Fix #1104: Improve lakeinfo structure and reporting

pdr_backend/cli/cli_module_lake.py

@@ -7,13 +7,11 @@
 from enforce_typing import enforce_types
 
 from pdr_backend.cli.cli_arguments_lake import LAKE_SUBCOMMANDS, LakeArgParser
-
+from pdr_backend.lake.duckdb_data_store import DuckDBDataStore
 from pdr_backend.lake.etl import ETL
 from pdr_backend.lake.gql_data_factory import GQLDataFactory
-from pdr_backend.lake.lake_info import LakeInfo
-from pdr_backend.lake.lake_validate import LakeValidate
-from pdr_backend.lake.duckdb_data_store import DuckDBDataStore
 from pdr_backend.lake.table import drop_tables_from_st
+from pdr_backend.lake_info.lake_info import LakeInfo
 from pdr_backend.ppss.ppss import PPSS
 from pdr_backend.util.time_types import UnixTimeMs
 
@@ -59,8 +57,8 @@ def do_lake_describe(args, ppss):
 
 @enforce_types
 def do_lake_validate(_, ppss):
-    lake_validate = LakeValidate(ppss)
-    lake_validate.run()
+    lake_info = LakeInfo(ppss)
+    lake_info.run_validation()
 
 
 @enforce_types

pdr_backend/cli/test/test_cli_module_lake.py

@@ -63,7 +63,7 @@ def test_do_lake_validate():
 
     ppss = Mock()
 
-    with patch("pdr_backend.cli.cli_module_lake.LakeValidate") as mock_lake_info:
+    with patch("pdr_backend.cli.cli_module_lake.LakeInfo") as mock_lake_info:
         do_lake_validate(args, ppss)
 
     mock_lake_info.assert_called_once_with(ppss)

pdr_backend/lake/renderers/cli.py → pdr_backend/lake_info/cli.py

@@ -7,12 +7,13 @@
 import polars as pl
 from polars.dataframe.frame import DataFrame
 
+from pdr_backend.lake_info.overview import TableViewsOverview
 from pdr_backend.util.time_types import UnixTimeMs
 
 
 class CliRenderer:
-    def __init__(self, lake_info):
-        self.lake_info = lake_info
+    def __init__(self, overview: TableViewsOverview):
+        self.overview = overview
 
     def print_table_info(self, source_name: str, source: Dict[str, DataFrame]):
         table_summary = DataFrame()
@@ -87,17 +88,17 @@ def print_table_info(self, source_name: str, source: Dict[str, DataFrame]):
             print(table_summary)
 
     def show(self):
-        if len(self.lake_info.all_table_names) == 0:
+        if len(self.overview.all_table_names) == 0:
             print("No Lake Tables")
         else:
             print("Lake Tables:")
-            print(self.lake_info.all_table_names)
-            self.print_table_info("tables", self.lake_info.table_info)
+            print(self.overview.all_table_names)
+            self.print_table_info("tables", self.overview.table_info)
 
-        if len(self.lake_info.all_view_names) == 0:
+        if len(self.overview.all_view_names) == 0:
             print("Great - No views in Lake.")
         else:
             print("=" * 80)
             print("Lake Views:")
-            print(self.lake_info.all_view_names)
+            print(self.overview.all_view_names)
             self.print_table_info("views", self.lake_info.view_info)