[API] Adding printing function to HeteroCL

[seanlatias]

Main Features

1. In this PR, we introduce a new API hcl.print, which can print HeteroCL objects. Currently, the functionality is still very basic. In other words, we cannot generate pretty print like NumPy.
2. I also removed the need for using hcl.Stage at the top-level. A default top-level Stage will be generated automatically. Please refer to test_dsl_basic.py for more examples.

Examples for hcl.print

1. Print an expression

A = hcl.compute((10,), lambda x: x)
hcl.print(A[0])
# output: 0
hcl.print(A[0], "A[0]: %d\n")
# output: A[0]: 0
hcl.print([A[0], A[1]])
# output: A[0] A[1]
hcl.print([A[0], A[1]], "0: %d, 1: %d\n")
# output: 0: 0, 1: 1
hcl.print(5)
# output: 5

More details: for integers, by default we use "%d". For all other data types including floating points and fixed points, we use "%f".

2. Print a Tensor or a slice of Tensor

A = hcl.compute((3, 3), lambda x, y: x+y)
hcl.print(A[0])
# output: [0, 1, 2]
hcl.print(A)
# output:
# [[0, 1, 2],
# [1, 2, 3],
# [2, 3, 4]]

As can be seen from the example above, currently we do not have pretty print with indentations and also find the largest common number of digits. This will be introduced later.

[seanlatias]

Solve issue #174.