From 325a5376feef0bbae03fc3c4ac0a6ce9710a7c94 Mon Sep 17 00:00:00 2001 From: Hossein Moein Date: Fri, 3 May 2024 10:14:18 -0400 Subject: [PATCH] Implemented PeaksAndValleysVisitor visitor --- docs/HTML/DataFrame.html | 4 + docs/HTML/MACDVisitor.html | 2 +- docs/HTML/PeaksAndValleysVisitor.html | 163 ++++++++++++++++++ .../DataFrame/DataFrameFinancialVisitors.h | 86 ++++----- include/DataFrame/DataFrameStatsVisitors.h | 2 +- test/dataframe_tester_3.cc | 75 +++++++- 6 files changed, 285 insertions(+), 47 deletions(-) create mode 100644 docs/HTML/PeaksAndValleysVisitor.html diff --git a/docs/HTML/DataFrame.html b/docs/HTML/DataFrame.html index 827992e0..5a9d6d11 100644 --- a/docs/HTML/DataFrame.html +++ b/docs/HTML/DataFrame.html @@ -956,6 +956,10 @@

API Reference with code samples

struct ParkinsonVolVisitor{} + + struct PeaksAndValleysVisitor{} + + struct PercentPriceOSCIVisitor{} diff --git a/docs/HTML/MACDVisitor.html b/docs/HTML/MACDVisitor.html index 21c7934d..c0623ff4 100644 --- a/docs/HTML/MACDVisitor.html +++ b/docs/HTML/MACDVisitor.html @@ -41,7 +41,7 @@ - 
#include <DataFrame/DataFrameStatsVisitors.h>
+        
#include <DataFrame/DataFrameFinancialVisitors.h>
 
 template<typename T, typename I = unsigned long,
          std::size_t A = 0>
diff --git a/docs/HTML/PeaksAndValleysVisitor.html b/docs/HTML/PeaksAndValleysVisitor.html
new file mode 100644
index 00000000..c8e60571
--- /dev/null
+++ b/docs/HTML/PeaksAndValleysVisitor.html
@@ -0,0 +1,163 @@
+
+
+
+
+
+
+
+
+
+  
+
+    
+      
+    
+    
+      
+      
+      
+    
+
+  
Signature Description Parameters
 
+        
#include <DataFrame/DataFrameFinancialVisitors.h>
+
+template<typename T, typename I = unsigned long,
+         std::size_t A = 0>
+struct PeaksAndValleysVisitor;
+
+// -------------------------------------
+
+template<typename T, typename I = unsigned long,
+         std::size_t A = 0>
+using pav_v = PeaksAndValleysVisitor<T, I, A>;
+        

+      		
+        This calculates peaks and valleys in a time-series. A peak is a datapoint that is higher than the previous and next datapoints. Similarly, a valley is a datapoint that is lower than the previous and next datapoints. If these conditions are not met, the previous values of peaks and valleys are copied over.

+        The results are two vectors, namely peaks and valleys with the same size as the input time-series.

+        get_result() returns the vector of peaks.

+        get_peaks() returns the vector of peaks.

+        get_valleys() returns the vector of valleys.


+
+        Result vectors are vectors of this struct:
+        

+  struct ResultItem  {
+
+      T   value { };  // Peak or valley
+      I   index { };  // Index corresponding to the value
+
+      // Convenient print function
+      template<typename S>
+      friend S &operator << (S &stream, const ResultItem &ri)  {
+
+          stream << ri.value << '@' << ri.index;
+          return (stream);
+      }
+  };
+        

+      		
+        T: Column data type

+        I: Index type

+        A: Memory alignment boundary for vectors. Default is system default alignment

+      

+
+
static void test_PeaksAndValleysVisitor()  {
+
+    std::cout << "\nTesting PeaksAndValleysVisitor{ } ..." << std::endl;
+
+    typedef StdDataFrame64<std::string> StrDataFrame;
+
+    StrDataFrame    df;
+
+    try  {
+        df.read("SHORT_IBM.csv", io_format::csv2);
+
+        pav_v<double, std::string, 64>  pav;
+
+        df.single_act_visit<double>("IBM_Close", pav);
+
+        assert(pav.get_peaks().size() == 1721);
+        assert(std::abs(pav.get_peaks()[0].value - 185.53) < 0.0001);
+        assert(pav.get_peaks()[0].index == "2014-01-02");
+        assert(std::abs(pav.get_peaks()[1].value - 186.64) < 0.0001);
+        assert(pav.get_peaks()[1].index == "2014-01-03");
+        assert(std::abs(pav.get_peaks()[19].value - 182.73) < 0.0001);
+        assert(pav.get_peaks()[19].index == "2014-01-23");
+        assert(std::abs(pav.get_peaks()[20].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[20].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[24].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[24].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[25].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[25].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[1720].value - 116.0) < 0.0001);
+        assert(pav.get_peaks()[1720].index == "2020-10-23");
+        assert(std::abs(pav.get_peaks()[1712].value - 125.93) < 0.0001);
+        assert(pav.get_peaks()[1712].index == "2020-10-16");
+        assert(std::abs(pav.get_peaks()[1707].value - 131.49) < 0.0001);
+        assert(pav.get_peaks()[1707].index == "2020-10-08");
+
+        assert(pav.get_valleys().size() == 1721);
+        assert(std::abs(pav.get_valleys()[0].value - 185.53) < 0.0001);
+        assert(pav.get_valleys()[0].index == "2014-01-02");
+        assert(std::abs(pav.get_valleys()[1].value - 186.64) < 0.0001);
+        assert(pav.get_valleys()[1].index == "2014-01-03");
+        assert(std::abs(pav.get_valleys()[19].value - 176.4) < 0.0001);
+        assert(pav.get_valleys()[19].index == "2014-01-29");
+        assert(std::abs(pav.get_valleys()[20].value - 176.4) < 0.0001);
+        assert(pav.get_valleys()[20].index == "2014-01-29");
+        assert(std::abs(pav.get_valleys()[24].value - 172.84) < 0.0001);
+        assert(pav.get_valleys()[24].index == "2014-02-04");
+        assert(std::abs(pav.get_valleys()[25].value - 172.84) < 0.0001);
+        assert(pav.get_valleys()[25].index == "2014-02-04");
+        assert(std::abs(pav.get_valleys()[1720].value - 106.65) < 0.0001);
+        assert(pav.get_valleys()[1720].index == "2020-10-28");
+        assert(std::abs(pav.get_valleys()[1712].value - 124.89) < 0.0001);
+        assert(pav.get_valleys()[1712].index == "2020-10-15");
+        assert(std::abs(pav.get_valleys()[1707].value - 121.97) < 0.0001);
+        assert(pav.get_valleys()[1707].index == "2020-10-06");
+    }
+    catch (const DataFrameError &ex)  {
+        std::cout << ex.what() << std::endl;
+    }
+}
+

+
+  
C++ DataFrame
+
+
+
+
+
diff --git a/include/DataFrame/DataFrameFinancialVisitors.h b/include/DataFrame/DataFrameFinancialVisitors.h
index 9d9d68ce..d0392a82 100644
--- a/include/DataFrame/DataFrameFinancialVisitors.h
+++ b/include/DataFrame/DataFrameFinancialVisitors.h
@@ -8225,62 +8225,70 @@ struct  QuantQualEstimationVisitor  {
 template
 using qqe_v = QuantQualEstimationVisitor;
 
+// ----------------------------------------------------------------------------
 
+template
+struct  PeaksAndValleysVisitor  {
 
+    struct ResultItem  {
 
+        T   value { };
+        I   index { };
 
+        template
+        friend S &operator << (S &stream, const ResultItem &ri)  {
 
+            stream << ri.value << '@' << ri.index;
+            return (stream);
+        }
+    };
 
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-// ----------------------------------------------------------------------------
-
-template
-struct  PeaksAndValleysVisitor  {
-
-    DEFINE_VISIT_BASIC_TYPES_3
+    using value_type = T;
+    using index_type = I;
+    using size_type = std::size_t;
+    using result_type =
+        std::vector::type>;
 
     template 
     inline void
-    operator() (const K &/*idx_begin*/, const K &/*idx_end*/,
+    operator() (const K &idx_begin, const K &/*idx_end*/,
                 const H &prices_begin, const H &prices_end)  {
 
         const size_type col_s = std::distance(prices_begin, prices_end);
 
 #ifdef HMDF_SANITY_EXCEPTIONS
         if (col_s < 4)
-            throw DataFrameError("PeaksAndValleysVisitor: column size must "
-                                 "> 3");
+            throw DataFrameError(
+                "PeaksAndValleysVisitor: column size must > 3");
 #endif // HMDF_SANITY_EXCEPTIONS
 
+        result_type peaks (col_s);
+        result_type valleys (col_s);
+
+        peaks[0] = valleys[0] = ResultItem { *prices_begin, *idx_begin };
+        peaks[1] = valleys[1] =
+            ResultItem { *(prices_begin + 1), *(idx_begin + 1) };
+        for (size_type i = 2; i < col_s; ++i)  {
+            auto  &curr_peak = peaks[i];
+            auto  &curr_valley = valleys[i];
+
+            curr_peak = peaks[i - 1];
+            curr_valley = valleys[i - 1];
+
+            const auto  val = *(prices_begin + i);
+            const auto  val_1 = *(prices_begin + (i - 1));
+            const auto  idx_1 = *(idx_begin + (i - 1));
+            const auto  val_2 = *(prices_begin + (i - 2));
+
+            if (val_1 > val && val_1 > val_2)
+                curr_peak = ResultItem { val_1, idx_1 };
+            else if (val_1 < val && val_1 < val_2)
+                curr_valley = ResultItem { val_1, idx_1 };
+        }
+
+        peaks_.swap(peaks);
+        valleys_.swap(valleys);
     }
 
     inline void pre ()  {
diff --git a/include/DataFrame/DataFrameStatsVisitors.h b/include/DataFrame/DataFrameStatsVisitors.h
index 6752b31c..6c69290c 100644
--- a/include/DataFrame/DataFrameStatsVisitors.h
+++ b/include/DataFrame/DataFrameStatsVisitors.h
@@ -58,7 +58,7 @@ SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #include 
 #include 
 
-// 2.0 *  PI
+// 2PI
 //
 #define TAU 6.28318530717958647692528676655900576
 
diff --git a/test/dataframe_tester_3.cc b/test/dataframe_tester_3.cc
index 64831f4b..2335e22c 100644
--- a/test/dataframe_tester_3.cc
+++ b/test/dataframe_tester_3.cc
@@ -3053,7 +3053,7 @@ static void test_get_data_by_rand_from_view()  {
     }
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_get_data_by_like_from_view()  {
 
@@ -3203,7 +3203,7 @@ static void test_get_reindexed_from_view()  {
     assert(result.get_column("dbl_col_2")[10] == 112.0);
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_concat_from_view()  {
 
@@ -3261,7 +3261,7 @@ static void test_concat_from_view()  {
     assert(result.get_column("int_col")[15] == 1);
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_get_view_by_sel_from_view()  {
 
@@ -3393,7 +3393,7 @@ static void test_get_view_by_rand_from_view()  {
     }
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_get_view_by_like_from_view()  {
 
@@ -3543,7 +3543,7 @@ static void test_get_reindexed_view_from_view()  {
     assert(result.get_column("dbl_col_2")[10] == 112.0);
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_concat_view_from_view()  {
 
@@ -3603,7 +3603,7 @@ static void test_concat_view_from_view()  {
     assert(result2.get_index()[14] == 14);
 }
 
-// -----------------------------------------------------------------------------
+// ----------------------------------------------------------------------------
 
 static void test_client_csv_read_test()  {
 
@@ -3632,6 +3632,68 @@ static void test_client_csv_read_test()  {
 
 // ----------------------------------------------------------------------------
 
+static void test_PeaksAndValleysVisitor()  {
+
+    std::cout << "\nTesting PeaksAndValleysVisitor{ } ..." << std::endl;
+
+    typedef StdDataFrame64 StrDataFrame;
+
+    StrDataFrame    df;
+
+    try  {
+        df.read("SHORT_IBM.csv", io_format::csv2);
+
+        pav_v  pav;
+
+        df.single_act_visit("IBM_Close", pav);
+
+        assert(pav.get_peaks().size() == 1721);
+        assert(std::abs(pav.get_peaks()[0].value - 185.53) < 0.0001);
+        assert(pav.get_peaks()[0].index == "2014-01-02");
+        assert(std::abs(pav.get_peaks()[1].value - 186.64) < 0.0001);
+        assert(pav.get_peaks()[1].index == "2014-01-03");
+        assert(std::abs(pav.get_peaks()[19].value - 182.73) < 0.0001);
+        assert(pav.get_peaks()[19].index == "2014-01-23");
+        assert(std::abs(pav.get_peaks()[20].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[20].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[24].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[24].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[25].value - 177.36) < 0.0001);
+        assert(pav.get_peaks()[25].index == "2014-01-30");
+        assert(std::abs(pav.get_peaks()[1720].value - 116.0) < 0.0001);
+        assert(pav.get_peaks()[1720].index == "2020-10-23");
+        assert(std::abs(pav.get_peaks()[1712].value - 125.93) < 0.0001);
+        assert(pav.get_peaks()[1712].index == "2020-10-16");
+        assert(std::abs(pav.get_peaks()[1707].value - 131.49) < 0.0001);
+        assert(pav.get_peaks()[1707].index == "2020-10-08");
+
+        assert(pav.get_valleys().size() == 1721);
+        assert(std::abs(pav.get_valleys()[0].value - 185.53) < 0.0001);
+        assert(pav.get_valleys()[0].index == "2014-01-02");
+        assert(std::abs(pav.get_valleys()[1].value - 186.64) < 0.0001);
+        assert(pav.get_valleys()[1].index == "2014-01-03");
+        assert(std::abs(pav.get_valleys()[19].value - 176.4) < 0.0001);
+        assert(pav.get_valleys()[19].index == "2014-01-29");
+        assert(std::abs(pav.get_valleys()[20].value - 176.4) < 0.0001);
+        assert(pav.get_valleys()[20].index == "2014-01-29");
+        assert(std::abs(pav.get_valleys()[24].value - 172.84) < 0.0001);
+        assert(pav.get_valleys()[24].index == "2014-02-04");
+        assert(std::abs(pav.get_valleys()[25].value - 172.84) < 0.0001);
+        assert(pav.get_valleys()[25].index == "2014-02-04");
+        assert(std::abs(pav.get_valleys()[1720].value - 106.65) < 0.0001);
+        assert(pav.get_valleys()[1720].index == "2020-10-28");
+        assert(std::abs(pav.get_valleys()[1712].value - 124.89) < 0.0001);
+        assert(pav.get_valleys()[1712].index == "2020-10-15");
+        assert(std::abs(pav.get_valleys()[1707].value - 121.97) < 0.0001);
+        assert(pav.get_valleys()[1707].index == "2020-10-06");
+    }
+    catch (const DataFrameError &ex)  {
+        std::cout << ex.what() << std::endl;
+    }
+}
+
+// ----------------------------------------------------------------------------
+
 int main(int, char *[]) {
 
     MyDataFrame::set_optimum_thread_level();
@@ -3708,6 +3770,7 @@ int main(int, char *[]) {
     test_get_reindexed_view_from_view();
     test_concat_view_from_view();
     test_client_csv_read_test();
+    test_PeaksAndValleysVisitor();
 
     return (0);
 }