feat: Refactored utils, moved function to calculate file size (#170)

* feat: Refactored utils, moved function to calculate file size

* Update src/unstract/sdk/utils/common_utils.py

Co-authored-by: vishnuszipstack <[email protected]>
Signed-off-by: Chandrasekharan M <[email protected]>

---------

Signed-off-by: Chandrasekharan M <[email protected]>
Co-authored-by: vishnuszipstack <[email protected]>

Author: chandrasekharan-zipstack

src/unstract/sdk/tool/stream.py

@@ -7,7 +7,7 @@
 from deprecated import deprecated
 
 from unstract.sdk.constants import Command, LogLevel, LogStage, ToolEnv
-from unstract.sdk.utils import ToolUtils
+from unstract.sdk.utils import Utils
 from unstract.sdk.utils.common_utils import UNSTRACT_TO_PY_LOG_LEVEL
 
 
@@ -29,7 +29,7 @@ def __init__(self, log_level: LogLevel = LogLevel.INFO, **kwargs) -> None:
 
         """
         self.log_level = log_level
-        self._exec_by_tool = ToolUtils.str_to_bool(
+        self._exec_by_tool = Utils.str_to_bool(
             os.environ.get(ToolEnv.EXECUTION_BY_TOOL, "False")
         )
         if self.is_exec_by_tool:

src/unstract/sdk/tool/validator.py

@@ -8,6 +8,7 @@
 from unstract.sdk.constants import MetadataKey, PropKey
 from unstract.sdk.tool.base import BaseTool
 from unstract.sdk.tool.mime_types import EXT_MIME_MAP
+from unstract.sdk.utils import Utils
 
 
 def extend_with_default(validator_class: Any) -> Any:
@@ -119,7 +120,7 @@ def _validate_file_size(self, input_file: Path) -> None:
             f"Checking input file size... (max file size: {max_file_size})"
         )
         file_size = self.tool.workflow_filestorage.size(path=input_file)
-        self.tool.stream_log(f"Input file size: {self._human_readable_size(file_size)}")
+        self.tool.stream_log(f"Input file size: {Utils.pretty_file_size(file_size)}")
 
         if file_size > max_size_in_bytes:
             source_name = self.tool.get_exec_metadata.get(MetadataKey.SOURCE_NAME)
@@ -128,22 +129,6 @@ def _validate_file_size(self, input_file: Path) -> None:
                 f"allowed size of {max_file_size}"
             )
 
-    def _human_readable_size(self, num: float, suffix: str = "B") -> str:
-        """Gets the human readable size for a file,
-
-        Args:
-            num (int): Size in bytes to parse
-            suffix (str, optional): _description_. Defaults to "B".
-
-        Returns:
-            str: Human readable size
-        """
-        for unit in ("", "K", "M", "G", "T"):
-            if abs(num) < 1024.0:
-                return f"{num:3.1f}{unit}{suffix}"
-            num /= 1024.0
-        return f"{num:.1f}{suffix}"
-
     def _parse_size_string(self, size_string: str) -> int:
         """Parses the size string for validation.
 

src/unstract/sdk/utils/__init__.py

@@ -1,2 +1,3 @@
+from .common_utils import CommonUtils, Utils  # noqa
 from .file_storage_utils import FileStorageUtils  # noqa
 from .tool_utils import ToolUtils  # noqa

src/unstract/sdk/utils/common_utils.py

@@ -9,12 +9,48 @@
 logger = logging.getLogger(__name__)
 
 
-class CommonUtils:
+class Utils:
     @staticmethod
     def generate_uuid() -> str:
         """Class method to get uuid."""
         return str(uuid.uuid4())
 
+    @staticmethod
+    def str_to_bool(string: str) -> bool:
+        """String value of boolean to boolean.
+
+        Useful while parsing envs to bool.
+
+        Args:
+            string (str): value like "true", "True" etc..
+
+        Returns:
+            bool
+        """
+        return string.lower() == "true"
+
+    @staticmethod
+    def pretty_file_size(num: float, suffix: str = "B") -> str:
+        """Gets the human readable size for a file,
+
+        Args:
+            num (int): Size in bytes to parse
+            suffix (str, optional): _description_. Defaults to "B".
+
+        Returns:
+            str: Human readable size
+        """
+        for unit in ("", "K", "M", "G", "T"):
+            if abs(num) < 1024.0:
+                return f"{num:.2f} {unit}{suffix}"
+            num /= 1024.0
+        return f"{num:.2f} {suffix}"
+
+
+# TODO: Kept for backward compatibility, make use of Utils instead
+class CommonUtils(Utils):
+    pass
+
 
 # Mapping from python log level to Unstract counterpart
 PY_TO_UNSTRACT_LOG_LEVEL = {

src/unstract/sdk/utils/tool_utils.py

@@ -161,20 +161,6 @@ def get_file_size(
         file_length = fs.size(path=input_file)
         return file_length
 
-    @staticmethod
-    def str_to_bool(string: str) -> bool:
-        """String value of boolean to boolean.
-
-        Useful while parsing envs to bool.
-
-        Args:
-            string (str): value like "true", "True" etc..
-
-        Returns:
-            bool
-        """
-        return string.lower() == "true"
-
     # Used the same function from LLMWhisperer
     @staticmethod
     def calculate_page_count(