Changeset 1243 for cpp/frams/genetics

Timestamp:

05/19/23 19:12:22 (20 months ago)

Author:

Maciej Komosinski

Message:

Simplify modifiers 'r' and 'R' separately because they obey simpler rules than other modifiers and thus can be reduced more

Location:

cpp/frams/genetics

Files:

• genooperators.cpp (modified) (3 diffs)
• genooperators.h (modified) (1 diff)

cpp/frams/genetics/genooperators.cpp

@@ -472 +472 @@
 }
 
+string GenoOperators::simplifiedModifiers_rR(const string& str)
+{
+        int R = 0; //positive means more 'R', negative means more 'r'
+        for (char c : str)
+        {
+                if (c == 'R') R++; else
+                        if (c == 'r') R--;
+        }
+        R %= 8; // 8 * 45 degrees = 360 degrees. After this, we get R=-7..+7
+
+        /* now, simplify homogeneous sequences of rR longer than 4: for example, rrrrr == RRR and RRRRRR == rr
+        -7      1
+        -6      2
+        -5      3
+        -4      -4 (or 4; we choose +4 meaning we will never see rrrr)
+        -3..3      (no changes)
+        4       4 (or -4)
+        5       -3
+        6       -2
+        7       -1
+        */
+        if (R <= -4) R += 8; //-4 => +4
+        else if (R >= 5) R -= 8;
+
+        return R == 0 ? "" : (R > 0 ? string(R, 'R') : string(-R, 'r'));
+}
+
 //#include <cassert>
 string GenoOperators::simplifiedModifiersFixedOrder(const char *str_of_char_pairs, vector<int> &char_counts)
 {
-//      assert(strlen(str_of_char_pairs) == char_counts.size());
-//      assert(char_counts.size() % 2 == 0);
-        const int MAX_NUMBER_SAME_TYPE = 8; // max. number of modifiers of each type (case-sensitive) - mainly for rR, even though for rR, 4 would be sufficient if we assume lower or upper can be chosen as required for minimal length, e.g. rrrrr==RRR, RRRRRR==rr
+        //      assert(strlen(str_of_char_pairs) == char_counts.size());
+        //      assert(char_counts.size() % 2 == 0);
+        const int MAX_NUMBER_SAME_TYPE = 8; // max. number of modifiers of each type (case-sensitive) - mainly for rR, even though for rR, 4 would be sufficient if we assume lower or upper can be chosen as required for minimal length just as simplifiedModifiers_rR() does, e.g. rrrrr==RRR, RRRRRR==rr
         string simplified;
-        //#define CLUMP_IDENTICAL_MODIFIERS //not good because with the exception of rR properties are calculated incrementally, non-linearly, and their values are updated after each modifier character, so these values may for example saturate after a large number of identical modifier symbols. The order of modifiers is (with the exception of rR) relevant and extreme values of properties increase this relevance, so better keep the modifiers dispersed.
+        //#define CLUMP_IDENTICAL_MODIFIERS //if GeneProps::normalizeBiol4() is used, this is not good because properties are calculated incrementally, non-linearly, their values are updated after each modifier character and some properties interact with each other due to normalization so they can saturate when clumped, therefore it is better keep the modifiers dispersed to equalize their effects
 #ifdef CLUMP_IDENTICAL_MODIFIERS
         for (size_t i = 0; i < strlen(str_of_char_pairs); i++)
@@ -510 +537 @@
 string GenoOperators::simplifiedModifiers(const string & original)
 {
-        const int MAX_NUMBER_SAME_TYPE = 6; // max. number of modifiers of each type (case-insensitive). rR could be treated separately in simplification because their influence follows different (i.e., simple additive) logic - so the simplifiedModifiersFixedOrder() logic with cancelling out is appropriate for rR. However in this function, making no exception to rR does not cause any harm to these modifiers either - the only consequence is that we will not remove antagonistic letters and will not simplify sequences of rR longer than 4, while they could be simplified (e.g. rrrrr==RRR, RRRRRR==rr).
+        const int MAX_NUMBER_SAME_TYPE = 6; // max. number of modifiers of each type (case-insensitive). rR is treated separately in simplification because their influence follows different (i.e., simple additive) logic - so the simplifiedModifiersFixedOrder() logic with cancelling out antagonistic modifiers is appropriate for rR.
         int counter[256] = {}; //initialize with zeros; 256 is unnecessarily too big and redundant, but enables very fast access (indexed directly by the ascii code)
         string simplified = "";
-        for (int i = original.size() - 1; i >= 0; i--) //iterate from end to begin - easier to remove "oldest" = first modifiers
+        for (int i = original.size() - 1; i >= 0; i--) //iterate from end to begin so it is easier to remove "oldest" = first modifiers
         {
                 unsigned char c = original[i];
-                if (!std::isalpha(c))
+                if (!std::isalpha(c) || c == 'r' || c == 'R') //ignore non-alphabet characters; also, 'r' and 'R' are handled separately by simplifiedModifiers_rR()
                         continue;
                 unsigned char lower = std::tolower(c);
@@ -524 +551 @@
         }
         std::reverse(simplified.begin(), simplified.end()); //"simplified" was built in reverse order, so need to restore the order that corresponds to "original"
-        return simplified;
-}
+        return simplifiedModifiers_rR(original) + simplified;
+}

cpp/frams/genetics/genooperators.h

@@ -218 +218 @@
 
         static int getRandomChar(const char *choices, const char *excluded); ///<returns index of a random character from \a choices excluding \a excluded, or -1 when everything is excluded or \a choices is empty.
-        static string simplifiedModifiersFixedOrder(const char *str_of_char_pairs, vector<int> &char_counts); ///<returns a sequence of chars from \a str_of_char_pairs based on how many times each char occurred in \a char_counts. Assume that an even-index char and the following odd-index char have the opposite influence, so they cancel out. We don't use this function, because a fixed order imposed by this function means that the number of different parameter values produced by a sequence of modifiers is lowered (N same-letter upper- and lower-case chars yield only 2*N different values). Due to how modifiers work, the effect of aaA, aAa, Aaa etc. is different (N same-letter upper- and lower-case chars yield 2^N different values), so simplifying modifiers should not impose any order and should not interfere with their original order - see \a simplifiedModifiers().
+        static string simplifiedModifiers_rR(const string& str); ///<finds all 'r' and 'R' in \a str and returns the shortest sequence of 'r' and 'R that is equivalent to all these found in \a str.
+        static string simplifiedModifiersFixedOrder(const char *str_of_char_pairs, vector<int> &char_counts); ///<returns a sequence of chars from \a str_of_char_pairs based on how many times each char occurred in \a char_counts. Assume that an even-index char and the following odd-index char have the opposite influence, so they cancel out. We don't use this function, because a fixed order imposed by this function means that the number of different parameter values produced by a sequence of modifiers is lowered (N same-letter upper- and lower-case chars yield only 2*N different values). Due to how modifiers work, the effect of aaA, aAa, Aaa etc. is different (N same-letter upper- and lower-case chars yield 2^N different values), so simplifying modifiers should not impose any order, should not interfere with their original order, and should not cancel out antagonistic modifiers - see \a simplifiedModifiers() and geneprops_test.cpp.
         //@}
         static string simplifiedModifiers(const string &original); ///<from the \a original sequence removes modifiers that are too numerous (exceeding a defined threshold number), starting the removal from the leftmost (="oldest" when interpreting the sequence from left to right) ones.