sunholo 0.124.0__tar.gz → 0.125.1__tar.gz

{sunholo-0.124.0/src/sunholo.egg-info → sunholo-0.125.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sunholo
-Version: 0.124.0
+Version: 0.125.1
 Summary: Large Language Model DevOps - a package to help deploy LLMs to the Cloud.
 Author-email: Holosun ApS <multivac@sunholo.com>
 License: Apache License, Version 2.0

{sunholo-0.124.0 → sunholo-0.125.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sunholo"
-version = "0.124.0"
+version = "0.125.1"
 description = "Large Language Model DevOps - a package to help deploy LLMs to the Cloud."
 readme = "README.md"
 requires-python = ">=3.10"

{sunholo-0.124.0 → sunholo-0.125.1}/src/sunholo/database/alloydb_client.py

@@ -584,7 +584,8 @@ class AlloyDBClient:
 
     async def create_table_from_schema(self, table_name: str, schema_data: dict, users: list = None):
         """
-        Creates or ensures a table exists based on the structure of the provided schema data.
+        Creates or ensures a table exists based on the structure of the provided schema data,
+        with special handling for expandable lists.
         
         Args:
             table_name (str): Name of the table to create
@@ -596,22 +597,32 @@ class AlloyDBClient:
         """
         # Generate column definitions from schema data
         columns = []
+        
         for key, value in schema_data.items():
-            if isinstance(value, dict):
-                # For nested objects, store as JSONB
-                columns.append(f'"{key}" JSONB')
-            elif isinstance(value, list):
-                # For arrays, store as JSONB
-                columns.append(f'"{key}" JSONB')
-            elif isinstance(value, int):
-                columns.append(f'"{key}" INTEGER')
-            elif isinstance(value, float):
-                columns.append(f'"{key}" NUMERIC')
-            elif isinstance(value, bool):
-                columns.append(f'"{key}" BOOLEAN')
+            # Check if this is a specially marked expandable list
+            if isinstance(value, dict) and value.get("__is_expandable_list__", False):
+                # Handle expandable lists - we need to examine the first item to determine column types
+                items = value.get("items", [])
+                
+                # Add an index column for this list
+                columns.append(f'"{key}_index" INTEGER')
+                
+                if items and isinstance(items[0], dict):
+                    # If the first item is a dictionary, we need to create columns for all its keys
+                    sample_item = items[0]
+                    # Flatten the dictionary with the key as prefix
+                    flattened_item = self._flatten_dict_for_schema(sample_item, key, "_")
+                    
+                    # Add columns for all keys in the flattened dictionary
+                    for item_key, item_value in flattened_item.items():
+                        item_column = self._get_column_definition(item_key, item_value)
+                        columns.append(item_column)
+                else:
+                    # If items are simple values, just add a column for the list key itself
+                    columns.append(self._get_column_definition(key, items[0] if items else None))
             else:
-                # Default to TEXT for strings and other types
-                columns.append(f'"{key}" TEXT')
+                # Regular handling for non-list fields
+                columns.append(self._get_column_definition(key, value))
         
         # Add metadata columns
         columns.extend([
@@ -651,9 +662,159 @@ class AlloyDBClient:
         
         return result
 
+    def _get_column_definition(self, key, value):
+        """
+        Helper method to get SQL column definition from a key and value.
+        
+        Args:
+            key (str): The column name
+            value: The value to determine the column type
+            
+        Returns:
+            str: SQL column definition
+        """
+        if value is None:
+            # For unknown types (None), default to TEXT
+            return f'"{key}" TEXT'
+        elif isinstance(value, dict):
+            # For nested objects, store as JSONB
+            return f'"{key}" JSONB'
+        elif isinstance(value, list):
+            # For arrays, store as JSONB
+            return f'"{key}" JSONB'
+        elif isinstance(value, int):
+            return f'"{key}" INTEGER'
+        elif isinstance(value, float):
+            return f'"{key}" NUMERIC'
+        elif isinstance(value, bool):
+            return f'"{key}" BOOLEAN'
+        else:
+            # Default to TEXT for strings and other types
+            return f'"{key}" TEXT'
+
+    def _flatten_dict_for_schema(self, nested_dict, parent_key='', separator='.'):
+        """
+        Flatten a nested dictionary for schema creation.
+        
+        Args:
+            nested_dict (dict): The nested dictionary to flatten
+            parent_key (str): The parent key for the current recursion level
+            separator (str): The separator to use between key levels
+            
+        Returns:
+            dict: A flattened dictionary
+        """
+        flattened = {}
+        
+        for key, value in nested_dict.items():
+            # Create the new key with parent_key if it exists
+            new_key = f"{parent_key}{separator}{key}" if parent_key else key
+            
+            # If value is a dictionary, recursively flatten it
+            if isinstance(value, dict):
+                flattened.update(self._flatten_dict_for_schema(value, new_key, separator))
+            else:
+                # For simple values, just add them with the new key
+                flattened[new_key] = value
+                
+        return flattened
+
+    def _flatten_dict(self, nested_dict, parent_key='', separator='.'):
+        """
+        Flatten a nested dictionary into a single-level dictionary with dot notation for keys.
+        
+        Args:
+            nested_dict (dict): The nested dictionary to flatten
+            parent_key (str): The parent key for the current recursion level
+            separator (str): The separator to use between key levels (default: '.')
+            
+        Returns:
+            dict: A flattened dictionary with special handling for lists
+        """
+        flattened = {}
+        
+        for key, value in nested_dict.items():
+            # Create the new key with parent_key if it exists
+            new_key = f"{parent_key}{separator}{key}" if parent_key else key
+            
+            # If value is a dictionary, recursively flatten it
+            if isinstance(value, dict):
+                flattened.update(self._flatten_dict(value, new_key, separator))
+            # Handle lists containing dictionaries or other values
+            elif isinstance(value, list):
+                # Mark lists for special processing during database insertion
+                # We'll use a special format to indicate this is a list that needs expansion
+                flattened[new_key] = {
+                    "__is_expandable_list__": True,
+                    "items": value
+                }
+            else:
+                # For simple values, just add them with the new key
+                flattened[new_key] = value
+                
+        return flattened
+
     async def write_data_to_table(self, table_name: str, data: dict, metadata: dict = None):
         """
-        Writes data to the specified table.
+        Writes data to the specified table, with special handling for expandable lists.
+        
+        Args:
+            table_name (str): Name of the table
+            data (dict): Data to write to the table
+            metadata (dict, optional): Additional metadata to include
+                
+        Returns:
+            List of results from SQL executions
+        """
+        # Find any expandable lists in the data
+        expandable_lists = {}
+        regular_data = {}
+        
+        for key, value in data.items():
+            if isinstance(value, dict) and value.get("__is_expandable_list__", False):
+                expandable_lists[key] = value["items"]
+            else:
+                regular_data[key] = value
+        
+        # If no expandable lists are found, do a simple insert
+        if not expandable_lists:
+            return await self._insert_single_row(table_name, regular_data, metadata)
+        
+        # For expandable lists, we need to create multiple rows
+        results = []
+        
+        # Create combinations of rows based on expandable lists
+        if expandable_lists:
+            # Get the first expandable list to start with
+            primary_list_key = next(iter(expandable_lists))
+            primary_list = expandable_lists[primary_list_key]
+            
+            # For each item in the primary list, create a new row
+            for item_idx, item in enumerate(primary_list):
+                # Create a copy of the regular data
+                row_data = dict(regular_data)
+                
+                # Add the current item from the primary list
+                if isinstance(item, dict):
+                    # If it's a dictionary, flatten it with the primary key as prefix
+                    flattened_item = self._flatten_dict(item, primary_list_key, "_")
+                    row_data.update(flattened_item)
+                else:
+                    # If it's a simple value, just add it with the list key
+                    row_data[primary_list_key] = item
+                
+                # Add item index for reference
+                row_data[f"{primary_list_key}_index"] = item_idx
+                
+                # Insert this row
+                result = await self._insert_single_row(table_name, row_data, metadata)
+                results.append(result)
+        
+        return results
+
+    async def _insert_single_row(self, table_name: str, data: dict, metadata: dict = None):
+        """
+        Inserts a single row of data into the specified table.
         
         Args:
             table_name (str): Name of the table
@@ -663,14 +824,15 @@ class AlloyDBClient:
         Returns:
             Result of SQL execution
         """
+        
         # Create copies to avoid modifying the original data
         insert_data = dict(data)
         
         # Add metadata if provided
         if metadata:
-            insert_data["source"] = metadata.get("objectId", metadata.get("source", "unknown"))
-            insert_data["extraction_backend"] = metadata.get("extraction_backend", "unknown")
-            insert_data["extraction_model"] = metadata.get("extraction_model", "unknown")
+            insert_data["source"] = metadata.get("objectId", metadata.get("source", "not-in-metadata"))
+            insert_data["extraction_backend"] = metadata.get("extraction_backend", "not-in-metadata")
+            insert_data["extraction_model"] = metadata.get("extraction_model", "not-in-metadata")
         
         # Prepare column names and values for SQL
         columns = [f'"{key}"' for key in insert_data.keys()]

{sunholo-0.124.0 → sunholo-0.125.1}/src/sunholo/gcs/download_url.py

@@ -36,7 +36,7 @@ def get_image_from_gcs(gs_uri: str) -> Image.Image: # type: ignore
     except IOError as e:
         raise ValueError("Unable to open image from bytes:", e)
 
-def get_bytes_from_gcs(gs_uri) -> Optional[bytes]:
+def get_bytes_from_gcs(gs_uri: str) -> Optional[bytes]:
     """
     Downloads a file from Google Cloud Storage and returns its bytes.
 

{sunholo-0.124.0 → sunholo-0.125.1/src/sunholo.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sunholo
-Version: 0.124.0
+Version: 0.125.1
 Summary: Large Language Model DevOps - a package to help deploy LLMs to the Cloud.
 Author-email: Holosun ApS <multivac@sunholo.com>
 License: Apache License, Version 2.0