Additionally, Zhu, for now, I think we should have our codes return two objects after every post-processing step: the data (pd.DataFrame) and the metadata (dict). In discussions with Joseph, it turns out that the Nodes are immutable, so we shouldn’t try to modify them. Additionally, we should keep in mind that users may want to save/access the data directly in the DataFrame or dictionary form. Once we have the steps in place that go from Node -> (pd.DataFrame, dict), we can write a simple function that repackages the data/metadata back into a node for pushing back to the database.
See Slack link here.