Add support for namespaced attributes in DynamicAttributes

Add support for namespaced attributes in DynamicAttribute with a recursive approach (a DynamicAttribute can have an attribute of type DynamicAttribute). Usage example (both syntaxes are correct):

attrs.get_attr("mem.data.b")
attrs.get_attr("mem").get_attr("data").get_attr("b")

This is a preliminary work in order to introduce a general mechanism for specifiyng backend and export attributes.

• Needs more tests in Python.
• Use snake convention for all attributes naming in C++ and Python.
• Solve #142 (closed)

include/aidge/data/Tensor.hpp

@@ -103,6 +103,22 @@ class Tensor : public Data,
         resize(dims);
     }
 
+    /**
+     * @brief Construct a new Tensor object from the 1-dimension Vector helper.
+     * @tparam T datatype
+     */
+    template <typename T>
+    constexpr Tensor(Vector<T> &&arr)
+        : Data(Type),
+          mDataType(NativeType<T>::type),
+          mDims({arr.data.size()}),
+          mStrides({1}),
+          mImpl(Registrar<Tensor>::create({"cpu", NativeType<T>::type})(0, {arr.data.size()})),
+          mSize(arr.data.size())
+    {
+        mImpl->copyFromHost(&arr.data[0], arr.data.size());
+    }
+
     /**
      * @brief Construct a new Tensor object from the 1-dimension Array helper.
      * @tparam T datatype
@@ -203,6 +219,12 @@ class Tensor : public Data,
      */
     Tensor &operator=(const Tensor& other);
 
+    template <typename T>
+    constexpr Tensor &operator=(Vector<T> &&arr) {
+        *this = Tensor(std::move(arr));
+        return *this;
+    }
+
     template <typename T, std::size_t SIZE_0>
     constexpr Tensor &operator=(Array1D<T, SIZE_0> &&arr) {
         *this = Tensor(std::move(arr));