Natools

Check-in [3c8fd02a5c]
Login
Overview
Comment:smaz-tools: add a procedure to help evaluate and improve dictionaries
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: 3c8fd02a5c318a0f0c74a761488a7be16d7468cf
User & Date: nat on 2016-10-06 17:26:42
Other Links: manifest | tags
Context
2016-10-07
16:01
tools/smaz: first draft of CLI interface for Evaluate_Dictionary check-in: a27f42e127 user: nat tags: trunk
2016-10-06
17:26
smaz-tools: add a procedure to help evaluate and improve dictionaries check-in: 3c8fd02a5c user: nat tags: trunk
2016-10-05
14:49
tools/smaz: add a command-line option for maximum word size check-in: 2995e1835c user: nat tags: trunk
Changes

Modified src/natools-smaz-tools.adb from [a336f10fe0] to [627ca11783].

432
433
434
435
436
437
438























































439
440
441
442
443
444
445
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500







+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+








         if Next - First in Min_Size .. Max_Size then
            Add_Word (Counter, Phrase (First .. Next - 1));
         end if;
      end loop Main_Loop;
   end Add_Words;


   procedure Evaluate_Dictionary
     (Dict : in Dictionary;
      Corpus : in String_Lists.List;
      Compressed_Size : out Ada.Streams.Stream_Element_Count;
      Counts : out Dictionary_Counts)
   is
      Verbatim_Code_Count : constant Ada.Streams.Stream_Element_Offset
        := Ada.Streams.Stream_Element_Offset
           (Ada.Streams.Stream_Element'Last - Dict.Dict_Last);

      Verbatim_Length : Ada.Streams.Stream_Element_Offset;
      Input_Byte : Ada.Streams.Stream_Element;
   begin
      Compressed_Size := 0;

      for I in Counts'Range loop
         Counts (I) := 0;
      end loop;

      for S of Corpus loop
         declare
            use type Ada.Streams.Stream_Element_Offset;
            Compressed : constant Ada.Streams.Stream_Element_Array
              := Compress (Dict, S);
            Index : Ada.Streams.Stream_Element_Offset := Compressed'First;
         begin
            Compressed_Size := Compressed_Size + Compressed'Length;

            while Index in Compressed'Range loop
               Input_Byte := Compressed (Index);

               if Input_Byte in Dict.Offsets'Range then
                  Counts (Input_Byte) := Counts (Input_Byte) + 1;
                  Index := Index + 1;
               else
                  if not Dict.Variable_Length_Verbatim then
                     Verbatim_Length := Ada.Streams.Stream_Element_Offset
                       (Ada.Streams.Stream_Element'Last - Input_Byte) + 1;
                  elsif Input_Byte < Ada.Streams.Stream_Element'Last then
                     Verbatim_Length := Ada.Streams.Stream_Element_Offset
                       (Ada.Streams.Stream_Element'Last - Input_Byte);
                  else
                     Index := Index + 1;
                     Verbatim_Length := Ada.Streams.Stream_Element_Offset
                       (Compressed (Index)) + Verbatim_Code_Count - 1;
                  end if;

                  Index := Index + Verbatim_Length + 1;
               end if;
            end loop;
         end;
      end loop;
   end Evaluate_Dictionary;


   function Simple_Dictionary
     (Counter : in Word_Counter;
      Word_Count : in Natural)
     return String_Lists.List
   is
      use type Ada.Containers.Count_Type;

Modified src/natools-smaz-tools.ads from [a92069b766] to [d2561626f6].

100
101
102
103
104
105
106












107
108
109
110
111
112
113
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125







+
+
+
+
+
+
+
+
+
+
+
+








   function Simple_Dictionary
     (Counter : in Word_Counter;
      Word_Count : in Natural)
     return String_Lists.List;
      --  Return the Word_Count words in Counter that have the highest score,
      --  the score being count * length.

   type Dictionary_Counts is
     array (Ada.Streams.Stream_Element) of String_Count;

   procedure Evaluate_Dictionary
     (Dict : in Dictionary;
      Corpus : in String_Lists.List;
      Compressed_Size : out Ada.Streams.Stream_Element_Count;
      Counts : out Dictionary_Counts);
      --  Compress all strings of Corpus, returning the total number of
      --  compressed bytes and the number of uses for each dictionary
      --  element.

private

   package Word_Maps is new Ada.Containers.Indefinite_Ordered_Maps
     (String, String_Count);

   type Word_Counter is record