omnata-plugin-runtime 0.9.1a210__tar.gz → 0.9.1a211__tar.gz

{omnata_plugin_runtime-0.9.1a210 → omnata_plugin_runtime-0.9.1a211}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: omnata-plugin-runtime
-Version: 0.9.1a210
+Version: 0.9.1a211
 Summary: Classes and common runtime components for building and running Omnata Plugins
 Author: James Weakley
 Author-email: james.weakley@omnata.com

{omnata_plugin_runtime-0.9.1a210 → omnata_plugin_runtime-0.9.1a211}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "omnata-plugin-runtime"
-version = "0.9.1-a210"
+version = "0.9.1-a211"
 description = "Classes and common runtime components for building and running Omnata Plugins"
 authors = ["James Weakley <james.weakley@omnata.com>"]
 readme = "README.md"

{omnata_plugin_runtime-0.9.1a210 → omnata_plugin_runtime-0.9.1a211}/src/omnata_plugin_runtime/json_schema.py

@@ -207,13 +207,35 @@ class SnowflakeViewColumn(BaseModel):
             expression=f"""TO_{timestamp_type}({expression}::varchar,'{timestamp_format}')"""
         else:
             if not json_schema_property.snowflakeColumnExpression:
-                expression=f"""{expression}::{json_schema_property.type}"""
+                expression=f"""{expression}::{json_schema_property.snowflake_data_type}"""
         return cls(
             name=final_column_name,
             expression=expression,
             comment=comment,
             is_join_column=json_schema_property.isJoinColumn,
         )
+    
+    @classmethod
+    def order_by_reference(cls,join_columns:List[Self]) -> List[Self]:
+        """
+        In some situations, column expressions may reference the alias of another column
+        This is allowed in Snowflake, as long as the aliased column is defined before it's used in a later column
+        So we need to sort the columns so that if the name of the column appears (in quotes) in the expression of another column, it is ordered first
+        """
+        
+        # Collect columns to be moved
+        columns_to_move:List[Self] = []
+        for column in join_columns:
+            for other_column in join_columns:
+                if f'"{column.name}"' in other_column.expression:
+                    if column not in columns_to_move:
+                        columns_to_move.append(column)
+
+        # Move collected columns to the front
+        for column in columns_to_move:
+            join_columns.remove(column)
+            join_columns.insert(0, column)
+        return join_columns
 
 
 class SnowflakeViewJoin(BaseModel):
@@ -259,11 +281,41 @@ class SnowflakeViewJoin(BaseModel):
 ON "{self.left_alias}"."{self.left_column}" = "{self.join_stream_alias}"."{self.join_stream_column}" """
 
 
-class SnowflakeViewParts(BaseModel):
+class FullyQualifiedTable(BaseModel):
     """
-    Represents the definition of a Snowflake normalized view.
+    Represents a fully qualified table name in Snowflake, including database, schema, and table name.
+    This is not a template, it's a fully specified object.
     """
 
+    database_name: Optional[str] = Field(default=None, description="The database name")
+    schema_name: str = Field(..., description="The schema name")
+    table_name: str = Field(..., description="The table name")
+
+    def get_fully_qualified_name(self, table_override: Optional[str] = None) -> str:
+        """
+        If table_override is provided, it will be used instead of the table name
+        """
+        actual_table_name = (
+            self.table_name if table_override is None else table_override
+        )
+        # We try to make this resilient to quoting
+        schema_name = self.schema_name.replace('"', "")
+        table_name = actual_table_name.replace('"', "")
+        if self.database_name is None or self.database_name == "":
+            return f'"{schema_name}"."{table_name}"'
+        database_name = self.database_name.replace('"', "")
+        return f'"{database_name}"."{schema_name}"."{table_name}"'
+
+
+class SnowflakeViewPart(BaseModel):
+    """
+    Represents a stream within a normalized view.
+    Because a normalized view can be built from multiple streams, this is potentially only part of the view.
+    """
+    stream_name: str = Field(..., description="The name of the stream")
+    raw_table_location: FullyQualifiedTable = Field(
+        ..., description="The location of the raw table that the stream is sourced from"
+    )
     comment: Optional[str] = Field(
         None, description="The comment to assign to the view"
     )
@@ -284,7 +336,7 @@ class SnowflakeViewParts(BaseModel):
         """
         Returns the columns that are sourced from joins.
         """
-        return [c for c in self.columns if c.is_join_column]
+        return SnowflakeViewColumn.order_by_reference([c for c in self.columns if c.is_join_column])
 
     def comment_clause(self) -> str:
         """
@@ -298,31 +350,86 @@ class SnowflakeViewParts(BaseModel):
         return [
             c.name_with_comment() for c in (self.direct_columns() + self.join_columns())
         ]
+    
+    def cte_text(self) -> str:
+        """
+        Returns the CTE text for this view part.
+        """
+        return f""" "{self.stream_name}" as (
+    select {', '.join([c.definition() for c in self.direct_columns()])} 
+    from {self.raw_table_location.get_fully_qualified_name()}
+) """
 
-class FullyQualifiedTable(BaseModel):
+class SnowflakeViewParts(BaseModel):
     """
-    Represents a fully qualified table name in Snowflake, including database, schema, and table name.
-    This is not a template, it's a fully specified object.
+    Represents a set of streams within a normalized view.
+    This is the top level object that represents the whole view.
     """
 
-    database_name: Optional[str] = Field(default=None, description="The database name")
-    schema_name: str = Field(..., description="The schema name")
-    table_name: str = Field(..., description="The table name")
+    main_part: SnowflakeViewPart = Field(
+        ..., description="The main part of the view, which is the stream that the view is named after"
+    )
+    joined_parts: List[SnowflakeViewPart] = Field(
+        ..., description="The other streams that are joined to the main stream"
+    )
 
-    def get_fully_qualified_name(self, table_override: Optional[str] = None) -> str:
+    def view_body(self):
         """
-        If table_override is provided, it will be used instead of the table name
+        Creates a view definition from the parts
         """
-        actual_table_name = (
-            self.table_name if table_override is None else table_override
+        ctes = [self.main_part.cte_text()] + [part.cte_text() for part in self.joined_parts]
+        all_ctes = "\n,".join(ctes)
+        join_columns = self.main_part.join_columns()
+        join_column_clauses = [c.definition() for c in join_columns]
+        # we select * from the original view (in the CTE) and then add any expressions that come from the join columns
+        final_column_clauses = [f'"{self.main_part.stream_name}".*'] + join_column_clauses
+        view_body = f"""with {all_ctes}
+    select {', '.join(final_column_clauses)}
+    from "{self.main_part.stream_name}" """
+        if len(self.main_part.joins) > 0:
+            join_clauses = [join.definition() for join in self.main_part.joins]
+            view_body += "\n" + ("\n".join(join_clauses))
+        return view_body
+
+    @classmethod
+    def generate(cls,
+        raw_stream_locations: Dict[str,FullyQualifiedTable],
+        stream_schemas: Dict[str,Dict],
+        stream_name: str,
+        include_default_columns: bool = True,
+        column_name_environment: Environment = Environment(),
+        column_name_expression: str = "{{column_name}}"
+    ) -> Self:
+        """
+        Returns the building blocks required to create a normalized view from a stream.
+        This includes any joins that are required, via CTEs.
+        """
+        # we start with the view parts for the view we are building
+        main_stream_view_part = normalized_view_part(
+            stream_name=stream_name,
+            raw_table_location=raw_stream_locations[stream_name],
+            include_default_columns=include_default_columns,
+            stream_schema=stream_schemas.get(stream_name),
+            column_name_environment=column_name_environment,
+            column_name_expression=column_name_expression
         )
-        # We try to make this resilient to quoting
-        schema_name = self.schema_name.replace('"', "")
-        table_name = actual_table_name.replace('"', "")
-        if self.database_name is None or self.database_name == "":
-            return f'"{schema_name}"."{table_name}"'
-        database_name = self.database_name.replace('"', "")
-        return f'"{database_name}"."{schema_name}"."{table_name}"'
+        joined_parts = []
+        for join in main_stream_view_part.joins:
+            if join.join_stream_name not in raw_stream_locations:
+                raise ValueError(f"Stream {join.join_stream_name} is required as a join for stream {stream_name}, but its location was not provided")
+            if join.join_stream_name not in stream_schemas:
+                raise ValueError(f"Stream {join.join_stream_name} is required as a join for stream {stream_name}, but its schema was not provided")
+            joined_parts.append(normalized_view_part(
+                stream_name=join.join_stream_name,
+                raw_table_location=raw_stream_locations[join.join_stream_name],
+                include_default_columns=include_default_columns,
+                stream_schema=stream_schemas[join.join_stream_name],
+                column_name_environment=column_name_environment,
+                column_name_expression=column_name_expression
+            ))
+        return cls(main_part=main_stream_view_part, joined_parts=joined_parts)
+
+    
 
 class JsonSchemaTopLevel(BaseModel):
     """
@@ -341,9 +448,9 @@ class JsonSchemaTopLevel(BaseModel):
     )
 
     def build_view_columns(self,
-                            column_name_environment: Environment = Environment(),                
-                            column_name_expression: str = "{{column_name}}"
-                           ) -> List[SnowflakeViewColumn]:
+            column_name_environment: Environment,                
+            column_name_expression: str
+        ) -> List[SnowflakeViewColumn]:
         """
         Returns a list of column definitions from a json schema
         """
@@ -413,10 +520,14 @@ class JsonSchemaTopLevel(BaseModel):
         )]
     
 
-def normalized_view_parts(
+def normalized_view_part(
+    stream_name:str,
+    raw_table_location:FullyQualifiedTable,
     include_default_columns: bool,
+    column_name_environment: Environment,                
+    column_name_expression: str,
     stream_schema: Optional[Dict] = None,
-) -> SnowflakeViewParts:
+) -> SnowflakeViewPart:
     """
     Returns an object containing:
     - A top level comment for the view
@@ -461,78 +572,14 @@ def normalized_view_parts(
             )
         )
     json_schema = JsonSchemaTopLevel.model_validate(stream_schema)
-    return SnowflakeViewParts(
-        columns=snowflake_columns + json_schema.build_view_columns(), 
+
+    return SnowflakeViewPart(
+        stream_name=stream_name,
+        raw_table_location=raw_table_location,
+        columns=snowflake_columns + json_schema.build_view_columns(
+            column_name_environment=column_name_environment,
+            column_name_expression=column_name_expression
+        ), 
         joins=json_schema.joins or [],
         comment=json_schema.description
     )
-
-def normalized_view_body(
-    stream_locations: Dict[str,FullyQualifiedTable],
-    stream_schemas: Dict[str,Dict],
-    stream_name: str,
-    include_default_columns: bool = True,
-) -> str:
-    """
-    Returns the SQL for the body of a normalized view.
-    Because views are created over raw data (potentially several joined raw tables), we have
-    to pass in the locations of those raw tables, keyed by stream name.
-    The stream schema is also passed in, keyed by stream name, and used to build the columns and joins.
-    """
-    main_stream_raw_table_name_quoted = stream_locations[stream_name].get_fully_qualified_name()
-    # we start with the view parts for the view we are building
-    main_stream_view_part = normalized_view_parts(
-        include_default_columns=include_default_columns,
-        stream_schema=stream_schemas.get(stream_name)
-    )
-    # we use a CTE because we may need to use aliases in the joins
-    main_stream_cte = f""" "{stream_name}" as (
-    select {', '.join([c.definition() for c in main_stream_view_part.direct_columns()])} 
-    from {main_stream_raw_table_name_quoted}
-) """
-    ctes = [main_stream_cte]
-    # we also use CTEs that recreate the views that the joins reference.
-    # the reason for this is that we can't rely on the view being there,
-    # and it's also possible that they reference each other
-    for join in main_stream_view_part.joins:
-        join_view_part = normalized_view_parts(
-            include_default_columns=include_default_columns,
-            stream_schema=stream_schemas.get(join.join_stream_name)
-        )
-        join_stream_raw_table_name_quoted = stream_locations[stream_name].get_fully_qualified_name()
-        join_view_cte = f""" "{join.join_stream_name}" as (
-        select {', '.join([c.definition() for c in join_view_part.direct_columns()])}
-        from {join_stream_raw_table_name_quoted}
-        ) """
-        ctes.append(join_view_cte)
-
-    join_columns = main_stream_view_part.join_columns()
-    # in some situations, column expressions may reference the alias of another column
-    # this is allowed in Snowflake, as long as the aliased column is defined before it's used in a later column
-    # so we need to sort the columns so that if the name of the column appears (in quotes) in the expression of another column, it is ordered first
-
-    # Collect columns to be moved
-    columns_to_move = []
-
-    for column in join_columns:
-        for other_column in join_columns:
-            if f'"{column.name}"' in other_column.expression:
-                if column not in columns_to_move:
-                    columns_to_move.append(column)
-
-    # Move collected columns to the front
-    for column in columns_to_move:
-        join_columns.remove(column)
-        join_columns.insert(0, column)
-
-    join_column_clauses = [c.definition() for c in join_columns]
-    # we select * from the original view (in the CTE) and then add any expressions that come from the join columns
-    final_column_clauses = [f'"{stream_name}".*'] + join_column_clauses
-    all_ctes = "\n,".join(ctes)
-    view_body = f"""with {all_ctes}
-select {', '.join(final_column_clauses)} from "{stream_name}" """
-
-    if len(main_stream_view_part.joins) > 0:
-        join_clauses = [join.definition() for join in main_stream_view_part.joins]
-        view_body += "\n" + ("\n".join(join_clauses))
-    return view_body