Add DC verifier python bindings and basic example

examples/basic/verifying_dc.py

@@ -0,0 +1,20 @@
+import desbordante as db
+
+# # Given denial constraint tells us that if two persons
+# live in the same state, the one earning a lower
+# salary has a lower tax rate
+DC = "!(s.State == t.State and s.Salary < t.Salary and s.FedTaxRate > t.FedTaxRate)"
+
+TABLE = "test_input_data/TestDC1.csv"
+
+# Creating a verificator and loading data in algortihm
+verificator = db.dc_verification.algorithms.Default() 
+verificator.load_data(table=(TABLE, ',', True))
+
+# Algorithm execution
+verificator.execute(denial_constraint=DC)
+
+# Obtaining the result
+result: bool = verificator.dc_holds()   
+
+print("DC " + DC + " holds: " + str(result))

examples/expert/data_cleaning.py

@@ -0,0 +1,47 @@
+import desbordante as db
+import networkx as nx
+import matplotlib.pyplot as plt
+import time
+
+def verify_dc(verificator, dc):    
+    total_start = time.time()
+    print("Algo execution start")
+    iterations = 1
+
+    algo_start = time.time()
+    for _ in range(iterations):
+        verificator.execute(denial_constraint=dc)
+    algo_time = (time.time() - algo_start) / iterations
+
+    print("Algo execution end")
+
+    result = verificator.dc_holds()
+    total_time = time.time() - total_start
+
+    print("DC " + dc + " holds: " + str(result))
+    print("Algo average time elapsed: " + str(algo_time))
+    print("Total time elapsed: " + str(total_time))
+
+
+def main():
+    table = 'TestDC1.csv'
+
+    verificator = db.dc_verification.algorithms.Default()
+    verificator.load_data(table=(table, ',', True))
+
+    dc = "!(t.0 == s.0 and t.1 > s.1 and t.2 < s.2)"
+
+    verify_dc(verificator, dc)
+    viols = verificator.get_violations()
+
+    G = nx.Graph()
+    for edge in viols:
+        G.add_edge(*edge)
+    print(G)
+    nx.draw(G, with_labels=True)
+
+    plt.show()
+
+
+if __name__ == "__main__":
+    main()

src/core/algorithms/dc/model/point.h

@@ -39,7 +39,7 @@ class Point {
     }
 
     size_t GetIndex() const {
-        return index;
+        return index_;
     }
 
     Component& operator[](size_t i) {

src/core/algorithms/dc/verifier/dc_verifier.cpp

@@ -123,10 +123,9 @@ bool DCVerifier::VerifyMixed(const DC& dc) {
 
         ProcessMixed(s_predicates, s_tree, t_tree, mixed_dc, i, all_cols, res);
         ProcessMixed(t_predicates, t_tree, s_tree, mixed_dc, i, all_cols, res);
-        if (!res) return false;
     }
 
-    return true;
+    return res;
 }
 
 void DCVerifier::ProcessMixed(std::vector<Predicate> const& preds, tree& insert_tree,
@@ -174,9 +173,8 @@ bool DCVerifier::VerifyTwoTuples(const DC& dc) {
         auto [box, inv_box] = SearchRanges(dc, tuple);
         std::vector<point> search_res = hash[key].QuerySearch(box);
         std::vector<point> inv_search_res = hash[key].QuerySearch(inv_box);
-        if (!search_res.empty() or !inv_search_res.empty()) {
-            return false;
-        }
+
+        if (!search_res.empty() or !inv_search_res.empty()) return false;
 
         hash[key].Insert(MakePoint(tuple, ineq_cols, i + index_offset_));
     }
@@ -185,17 +183,20 @@ bool DCVerifier::VerifyTwoTuples(const DC& dc) {
 }
 
 bool DCVerifier::VerifyAllEquality(const DC& dc) {
+    bool res = true;
     std::unordered_set<size_t> res_tuples;
     std::vector<uint> const eq_cols = dc.GetColumnIndices();
     for (size_t i = 0; i < data_[0].GetNumRows(); i++) {
         if (ContainsNullOrEmpty(eq_cols, i)) continue;
         std::vector<std::byte const*> tuple = GetTuple(i);
         size_t key = HashTuple(tuple, eq_cols);
-        if (res_tuples.find(key) != res_tuples.end()) return false;
+        if (res_tuples.find(key) != res_tuples.end()) {
+            res = false;
+        }
         res_tuples.insert(key);
     }
 
-    return true;
+    return res;
 }
 
 bool DCVerifier::VerifyOneInequality(const DC& dc) {

src/python_bindings/bindings.cpp

@@ -8,6 +8,7 @@
 #include "bind_main_classes.h"
 #include "cfd/bind_cfd.h"
 #include "data/bind_data_types.h"
+#include "dc/bind_dc_verification.h"
 #include "dd/bind_split.h"
 #include "dynamic/bind_dynamic_fd_verification.h"
 #include "fd/bind_fd.h"
@@ -38,12 +39,15 @@ PYBIND11_MODULE(desbordante, module, pybind11::mod_gil_not_used()) {
         el::Loggers::reconfigureAllLoggers(conf);
     }
 
+    // clang-format off
     for (auto bind_func :
          {BindMainClasses, BindDataTypes, BindFd, BindCfd, BindAr, BindUcc, BindAc, BindOd, BindNd,
           BindFdVerification, BindMfdVerification, BindUccVerification, BindStatistics, BindInd,
-          BindGfdVerification, BindSplit, BindDynamicFdVerification, BindNdVerification, BindSFD}) {
+          BindGfdVerification, BindSplit, BindDynamicFdVerification, BindNdVerification, BindSFD,
+          BindDCVerification}) {
         bind_func(module);
     }
+    // clang-format on
 }
 
 }  // namespace python_bindings

src/python_bindings/dc/bind_dc_verification.cpp

@@ -0,0 +1,17 @@
+#include "bind_dc_verification.h"
+
+#include "algorithms/dc/verifier/dc_verifier.h"
+#include "py_util/bind_primitive.h"
+
+namespace python_bindings {
+
+namespace py = pybind11;
+
+void BindDCVerification(py::module_& main_module) {
+    auto dc_verification_module = main_module.def_submodule("dc_verification");
+
+    BindPrimitiveNoBase<algos::DCVerifier>(dc_verification_module, "DCVerification")
+            .def("dc_holds", &algos::DCVerifier::DCHolds);
+}
+
+}  // namespace python_bindings

src/python_bindings/dc/bind_dc_verification.h

@@ -0,0 +1,7 @@
+#pragma once
+
+#include <pybind11/pybind11.h>
+
+namespace python_bindings {
+void BindDCVerification(pybind11::module_& main_module);
+}  // namespace python_bindings

src/python_bindings/py_util/get_py_type.cpp

@@ -87,7 +87,7 @@ py::tuple GetPyType(std::type_index type_index) {
             PyTypePair<std::filesystem::path, kPyStr>,
             PyTypePair<std::vector<std::filesystem::path>, kPyList, kPyStr>,
             PyTypePair<std::unordered_set<size_t>, kPySet, kPyInt>,
-    };
+            PyTypePair<std::string, kPyStr>};
     return type_map.at(type_index)();
 }
 

src/python_bindings/py_util/py_to_any.cpp

@@ -130,7 +130,7 @@ std::unordered_map<std::type_index, ConvFunc> const kConverters{
         kNormalConvPair<std::filesystem::path>,
         kNormalConvPair<std::vector<std::filesystem::path>>,
         kNormalConvPair<std::unordered_set<size_t>>,
-};
+        kNormalConvPair<std::string>};
 
 }  // namespace
 

src/tests/all_csv_configs.cpp

@@ -123,4 +123,6 @@ CSVConfig const kLineItem = CreateCsvConfig("LineItem.csv", '|', true);
 CSVConfig const kTmpDC = CreateCsvConfig("tmp_dc.csv", ',', true);
 CSVConfig const kTestDC = CreateCsvConfig("TestDC.csv", ',', true);
 CSVConfig const kTestDC1 = CreateCsvConfig("TestDC1.csv", ',', true);
+CSVConfig const kTestDC2 = CreateCsvConfig("TestDC2.csv", ',', true);
+CSVConfig const kTestDC3 = CreateCsvConfig("TestDC3.csv", ',', true);
 }  // namespace tests

src/tests/all_csv_configs.h

@@ -103,4 +103,6 @@ extern CSVConfig const kLineItem;
 extern CSVConfig const kTmpDC;
 extern CSVConfig const kTestDC;
 extern CSVConfig const kTestDC1;
+extern CSVConfig const kTestDC2;
+extern CSVConfig const kTestDC3;
 }  // namespace tests

src/tests/test_dc_verifier.cpp

@@ -26,6 +26,10 @@ TEST(TestDCVerifier, TestUCC) {
     auto params = GetParamMap(kTestDC, dc_string);
     std::unique_ptr<DCVerifier> dc_verifier = algos::CreateAndLoadAlgorithm<DCVerifier>(params);
     dc_verifier->Execute();
+
+    size_t expected_size = 0;
+    std::vector<Point> res = dc_verifier->GetViolations();
+    EXPECT_EQ(expected_size, res.size());
     EXPECT_TRUE(dc_verifier->DCHolds());
 }
 
@@ -62,7 +66,6 @@ TEST(TestDCVerifier, TestMixedNamesColNums) {
 }
 
 TEST(TestDCVerifier, TestNoHeader) {
-    CSVConfig const kTestAR{"TestAR.csv", ',', false};
     std::string dc_string = "!(s.0 == t.1 and s.1 == t.2 and s.2 == t.3)";
     auto params = GetParamMap(kBernoulliRelation, dc_string);
     std::unique_ptr<DCVerifier> dc_verifier = algos::CreateAndLoadAlgorithm<DCVerifier>(params);

test_input_data/TestDC2.csv

@@ -0,0 +1,12 @@
+State,Salary,FedTaxRate
+NewYork,3000,0.2
+NewYork,4000,0.25
+NewYork,5000,0.3
+Wisconsin,5000,0.15
+Wisconsin,6000,0.2
+Wisconsin,4000,0.1
+Wisconsin,4500,0.12
+Texas,5000,0.05
+Texas,1000,0.15
+Texas,2000,0.25
+Texas,3000,0.3

test_input_data/TestDC3.csv

@@ -0,0 +1,14 @@
+NewYork,3000,0.2
+NewYork,4000,0.25
+NewYork,5000,0.3
+NewYork,10,10
+Wisconsin,5000,0.15
+Wisconsin,6000,0.2
+Wisconsin,20,20
+Wisconsin,4000,0.1
+Texas,5000,0.05
+Texas,1000,0.15
+Texas,2000,0.25
+Texas,3000,0.3
+Texas,3000,0.31
+Texas,30,30