I am unable to call the source and target attributes in the tuple in the module penman.code during AMR parsing

Question

I used the code modified by newbing when using AMR parsing to generate the spangraph, as follows:

# Define the AMR graph as a string
amr = """# ::id lpp_1943.2 ::date 2012-06-07T17:06:20 ::annotator ISI-AMR-05 ::preferred
# ::snt Once when I was six years old I saw a magnificent picture in a book , called True Stories from Nature , about the primeval forest .
# ::save-date Mon May 13, 2013 ::file lpp_1943_2.txt
(s / see-01
      :ARG0 (i / i)
      :ARG1 (p / picture
            :mod (m / magnificent)
            :location (b2 / book :wiki -
                  :name (n / name :op1 "True" :op2 "Stories" :op3 "from" :op4 "Nature")
                  :topic (f / forest
                        :mod (p2 / primeval))))
      :mod (o / once)
      :time (a / age-01
            :ARG1 i
            :ARG2 (t / temporal-quantity :quant 6
                  :unit (y / year))))"""

# Import the penman library to parse the AMR graph
import penman

# Parse the AMR graph into a triple representation
triples = penman.decode(amr).triples

# Define the sentence as a list of tokens
sentence = ["Once", "when", "I", "was", "six", "years", "old", "I", "saw", "a", "magnificent", "picture", "in", "a", "book", ",", "called", "True", "Stories", "from", "Nature", ",", "about", "the", "primeval", "forest", "."]

# Define a function to find the span of a concept in the sentence
def find_span(concept):
    # If the concept is a string, return the index of the first occurrence of the concept in the sentence
    if isinstance(concept, str):
        return sentence.index(concept)
    # If the concept is a tuple, return the index of the first occurrence of the first element of the concept in the sentence
    elif isinstance(concept, tuple):
        return sentence.index(concept[0])
    # Otherwise, return -1 to indicate no span found
    else:
        return -1

# Define a function to collapse a subgraph into a single label
def collapse_subgraph(subgraph):
    # If the subgraph is empty, return None
    if not subgraph:
        return None
    # If the subgraph has only one triple, return the concept of the source node
    elif len(subgraph) == 1:
        return subgraph[0].source
    # If the subgraph has more than one triple, check if it is a name subgraph
    elif all(triple.relation.startswith(":op") for triple in subgraph):
        # Return a tuple of the name parts in order
        return tuple(triple.target for triple in sorted(subgraph, key=lambda x: x.relation))
    # Otherwise, return None to indicate no collapse possible
    else:
        return None
# Initialize a node counter to assign indices to nodes
node_counter = 0
# Loop through each triple in the AMR graph
for triple in triples:
    # Get the source and target concepts of the triple
    source = triple.source
    target = triple.target

    # Check if the source concept is already mapped to a node
    if source not in concept_to_node:
        # Find the span of the source concept in the sentence
        span = find_span(source)
        # If no span is found, skip this triple and continue with the next one
        if span == -1:
            continue

        # Check if there is already a node for this span
        if span in span_to_node:
            # Get the existing node for this span
            node = span_to_node[span]
        else:
            # Create a new node for this span
            node = "s" + str(node_counter)
            # Increment the node counter
            node_counter += 1
            # Add the node to the nodes list
            nodes.append(node)
            # Add the span to the node to span mapping
            node_to_span[node] = span
            # Add the node to the span to node mapping
            span_to_node[span] = node

        # Add the source concept to the node to concept mapping
        node_to_concept[node] = source
        # Add the node to the concept to node mapping
        concept_to_node[source] = node

I am unable to call the source and target attributes in the tuple. This code attempts to build a span diagram from AMR, but cannot run successfully because of this error:

D:\Anaconda3\python.exe D:/360MoveData/Users/Dell/Desktop/webhomework/NLP/AMRPASING/decode_from_txt_to_graph.py
Traceback (most recent call last):
  File "D:\360MoveData\Users\Dell\Desktop\webhomework\NLP\AMRPASING\decode_from_txt_to_graph.py", line 80, in <module>
    source = triple.source
AttributeError: 'tuple' object has no attribute 'source'

I tried to check the program source code and asked newbing, ChatGPT, etc., but found nothing