Build against datasketches-cpp 5.3.0

C++ 5.3.0 fixes five deserialization vulnerabilities (CVE-2026-103501,
-103513, -103634, -103635, -103636), which this package inherits.

Two C++ member functions gained a defaulted parameter, which breaks the
bindings that take their address:
- update_theta_sketch::compact(ordered, trim): expose trim=False. Without
  a second nb::arg the build fails.
- hll_sketch::reset(full_size): expose full_size=False. Without an
  nb::arg, reset() required an argument and test_hll_* failed.

Also bind the new frequent_items_sketch::reset(). Adds tests for all three.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Author: leerho

CMakeLists.txt

@@ -117,7 +117,7 @@ cmake_policy(SET CMP0097 NEW)
 include(ExternalProject)
 ExternalProject_Add(datasketches
   GIT_REPOSITORY https://github.com/apache/datasketches-cpp.git
-  GIT_TAG 5.2.0
+  GIT_TAG 5.3.0
   GIT_SHALLOW true
   GIT_SUBMODULES ""
   INSTALL_DIR /tmp/datasketches

src/fi_wrapper.cpp

@@ -62,6 +62,8 @@ void bind_fi_sketch(nb::module_ &m, const char* name) {
          "Merges the given sketch into this one")
     .def("is_empty", &frequent_items_sketch<T, W, H, E>::is_empty,
          "Returns True if the sketch is empty, otherwise False")
+    .def("reset", &frequent_items_sketch<T, W, H, E>::reset,
+         "Resets the sketch to the initial empty state")
     .def_prop_ro("num_active_items", &frequent_items_sketch<T, W, H, E>::get_num_active_items,
          "The number of active items in the sketch")
     .def_prop_ro("total_weight", &frequent_items_sketch<T, W, H, E>::get_total_weight,

src/hll_wrapper.cpp

@@ -65,8 +65,8 @@ void init_hll(nb::module_ &m) {
          "Returns the size of the serialized sketch")
     .def("get_compact_serialization_bytes", &hll_sketch::get_compact_serialization_bytes,
          "Returns the size of the serialized sketch when compressing the exception table if HLL_4")
-    .def("reset", &hll_sketch::reset,
-         "Resets the sketch to the empty state in coupon collection mode")
+    .def("reset", &hll_sketch::reset, nb::arg("full_size")=false,
+         "Resets the sketch to the empty state, in coupon collection mode or, if full_size is True, as a full-size HLL array")
     .def("update", (void (hll_sketch::*)(int64_t)) &hll_sketch::update, nb::arg("datum"),
          "Updates the sketch with the given integral value")
     .def("update", (void (hll_sketch::*)(double)) &hll_sketch::update, nb::arg("datum"),

src/theta_wrapper.cpp

@@ -88,8 +88,8 @@ void init_theta(nb::module_ &m) {
          "Updates the sketch with the given floating point value")
     .def("update", (void (update_theta_sketch::*)(const std::string&)) &update_theta_sketch::update, nb::arg("datum"),
          "Updates the sketch with the given string")
-    .def("compact", &update_theta_sketch::compact, nb::arg("ordered")=true,
-         "Returns a compacted form of the sketch, optionally sorting it")
+    .def("compact", &update_theta_sketch::compact, nb::arg("ordered")=true, nb::arg("trim")=false,
+         "Returns a compacted form of the sketch, optionally sorting it and trimming it to at most the nominal size k")
     .def("trim", &update_theta_sketch::trim, "Removes retained entries in excess of the nominal size k (if any)")
     .def("reset", &update_theta_sketch::reset, "Resets the sketch to the initial empty state")
   ;

tests/fi_test.py

@@ -148,5 +148,12 @@ def test_fi_sketch(self):
     reference_apriori_error = frequent_strings_sketch.get_apriori_error(k, wt)
     self.assertAlmostEqual(sk_apriori_error, reference_apriori_error, delta=1e-6)
 
+    fi.update("a", 5)
+    self.assertFalse(fi.is_empty())
+    fi.reset()
+    self.assertTrue(fi.is_empty())
+    self.assertEqual(fi.total_weight, 0)
+    self.assertEqual(fi.num_active_items, 0)
+
 if __name__ == '__main__':
   unittest.main()

tests/hll_test.py

@@ -101,6 +101,12 @@ def test_hll_sketch(self):
         hll.reset()
         self.assertTrue(hll.is_empty())
 
+        # reset can also return to an empty full-size HLL array
+        hll.reset(full_size=True)
+        self.assertTrue(hll.is_empty())
+        full = hll_sketch(lgk, tgt_hll_type.HLL_6, True)
+        self.assertEqual(hll.get_updatable_serialization_bytes(), full.get_updatable_serialization_bytes())
+
     def test_hll_union(self):
         lgk = 7
         n = 53

tests/theta_test.py

@@ -68,6 +68,11 @@ def test_theta_basic_example(self):
           count = count + 1
         self.assertEqual(count, new_sk.num_retained)
 
+        # compact can also trim the result to at most k entries
+        trimmed = sk.compact(trim=True)
+        self.assertLessEqual(trimmed.num_retained, 1 << lgk)
+        self.assertLessEqual(trimmed.num_retained, sk.num_retained)
+
         num = sk.num_retained
         sk.trim()
         self.assertLessEqual(sk.num_retained, num)