[WIP] Improve Execution Serialization

[pedroigorjs]

This pull request introduces enhancements to execution tracking and output normalization in the retrack engine. The main changes add support for node aliases, improve serialization of execution data, and introduce a utility for DataFrame normalization. These updates make execution outputs more readable and structured, and provide better context through node aliases.

Execution Tracking and Serialization Improvements:

• Added an aliases attribute to the Execution class, along with methods to set and store node aliases during execution. This allows each node's output to be associated with a human-readable alias.
• Implemented the to_normalized_dict method in Execution, which serializes inputs, outputs, child executions, and results into a normalized dictionary format for easier downstream processing and inspection.

Node Alias Support:

• Added an alias() method to the BaseNode class, which returns a node's alias if available, falling back to its name. This ensures consistent alias retrieval for execution tracking.

Utility Functions:

• Added the to_normalized_dict function in retrack/utils/transformers.py to convert DataFrame columns into a list of dictionaries with names and values, supporting the new serialization logic.

New Normalized JSON Example:

{
  "inputs": [
    {
      "name": "a",
      "values": [
        "1"
      ]
    },
    {
      "name": "b",
      "values": [
        "2"
      ]
    },
    {
      "name": "context",
      "values": [
        null
      ]
    },
    {
      "name": "c",
      "values": [
        "3"
      ]
    },
    {
      "name": "d",
      "values": [
        "4"
      ]
    }
  ],
  "outputs": [
    {
      "node_id": "c1ec8f0109c209bb",
      "alias": "Input",
      "values": [
        "1"
      ]
    },
    {
      "node_id": "fcc373ce66014558",
      "alias": "Input",
      "values": [
        "2"
      ]
    },
    {
      "node_id": "1b4b55ed6d5c9d46",
      "alias": "Math",
      "values": [
        3.0
      ]
    }
  ],
  "executions": [],
  "results": [
    {
      "name": "output",
      "values": [
        3.0
      ]
    },
    {
      "name": "message",
      "values": [
        null
      ]
    }
  ]
}