Feature: support for typed features in basic reader configuration

jmhorcas · jmhorcas · commit 2cda1e9ce6b8 · 2025-07-11T02:27:27.000+02:00
diff --git a/flamapy/metamodels/configuration_metamodel/transformations/configuration_basic_reader.py b/flamapy/metamodels/configuration_metamodel/transformations/configuration_basic_reader.py
@@ -1,35 +1,92 @@
 import csv
+import ast
+from typing import Any
+import pathlib
 
 from flamapy.core.transformations.text_to_model import TextToModel
+from flamapy.core.exceptions import ConfigurationNotFound
 
 from flamapy.metamodels.configuration_metamodel.models.configuration import Configuration
-from flamapy.core.utils import file_exists
-from flamapy.core.exceptions import ConfigurationNotFound
 
 
 class ConfigurationBasicReader(TextToModel):
+    """Reads a configuration from a csvconf file and transforms it into a Configuration model.
+    
+    The csvconf file should have two columns (separated by commas):
+    - The first column contains the element names (e.g., features).
+    - The second column contains the values for those elements.
+    Values can be of various types, including Boolean, numeric, strings, lists, or dictionaries.
+
+    If the path is a directory, it will read all csvconf files in that directory and concatenate
+    their configurations into a single Configuration object.
+    This allows for managing multiple configurations associated with variability models split in
+    several files (e.g., imports in UVL models).
+
+    If the csvconf file or directory does not exist, it raises a ConfigurationNotFound exception.
+    """
+
     @staticmethod
     def get_source_extension() -> str:
         return 'csvconf'
 
     def __init__(self, path: str) -> None:
-        self._path = path
+        self._path: str = path
 
     def transform(self) -> Configuration:
-        csv_reader = self.get_configuration_from_csv(self._path)
+        path = pathlib.Path(self._path)
+        if path.is_file():
+            return self.get_configuration_from_csv(path)
+        elif path.is_dir():
+            return self.get_configuration_from_directory(path)
+        else:
+            raise ConfigurationNotFound
+
+    def get_configuration_from_csv(self, path: pathlib.Path) -> Configuration:
         elements = {}
-        for row in csv_reader:
-            # Assuming that row[1] is supposed to represent a boolean value
-            # Convert 'true'/'false' strings to actual boolean values
-            elements[row[0]] = row[1].lower() == 'true'
+        with open(path, 'r', encoding='utf-8') as csvfile:
+            csv_reader = list(csv.reader(csvfile))
+            for row in csv_reader:
+                element = row[0]
+                value = convert(row[1])
+                elements[element] = value
+        return Configuration(elements)
+    
+    def get_configuration_from_directory(self, directory: pathlib.Path) -> Configuration:
+        """Reads all CSV files in a directory and returns a Configuration object as
+        result of concatenating all configurations."""
+        elements = {}
+        for filepath in directory.rglob('*.csvconf'):
+            if filepath.is_file():
+                config = self.get_configuration_from_csv(filepath)
+                elements.update(config.elements)
         return Configuration(elements)
 
-    def get_configuration_from_csv(self, path: str) -> list[list[str]]:
-        # Returns a list of list
-        if not file_exists(path):
-            raise ConfigurationNotFound
 
-        with open(path, 'r', encoding='utf-8') as csvfile:
-            csvreader = list(csv.reader(csvfile))
+def convert(value: str) -> Any:
+    """
+    Converts a string to the most appropriate Python type.
+    
+    Handles:
+    - Boolean values like 'true', 'FALSE', etc.
+    - Numeric values (int, float)
+    - Python literals (lists, dicts, tuples, None)
+    - Returns the original string if no conversion is possible
+    """
+    stripped = value.strip()
+    # Empty or whitespace-only → None
+    if stripped == '':
+        return None
+
+    # Normalize and check for boolean values
+    val_lower = value.strip().lower()
+    if val_lower == 'true':
+        return True
+    if val_lower == 'false':
+        return False
 
-        return csvreader
+    try:
+        # Try to evaluate safe Python literals (e.g. numbers, lists, tuples)
+        return ast.literal_eval(value)
+    except (ValueError, SyntaxError):
+        # If evaluation fails, return the string itself
+        return value
