACY_ENGINE
 
   // If a language specific config file (lang.config) exists, load it in.
-  TFile fp;
   if (mgr->GetComponent(TESSDATA_LANG_CONFIG, &fp)) {
     ParamUtils::ReadParamsFromFp(SET_PARAM_CONSTRAINT_NONE, &fp,
                                  this->params());
diff --git a/src/ccutil/ccutil.h b/src/ccutil/ccutil.h
index 71e89c60..bdeccc14 100644
--- a/src/ccutil/ccutil.h
+++ b/src/ccutil/ccutil.h
@@ -80,6 +80,13 @@ class CCUtil {
   // Member parameters.
   // These have to be declared and initialized after params_ member, since
   // params_ should be initialized before parameters are added to it.
+/* willus mod */
+/*
+  #ifdef _WIN32
+  STRING_VAR_H(tessedit_module_name, WINDLLNAME,
+               "Module colocated with tessdata dir");
+  #endif
+*/
   INT_VAR_H(ambigs_debug_level, 0, "Debug level for unichar ambiguities");
   BOOL_VAR_H(use_definite_ambigs_for_classifier, false,
              "Use definite ambiguities when running character classifier");
diff --git a/src/ccutil/genericvector.h b/src/ccutil/genericvector.h
index 3556d153..3a5e8662 100644
--- a/src/ccutil/genericvector.h
+++ b/src/ccutil/genericvector.h
@@ -382,7 +382,26 @@ inline bool LoadDataFromFile(const char* filename, GenericVector<char>* data) {
       // reserve an extra byte in case caller wants to append a '\0' character
       data->reserve(size + 1);
       data->resize_no_init(size);
-      result = static_cast<long>(fread(&(*data)[0], 1, size, fp)) == size;
+    /* willus mod Dec 2018--weird issue with Win XP and MinGW gcc 7.3.0 */
+    /* Can't read entire file at once -- need to break up into smaller blocksize reads */
+    {
+    int frs,n;
+    int blocksize;
+    blocksize=1024*1024;
+    for (n=0;1;)
+        {
+        int bs;
+        bs= size-n > blocksize ? blocksize : size-n;
+        frs=(int)fread(&(*data)[n],1,bs,fp);
+        n+=frs;
+        if (frs<bs || bs<blocksize || n>=size)
+            break;
+        }
+    result = static_cast<long>((long)n==size);
+    }
+    /*
+    result = static_cast<long>(fread(&(*data)[0], 1, size, fp)) == size;
+    */
     }
     fclose(fp);
   }
diff --git a/src/ccutil/mainblk.cpp b/src/ccutil/mainblk.cpp
index 52b04b04..80b26044 100644
--- a/src/ccutil/mainblk.cpp
+++ b/src/ccutil/mainblk.cpp
@@ -55,8 +55,22 @@ void CCUtil::main_setup(const char *argv0, const char *basename) {
 #if defined(_WIN32)
   } else if (datadir == nullptr || _access(datadir.string(), 0) != 0) {
     /* Look for tessdata in directory of executable. */
+    /*
+    char drive[_MAX_DRIVE];
+    char dir[_MAX_DIR];
+    */
     char path[_MAX_PATH];
-    DWORD length = GetModuleFileName(nullptr, path, sizeof(path));
+    int i;
+    /* DWORD length = */ GetModuleFileName(nullptr, path, sizeof(path));
+    /* willus mod--avoid _splitpath_s -- not in XP */
+    for (i=strlen(path)-1;i>=0 && path[i]!='/' && path[i]!='\\';i--);
+    if (i>=0)
+        {
+        path[i]='\0';
+        datadir=path;
+        datadir += "/tessdata";
+        }
+    /*
     if (length > 0 && length < sizeof(path)) {
       char* separator = std::strrchr(path, '\\');
       if (separator != nullptr) {
@@ -65,6 +79,7 @@ void CCUtil::main_setup(const char *argv0, const char *basename) {
         datadir += "/tessdata";
       }
     }
+    */
 #endif /* _WIN32 */
 #if defined(TESSDATA_PREFIX)
   } else {
diff --git a/src/ccutil/params.cpp b/src/ccutil/params.cpp
index 00bf2563..486c5ce0 100644
--- a/src/ccutil/params.cpp
+++ b/src/ccutil/params.cpp
@@ -82,7 +82,8 @@ bool ParamUtils::ReadParamsFromFp(SetParamConstraint constraint, TFile *fp,
 
       if (!foundit) {
         anyerr = true;         // had an error
-        tprintf("Warning: Parameter not found: %s\n", line);
+        /* willus mod */
+        tprintf("Tesseract warning: Parameter %s not found in file %s.\n",line,fp->tfile_filename);
       }
     }
   }
diff --git a/src/ccutil/serialis.cpp b/src/ccutil/serialis.cpp
index 7def011f..6107a494 100644
--- a/src/ccutil/serialis.cpp
+++ b/src/ccutil/serialis.cpp
@@ -201,6 +201,9 @@ bool TFile::Open(const STRING& filename, FileReader reader) {
   offset_ = 0;
   is_writing_ = false;
   swap_ = false;
+  /* willus mod */
+  strncpy(tfile_filename,filename.string(),511);
+  tfile_filename[511]='\0';
   if (reader == nullptr)
     return LoadDataFromFile(filename, data_);
   else
diff --git a/src/ccutil/serialis.h b/src/ccutil/serialis.h
index 095b9227..4cc8251e 100644
--- a/src/ccutil/serialis.h
+++ b/src/ccutil/serialis.h
@@ -77,6 +77,8 @@ class TFile {
  public:
   TFile();
   ~TFile();
+  /* willus mod */
+  char tfile_filename[512];
 
   // All the Open methods load the whole file into memory for reading.
   // Opens a file with a supplied reader, or nullptr to use the default.
diff --git a/src/lstm/input.cpp b/src/lstm/input.cpp
index 73b584b3..0b0b54c3 100644
--- a/src/lstm/input.cpp
+++ b/src/lstm/input.cpp
@@ -93,8 +93,11 @@ Pix* Input::PrepareLSTMInputs(const ImageData& image_data,
     return nullptr;
   }
   if (width < min_width || height < min_width) {
+    /* willus mod -- no warning */
+    /*
     tprintf("Image too small to scale!! (%dx%d vs min width of %d)\n", width,
             height, min_width);
+    */
     pixDestroy(&pix);
     return nullptr;
   }
-- 
2.22.0