docling-parse 2.0.0__tar.gz → 2.0.1__tar.gz

{docling_parse-2.0.0 → docling_parse-2.0.1}/CHANGELOG.md

@@ -1,3 +1,9 @@
+## [v2.0.1](https://github.com/DS4SD/docling-parse/releases/tag/v2.0.1) - 2024-10-25
+
+### Fix
+
+* Robustify parser v2 ([#49](https://github.com/DS4SD/docling-parse/issues/49)) ([`1815e7d`](https://github.com/DS4SD/docling-parse/commit/1815e7d9400bd2551e99efb475fd59a7bf81069a))
+
 ## [v2.0.0](https://github.com/DS4SD/docling-parse/releases/tag/v2.0.0) - 2024-10-23
 
 ### Feature

{docling_parse-2.0.0 → docling_parse-2.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-parse
-Version: 2.0.0
+Version: 2.0.1
 Summary: Simple package to extract text with coordinates from programmatic PDFs
 Home-page: https://github.com/DS4SD/docling-parse
 License: MIT

{docling_parse-2.0.0 → docling_parse-2.0.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling-parse"
-version = "2.0.0"
+version = "2.0.1"
 description = "Simple package to extract text with coordinates from programmatic PDFs"
 authors = ["Peter Staar <taa@zurich.ibm.com>"]
 maintainers = [

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font/base_fonts.h

@@ -43,12 +43,15 @@ namespace pdflib
 
   private:
 
+    bool initialized;
+    
     std::set<std::string> core_14_fonts;
 
     std::map<std::string, base_font_type> name_to_basefont;
   };
 
-  base_fonts::base_fonts()
+  base_fonts::base_fonts():
+    initialized(false)
   {}
 
   base_fonts::~base_fonts()
@@ -159,6 +162,12 @@ namespace pdflib
   template<typename glyphs_type>
   void base_fonts::initialise(std::string dirname, glyphs_type& glyphs)
   {
+    if(initialized)
+      {
+	LOG_S(WARNING) << "skipping base_fonts::initialise, already initialized ...";
+	return;
+      }
+    
     std::vector<std::string> standard = utils::filesystem::list_files(dirname+"/standard");
     std::sort(standard.begin(), standard.end());
 
@@ -217,7 +226,9 @@ namespace pdflib
 	  {
 	    //LOG_S(WARNING) << "\t font-name (=" << fontname << ") already read";
 	  }
-      }    
+      }
+
+    initialized = true;
   }
 
   std::string base_fonts::read_fontname(std::string filename)

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font/cmap.h

@@ -504,7 +504,7 @@ namespace pdflib
 
         if(_map.count(begin+i)==1)
           {
-            LOG_S(FATAL) << "overwriting number c=" << begin+i;
+            LOG_S(ERROR) << "overwriting number c=" << begin+i;
           }
 
         _map[begin + i] = tgt.at(i);
@@ -514,18 +514,43 @@ namespace pdflib
   void cmap_parser::parse_beginbfchar(std::vector<qpdf_instruction>& parameters)
   {
     LOG_S(INFO) << __FUNCTION__;
-    assert(parameters.size()==1);
+    //assert(parameters.size()==1);
 
-    char_count = parameters[0].to_int();
+    if(parameters.size()==1)
+      {
+	char_count = parameters[0].to_int();
+      }
+    else if(parameters.size()>0)
+      {
+	LOG_S(WARNING) << "parameters.size()>0 for parse_beginbfchar";
+	char_count = parameters[0].to_int();
+      }    
+    else
+      {
+	LOG_S(ERROR) << "parameters.size()!=1 for parse_beginbfchar";
+      }
   }
 
   void cmap_parser::parse_endbfchar(std::vector<qpdf_instruction>& parameters)
   {
     LOG_S(INFO) << __FUNCTION__ << ": starting ...";
-    assert(parameters.size()==2*char_count);    
 
+    if(parameters.size()!=2*char_count)
+      {
+	LOG_S(WARNING) << "parameters.size()!=2*char_count -> "
+		       << "parameters: " << parameters.size() << ", "
+		       << "char_count: " << char_count;
+      }
+    //assert(parameters.size()==2*char_count);
+    
     for(size_t i=0; i<char_count; i++)
       {
+	if(2*i>=parameters.size())
+	  {
+	    LOG_S(ERROR) << "going out of bounds: skipping parse_endbfchar";
+	    continue;
+	  }
+	
         QPDFObjectHandle source_ = parameters[2*i+0].obj;
         QPDFObjectHandle target_ = parameters[2*i+1].obj;
 

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font/encodings.h

@@ -21,10 +21,13 @@ namespace pdflib
 
   private:
 
+    bool initialized;
+    
     std::map<font_encoding_name, font_encoding> name_to_encoding;
   };
 
-  font_encodings::font_encodings()
+  font_encodings::font_encodings():
+    initialized(false)
   {}
 
   font_encodings::~font_encodings()
@@ -38,6 +41,12 @@ namespace pdflib
   template<typename glyphs_type>
   void font_encodings::initialise(std::string dirname, glyphs_type& glyphs)
   {
+    if(initialized)
+      {
+	LOG_S(WARNING) << "skipping font_encodings::initialise, already initialized ...";
+	return;
+      }
+    
     std::vector<std::pair<font_encoding_name, std::string> > items = {
       {STANDARD, "std.dat"},
       {MACROMAN, "macroman.dat"},
@@ -50,6 +59,8 @@ namespace pdflib
         font_encoding& encoding = name_to_encoding[item.first];
         encoding.initialise(item.first, dirname+"/"+item.second, glyphs);
       }
+
+    initialized = true;
   }
 
 }

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font/font_cids.h

@@ -33,6 +33,7 @@ namespace pdflib
 
   private:
 
+    bool initialized;
     std::string directory;
 
     std::map<std::string, int>                       ro_2_sup;
@@ -45,7 +46,8 @@ namespace pdflib
     std::map<std::string, font_cid> cids;
   };
 
-  font_cids::font_cids()
+  font_cids::font_cids():
+    initialized(false)
   {}
 
   font_cids::~font_cids()
@@ -84,7 +86,13 @@ namespace pdflib
 
   void font_cids::initialise(std::string dirname)
   {
-    LOG_S(INFO) << __FUNCTION__;
+    if(initialized)
+      {
+	LOG_S(WARNING) << "skipping font_cids::initialise, already initialized ...";
+	return;
+      }
+    
+    LOG_S(INFO) << "initialise font_cids";
 
     directory = dirname;
     directory += (directory.back()=='/'? "" : "/");
@@ -126,6 +134,8 @@ namespace pdflib
 	    cmap_2_filename[file] = cdir+"/"+file;
 	  }
       }
+
+    initialized = true;
   }
 
   bool font_cids::decode_cmap_resource(std::string cmap_name)

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font/glyphs.h

@@ -28,7 +28,7 @@ namespace pdflib
     std::string operator[](std::string key);
 
     void initialise(std::string dirname);
-
+    
   private:
 
     void read_file_hex(std::string filename);
@@ -39,13 +39,16 @@ namespace pdflib
     
   private:
 
+    bool initialized;
+    
     std::set<std::string> unknown_glyphs;
 
     std::map<std::string, std::string> name_to_code;
     std::map<std::string, std::string> name_to_utf8;
   };
 
-  font_glyphs::font_glyphs()
+  font_glyphs::font_glyphs():
+    initialized(false)
   {}
 
   font_glyphs::~font_glyphs()
@@ -103,6 +106,12 @@ namespace pdflib
 
   void font_glyphs::initialise(std::string dirname)
   {
+    if(initialized)
+      {
+	LOG_S(WARNING) << "skipping font_glyphs::initialise, already initialized ...";
+	return;
+      }
+    
     LOG_S(INFO) << "font-glyphs initialise from directory: " 
 		<< dirname;
 
@@ -116,7 +125,7 @@ namespace pdflib
       "/custom/MathematicalPi/MathematicalPi.hex.dat"
     };
 
-    for(auto path : paths_hex)
+    for(auto path:paths_hex)
       {
         std::string fpath = dirname + path; 
         read_file_hex(fpath);
@@ -126,11 +135,13 @@ namespace pdflib
       "/custom/MathematicalPi/MathematicalPi.uni.dat"
     };
 
-    for(auto path : paths_uni)
+    for(auto path:paths_uni)
       {
         std::string fpath = dirname + path; 
         read_file_uni(fpath);
       }
+
+    initialized = true;
   }
 
   void font_glyphs::read_file_hex(std::string filename)
@@ -165,7 +176,7 @@ namespace pdflib
               {
                 name_to_utf8[key] = utils::string::hex_to_utf8(val_, 4);
               }
-            else if(name_to_utf8.count(key)==1)
+            else if(name_to_utf8.count(key)==1) // already present
               {
                 LOG_S(ERROR) << "key [" << key << "] is defined twice";               
               }

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_resources/page_font.h

@@ -688,7 +688,7 @@ namespace pdflib
     else
       {
         subtype=NULL_TYPE;
-        LOG_S(FATAL) << "could not find subtype in font: " << json_font.dump(2);
+        LOG_S(ERROR) << "could not find subtype in font: " << json_font.dump(2);
       }
   }
 
@@ -1050,15 +1050,21 @@ namespace pdflib
 
     if(values.size()!=(lchar-fchar+1))
       {
-        LOG_S(FATAL) << "values.size()!=(lchar-fchar+1) -> "
+        LOG_S(ERROR) << "values.size()!=(lchar-fchar+1) -> "
                      << values.size() << "!=" << lchar << "-" << fchar << "+1";
       }
 
     int cnt=0;
     for(int ind=fchar; ind<=lchar; ind++)
       {
+	if(cnt>=values.size())
+	  {
+	    LOG_S(ERROR) << "going out of bounds with " << cnt << " >= " << values.size();
+	    continue;
+	  }
+	
         numb_to_widths[ind] = values[cnt++];
-	LOG_S(INFO) << "index: " << ind << " -> width: " << numb_to_widths.at(ind);
+	//LOG_S(INFO) << "index: " << ind << " -> width: " << numb_to_widths.at(ind);
       }
   }
 

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_states/grph.h

@@ -72,7 +72,7 @@ namespace pdflib
     int line_cap;
     int line_join;
 
-    int                 dash_phase;
+    double              dash_phase;
     std::vector<double> dash_array;
 
     double flatness;
@@ -213,8 +213,20 @@ namespace pdflib
 	dash_array.push_back(val);
       }
 
-    assert(instructions[1].is_integer());
-    dash_phase = instructions[1].to_int();
+    if(instructions[1].is_integer())
+      {
+	dash_phase = instructions[1].to_int();
+      }
+    else if(instructions[1].is_number())
+      {
+	dash_phase = instructions[1].to_double();
+      }
+    else
+      {
+	dash_phase = 0;
+	LOG_S(ERROR) << "failed instructions[1] with is_integer() and is_number"
+		       << instructions[1].unparse();
+      }
   }
 
   void pdf_state<GRPH>::ri(std::vector<qpdf_instruction>& instructions)
@@ -227,8 +239,16 @@ namespace pdflib
     //assert(instructions.size()==1);
     if(not verify(instructions, 1, __FUNCTION__) ) { return; }
     
-    assert(instructions[0].is_number());    
-    flatness = instructions[0].to_double();    
+    if(instructions[0].is_number())
+      {
+	flatness = instructions[0].to_double();
+      }
+    else
+      {
+	flatness = 0;
+	LOG_S(ERROR) << "failed instructions[0].is_number(): "
+		       << instructions[0].unparse();	
+      }
   }
 
   void pdf_state<GRPH>::gs(std::vector<qpdf_instruction>& instructions)

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_states/line.h

@@ -518,6 +518,13 @@ namespace pdflib
 
     // first close
     auto& line = curr_lines.back();
+
+    if(line.size()==0)
+      {
+        LOG_S(WARNING) << "applying 'h' on empty line";
+        return;
+      }
+    
     std::pair<double, double> coor = line.front();
 
     line.append(coor.first, coor.second);

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/pdf_states/text.h

@@ -260,7 +260,7 @@ namespace pdflib
       }
     else
       {
-        LOG_S(FATAL) << "unknown page-font: " << font_name;
+        LOG_S(ERROR) << "unknown page-font: " << font_name;
       }
   }
 
@@ -313,8 +313,9 @@ namespace pdflib
           }
         else
           {
-            LOG_S(FATAL) << "item is not a string nor a value: "
-                         << item.unparse() << " [" << item.getTypeName() << "]";
+            LOG_S(ERROR) << "item is not a string nor a value: "
+                         << item.unparse() << " [" << item.getTypeName() << "]"
+			 << " -> skipping for now ...";
           }
       }
   }

{docling_parse-2.0.0 → docling_parse-2.0.1}/src/v2/qpdf/to_json.h

@@ -16,8 +16,8 @@ namespace pdflib
   // FIXME: add a begin time to cap the max time spent in this routine
   nlohmann::json to_json(QPDFObjectHandle obj, std::set<std::string> prev_objs={}, int level=0)
   {
-    //const static int max_level=32;
-    const static int max_level=128;
+    const static int max_level=32;
+    //const static int max_level=128;
     
     LOG_S(INFO) << "to_json (level=" << level << "): " << prev_objs.size();
     
@@ -47,7 +47,7 @@ namespace pdflib
     
     if(level<max_level)
       {
-	const static std::set<std::string> keys_to_be_skipped = {"/Parent", "/P", "/Annots"};
+	const static std::set<std::string> keys_to_be_skipped = {"/Parent", "/P", "/Annots", "/B"};
 	
         if(obj.isDictionary())
           {