semacode 0.7.0

data/ext/iec16022ecc200.h

@@ -0,0 +1,27 @@
+// IEC16022 bar code generation library
+// Adrian Kennard, Andrews & Arnold Ltd
+// with help from Cliff Hones on the RS coding
+//
+// Revision 1.3  2004/09/09 07:45:09  cvs
+// Added change history to source files
+// Added "info" type to IEC16022
+// Added exact size checking shortcodes on encoding generation for iec16022
+//
+
+// Main encoding function
+// Returns the grid (malloced) containing the matrix. L corner at 0,0.
+// Takes suggested size in *Wptr, *Hptr, or 0,0. Fills in actual size.
+// Takes barcodelen and barcode to be encoded
+// Note, if *encodingptr is null, then fills with auto picked (malloced) encoding
+// If lenp not null, then the length of encoded data before any final unlatch or pad is stored
+// If maxp not null, then the max storage of this size code is stored
+// If eccp not null, then the number of ecc bytes used in this size is stored
+// Returns 0 on error (writes to stderr with details).
+
+#define MAXBARCODE 3116
+
+unsigned char *
+iec16022ecc200 (int *Wptr, int *Hptr, char **encodingptr, int barcodelen, unsigned char *barcode, int *lenp,int *maxp,int *eccp);
+
+
+

data/ext/reedsol.c

@@ -0,0 +1,170 @@
+// reedsol.c
+//
+// This is a simple Reed-Solomon encoder
+// (C) Cliff Hones 2004
+//
+// It is not written with high efficiency in mind, so is probably
+// not suitable for real-time encoding.  The aim was to keep it
+// simple, general and clear.
+//
+// <Some notes on the theory and implementation need to be added here>
+
+// Usage:
+// First call rs_init_gf(poly) to set up the Galois Field parameters.
+// Then  call rs_init_code(size, index) to set the encoding size
+// Then  call rs_encode(datasize, data, out) to encode the data.
+//
+// These can be called repeatedly as required - but note that
+// rs_init_code must be called following any rs_init_gf call.
+//
+// If the parameters are fixed, some of the statics below can be
+// replaced with constants in the obvious way, and additionally
+// malloc/free can be avoided by using static arrays of a suitable
+// size.
+
+#include <stdio.h>              // only needed for debug (main)
+#include <stdlib.h>             // only needed for malloc/free
+
+static int gfpoly;
+static int symsize;             // in bits
+static int logmod;              // 2**symsize - 1
+static int rlen;
+
+static int *log = NULL,
+   *alog = NULL,
+   *rspoly = NULL;
+
+// rs_init_gf(poly) initialises the parameters for the Galois Field.
+// The symbol size is determined from the highest bit set in poly
+// This implementation will support sizes up to 30 bits (though that
+// will result in very large log/antilog tables) - bit sizes of
+// 8 or 4 are typical
+//
+// The poly is the bit pattern representing the GF characteristic
+// polynomial.  e.g. for ECC200 (8-bit symbols) the polynomial is
+// a**8 + a**5 + a**3 + a**2 + 1, which translates to 0x12d.
+
+void
+rs_init_gf (int poly)
+{
+   int m,
+     b,
+     p,
+     v;
+
+   // Return storage from previous setup
+   if (log)
+   {
+      free (log);
+      free (alog);
+      free (rspoly);
+      rspoly = NULL;
+   }
+   // Find the top bit, and hence the symbol size
+   for (b = 1, m = 0; b <= poly; b <<= 1)
+      m++;
+   b >>= 1;
+   m--;
+   gfpoly = poly;
+   symsize = m;
+
+   // Calculate the log/alog tables
+   logmod = (1 << m) - 1;
+   log = (int *) malloc (sizeof (int) * (logmod + 1));
+   alog = (int *) malloc (sizeof (int) * logmod);
+
+   for (p = 1, v = 0; v < logmod; v++)
+   {
+      alog[v] = p;
+      log[p] = v;
+      p <<= 1;
+      if (p & b)
+         p ^= poly;
+   }
+}
+
+// rs_init_code(nsym, index) initialises the Reed-Solomon encoder
+// nsym is the number of symbols to be generated (to be appended
+// to the input data).  index is usually 1 - it is the index of
+// the constant in the first term (i) of the RS generator polynomial:
+// (x + 2**i)*(x + 2**(i+1))*...   [nsym terms]
+// For ECC200, index is 1.
+
+void
+rs_init_code (int nsym, int index)
+{
+   int i,
+     k;
+
+   if (rspoly)
+      free (rspoly);
+   rspoly = (int *) malloc (sizeof (int) * (nsym + 1));
+
+   rlen = nsym;
+
+   rspoly[0] = 1;
+   for (i = 1; i <= nsym; i++)
+   {
+      rspoly[i] = 1;
+      for (k = i - 1; k > 0; k--)
+      {
+         if (rspoly[k])
+            rspoly[k] = alog[(log[rspoly[k]] + index) % logmod];
+         rspoly[k] ^= rspoly[k - 1];
+      }
+      rspoly[0] = alog[(log[rspoly[0]] + index) % logmod];
+      index++;
+   }
+}
+
+// Note that the following uses byte arrays, so is only suitable for
+// symbol sizes up to 8 bits.  Just change the data type of data and res
+// to unsigned int * for larger symbols.
+
+void
+rs_encode (int len, unsigned char *data, unsigned char *res)
+{
+   int i,
+     k,
+     m;
+   for (i = 0; i < rlen; i++)
+      res[i] = 0;
+   for (i = 0; i < len; i++)
+   {
+      m = res[rlen - 1] ^ data[i];
+      for (k = rlen - 1; k > 0; k--)
+      {
+         if (m && rspoly[k])
+            res[k] = res[k - 1] ^ alog[(log[m] + log[rspoly[k]]) % logmod];
+         else
+            res[k] = res[k - 1];
+      }
+      if (m && rspoly[0])
+         res[0] = alog[(log[m] + log[rspoly[0]]) % logmod];
+      else
+         res[0] = 0;
+   }
+}
+
+#ifndef LIB
+// The following tests the routines with the ISO/IEC 16022 Annexe R data
+int
+main (void)
+{
+   register int i;
+
+   unsigned char data[9] = { 142, 164, 186 };
+   unsigned char out[5];
+
+   rs_init_gf (0x12d);
+   rs_init_code (5, 1);
+
+   rs_encode (3, data, out);
+
+   printf ("Result of Annexe R encoding:\n");
+   for (i = 4; i >= 0; i--)
+      printf ("  %d\n", out[i]);
+
+   return 0;
+}
+#endif

data/ext/reedsol.h

@@ -0,0 +1,5 @@
+
+void rs_init_gf(int poly);
+void rs_init_code(int nsym, int index);
+void rs_encode(int len, unsigned char *data, unsigned char *res);
+

data/ext/semacode.c

@@ -0,0 +1,359 @@
+/*
+
+  == Introduction
+  
+  This Ruby extension implements a DataMatrix encoder for Ruby. It is
+  typically used to create semacodes, which are barcodes, that contain URLs.
+  This encoder does not create image files or visual representations of the
+  semacode. This is because it can be used for more than creating images, such
+  as rendering semacodes to HTML, SVG, PDF or even stored in a database or
+  file for later use.
+
+  Author::    Guido Sohne  (mailto:guido@sohne.net)
+  Copyright:: Copyright (c) 2007 Guido Sohne
+  License::   GNU GPL version 2, available from http://www.gnu.org
+*/
+
+#include "ruby.h"
+#include "semacode.h"
+
+/*
+
+Internal function that encodes a string of given length, storing
+the encoded result into an internal, private data structure. This
+structure is consulted for any operations, such as to get the 
+semacode dimensions. It deallocates any previous data before 
+generating a new encoding.
+
+Due to a bug in the underlying encoder, we do two things
+
+ * append a space character before encoding, to get around
+   an off by one error lurking in the C code
+   
+ * manually select the best barcode dimensions, to avoid
+   an encoder bug where sometimes no suitable encoding would
+   be found
+
+*/
+semacode_t* 
+encode_string(semacode_t *semacode, int message_length, char *message)
+{
+  /* deallocate if exists already */
+  if(semacode !=NULL && semacode->data != NULL)
+    free(semacode->data);
+    
+  bzero(semacode, sizeof(semacode_t));
+  
+  // work around encoding bug by appending an extra character.
+  strcat(message, " ");
+  message_length++;
+  
+  // choose the best grid that will hold our message
+  iec16022init(&semacode->width, &semacode->height, message);
+  
+  // encode the actual data
+  semacode->data = (char *) iec16022ecc200(
+    &semacode->width, 
+    &semacode->height, 
+    NULL, 
+    message_length, 
+    (unsigned char *) message, 
+    &semacode->raw_encoded_length,
+    &semacode->symbol_capacity, 
+    &semacode->ecc_bytes);
+
+  return semacode;
+}
+
+/* our module and class, respectively */
+
+static VALUE rb_mSemacode;
+static VALUE rb_cEncoder;
+
+static void
+semacode_mark(semacode_t *semacode)
+{
+  /* need this if we ever hold any other Ruby objects. so let's not go there. */
+}
+
+static void
+semacode_free(semacode_t *semacode)
+{
+  if(semacode != NULL) {
+    if(semacode->data != NULL)
+      free(semacode->data);
+    bzero(semacode, sizeof(semacode));
+    free(semacode);
+  }
+}
+
+static VALUE
+semacode_allocate(VALUE klass)
+{
+  semacode_t *semacode;
+  return Data_Make_Struct(klass, semacode_t, semacode_mark, semacode_free, semacode); 
+}
+
+/* 
+  Initialize the semacode. This function is called after a semaode is
+  created. Ruby objects are created using a new method, and then initialized
+  via the 'initialize' method once they have been allocated.
+  
+  The initializer takes a single argument, which can be anything that 
+  responds to the 'to_s' method - that is, anything string like.
+  
+  The string in the argument is encoded and the semacode is returned
+  initialized and ready for use.
+  
+*/
+static VALUE
+semacode_init(VALUE self, VALUE message)
+{
+  semacode_t *semacode;
+  
+  if (!rb_respond_to(message, rb_intern ("to_s")))
+      rb_raise(rb_eRuntimeError, "target must respond to 'to_s'");
+
+  Data_Get_Struct(self, semacode_t, semacode);
+  encode_string(semacode, StringValueLen(message), StringValuePtr(message));
+  
+  return self;
+}
+
+/*
+  This function turns the raw output from an encoding into a more
+  friendly format organized by rows and columns.
+  
+  It returns the semacode matrix as an array of arrays of boolean. The
+  first element in the array is the top row, the last is the bottom row.
+  
+  Each row is also an array, containing boolean values. The length of each
+  row is the same as the semacode width, and the number of rows is the same
+  as the semacode height.
+
+*/
+static VALUE
+semacode_grid(semacode_t *semacode)
+{
+  int w = semacode->width;
+  int h = semacode->height;
+  
+  VALUE ret = rb_ary_new2(h);
+
+  int x, y;
+	for (y = h - 1; y >= 0; y--) {
+	  VALUE ary = rb_ary_new2(w);
+		for (x = 0; x < w; x++) {
+		  if(semacode->data[y * w + x])
+		    rb_ary_push(ary, Qtrue);
+		  else
+		    rb_ary_push(ary, Qfalse);
+		}
+		rb_ary_push(ret, ary);
+	}
+	
+	return ret;
+}
+
+/*
+  This function turns the raw output from an encoding into a string
+  representation.
+  
+  It returns the semacode matrix as a comma-separated list of character
+  vectors (sequence of characters). The top row is the first vector and
+  the bottow row is the last vector.
+  
+  Each vector is a sequence of characters, either '1' or '0', to represent
+  the bits of the semacode pattern. The length of a vector is the semacode
+  width, and the number of vectors is the same as the semacode height.
+
+*/
+static VALUE
+semacode_to_s(VALUE self)
+{
+  semacode_t *semacode;
+  VALUE str;
+  int x, y;  
+  int w, h;
+  
+  Data_Get_Struct(self, semacode_t, semacode);
+  
+  if(semacode == NULL || semacode->data == NULL)
+    return Qnil;
+  
+  w = semacode->width;
+  h = semacode->height;
+  
+  str = rb_str_new2("");
+  
+	for (y = h - 1; y >= 0; y--) {
+		for (x = 0; x < w; x++) {
+		  if(semacode->data[y * w + x])
+		    rb_str_cat(str, "1", 1);
+		  else
+		    rb_str_cat(str, "0", 1);
+		}
+		rb_str_cat(str, ",", 1);
+	}
+	
+	return str;
+}
+/*
+
+  After creating a semacode, it is possible to reuse the semacode object
+  if you want to encode another URL. You should call the 'encode' method
+  at any time to create a replacement semecode for the current object.
+  
+  It returns the semacode matrix as an array of arrays of boolean. The
+  first element in the array is the top row, the last is the bottom row.
+  
+  Each row is also an array, containing boolean values. The length of each
+  row is the same as the semacode width, and the number of rows is the same
+  as the semacode height.
+
+*/
+static VALUE
+semacode_encode(VALUE self, VALUE message)
+{
+  semacode_t *semacode;
+  
+  if (!rb_respond_to(message, rb_intern ("to_s")))
+      rb_raise(rb_eRuntimeError, "target must respond to 'to_s'");
+  
+  Data_Get_Struct(self, semacode_t, semacode);
+  
+  /* free previous string if that exists */
+  if(semacode->data != NULL) {
+    free(semacode->data);
+    semacode->data == NULL;
+  }
+  
+  /* do a new encoding */
+  DATA_PTR(self) = encode_string(semacode, StringValueLen(message), StringValuePtr(message));
+
+  return semacode_grid(semacode);
+}
+
+/*
+  This function gives the encoding organized by rows and columns.
+  
+  It returns the semacode matrix as an array of arrays of boolean. The
+  first element in the array is the top row, the last is the bottom row.
+  
+  Each row is also an array, containing boolean values. The length of each
+  row is the same as the semacode width, and the number of rows is the same
+  as the semacode height.
+
+*/
+static VALUE
+semacode_data(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  if(semacode->data == NULL)
+    return Qnil;
+  else
+    return semacode_grid(semacode);
+}
+
+/*  
+  This returns the width of the semacode.
+*/
+static VALUE
+semacode_width(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->width);
+}
+
+/*  
+  This returns the height of the semacode.
+*/
+static VALUE
+semacode_height(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->height);
+}
+
+/*  
+  This returns the length of the semacode. It is
+  the same as the product of the height and the width.
+*/
+static VALUE
+semacode_length(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->height * semacode->width);
+}
+
+/*  
+  This returns the length of the raw underlying encoding
+  representing the data, before padding, error correction
+  or any other operations on the raw encoding.
+*/
+static VALUE
+semacode_raw_encoded_length(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->raw_encoded_length);
+}
+
+/*  
+  This returns the maximum number of characters that can
+  be stored in a symbol of the given width and height. You
+  can use this to decide if it is worth packing in extra
+  information while keeping the symbol size the same.
+*/
+static VALUE
+semacode_symbol_size(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->symbol_capacity);
+}
+
+/*  
+  This returns the number of bytes that are being devoted to
+  error correction.
+*/
+static VALUE
+semacode_ecc_bytes(VALUE self)
+{
+  semacode_t *semacode;
+  Data_Get_Struct(self, semacode_t, semacode);
+
+  return INT2FIX(semacode->ecc_bytes);
+}
+
+void 
+Init_semacode()
+{
+  rb_mSemacode = rb_define_module ("DataMatrix");
+  rb_cEncoder = rb_define_class_under(rb_mSemacode, "Encoder", rb_cObject);
+  rb_define_alloc_func(rb_cEncoder, semacode_allocate);
+  
+  rb_define_method(rb_cEncoder, "initialize", semacode_init, 1);
+  rb_define_method(rb_cEncoder, "encode", semacode_encode, 1);  
+  rb_define_method(rb_cEncoder, "to_a", semacode_data, 0);
+  rb_define_method(rb_cEncoder, "data", semacode_data, 0);  
+  rb_define_method(rb_cEncoder, "to_s", semacode_to_s, 0);
+  rb_define_method(rb_cEncoder, "to_str", semacode_to_s, 0);  
+  rb_define_method(rb_cEncoder, "width", semacode_width, 0);    
+  rb_define_method(rb_cEncoder, "height", semacode_height, 0);
+  rb_define_method(rb_cEncoder, "length", semacode_length, 0);
+  rb_define_method(rb_cEncoder, "size", semacode_length, 0);  
+  rb_define_method(rb_cEncoder, "raw_encoded_length", semacode_raw_encoded_length, 0);    
+  rb_define_method(rb_cEncoder, "symbol_size", semacode_symbol_size, 0);    
+  rb_define_method(rb_cEncoder, "ecc_bytes", semacode_ecc_bytes, 0);    
+}