Add an `atomic_vector` wrapper for `vector` where `push_back` is thread-safe

pyfastani/_atomic_vector.hpp

+#ifndef _SAFEVEC_HPP
+#define _SAFEVEC_HPP
+
+#include <mutex>
+#include <vector>
+
+
+template <class T>
+class atomic_vector: public std::vector<T> {
+protected:
+    std::mutex mutex;
+public:
+    atomic_vector(): std::vector<T>(), mutex() {}
+    void push_back(const T& val) {
+        mutex.lock();
+        std::vector<T>::push_back(val);
+        mutex.unlock();
+    }
+};
+
+#endif

pyfastani/_atomic_vector.pxd

+from libcpp.vector cimport vector
+
+cdef extern from "_atomic_vector.hpp" nogil:
+    cdef cppclass atomic_vector[T](vector[T]):
+        pass