mingtan888

https://github.com/salesforce/TabularSemanticParsing/blob/0b094d329a5f0b5c038a820117ba4b6e7b6215cc/src/data_processor/schema_loader.py#L186 `schema_graph.compute_field_picklist` is executed before `load_data_from_spider_json`, when `self.field_rev_index` is still empty. Not sure if this is intentional, or happens to be a bug?