update default time unit from ps to s, add scientific notation to val…

biosim_extractor/helpers/metadata_utils.py

@@ -1,24 +1,33 @@
 #!/usr/bin/env python3
 
 
-def round_floats(obj, decimals=3):
+def round_floats(obj, decimals=3, preserve_below=1e-3):
     """
-    Recursively round all floats in a nested structure (dict, list, tuple) to the given decimals.
+    Recursively round floats in nested dicts, lists, and tuples.
+
+    Very small non-zero floats are rounded in scientific notation so they are
+    not rounded to zero. They remain numbers.
 
     Args:
-        obj: The object to process (dict, list, tuple, float, etc.).
-        decimals: Number of decimal places to round to.
+        obj: Object to process.
+        decimals: Decimal places to keep.
+        preserve_below: Lower absolute-value threshold for preserving floats.
 
     Returns:
-        The processed object with all floats rounded.
+        Object with floats rounded while preserving numeric types.
     """
     if isinstance(obj, float):
+        if obj != 0 and abs(obj) < preserve_below:
+            return float(f"{obj:.{decimals}e}")
         return round(obj, decimals)
+
     elif isinstance(obj, list):
-        return [round_floats(item, decimals) for item in obj]
+        return [round_floats(item, decimals, preserve_below) for item in obj]
+
     elif isinstance(obj, tuple):
-        return tuple(round_floats(item, decimals) for item in obj)
+        return tuple(round_floats(item, decimals, preserve_below) for item in obj)
+
     elif isinstance(obj, dict):
-        return {k: round_floats(v, decimals) for k, v in obj.items()}
-    else:
-        return obj
+        return {k: round_floats(v, decimals, preserve_below) for k, v in obj.items()}
+
+    return obj

biosim_extractor/units/unitconversion.py

@@ -25,7 +25,7 @@ def __init__(self, standard_units: Optional[Dict[str, str]] = None) -> None:
         self.standards: Dict[str, str] = standard_units or {
             "length": "nm",  # nanometers
             "volume": "nm³",  # nanometers cubed
-            "time": "ps",  # picoseconds
+            "time": "s",  # seconds
             "energy": "kJ/mol",  # kilojoules per mole
             "temperature": "K",  # Kelvin
             "pressure": "bar",  # bar
@@ -37,6 +37,7 @@ def __init__(self, standard_units: Optional[Dict[str, str]] = None) -> None:
             "angle": "degree",  # degree
             "charge": "e",  # atomic charge units
             "molecular_weight": "g/mol",  # grams per mol
+            "information_storage": "GB",  # gigabytes
         }
 
         # Conversion factors to standard units
@@ -59,10 +60,12 @@ def __init__(self, standard_units: Optional[Dict[str, str]] = None) -> None:
                 "ų": 1000,
             },
             "time": {
-                "fs": 0.001,
-                "ps": 1.0,
-                "ns": 1000.0,
-                "s": 1e12,
+                "fs": 1e-15,
+                "ps": 1e-12,
+                "ns": 1e-9,
+                "μs": 1e-6,
+                "ms": 0.001,
+                "s": 1.0,
             },
             "energy": {
                 "kcal/mol": 4.184,
@@ -120,6 +123,10 @@ def __init__(self, standard_units: Optional[Dict[str, str]] = None) -> None:
                 "amu": 1.0,
                 "Da": 1.0,
             },
+            "information_storage": {
+                "GB": 1.0,
+                "MB": 0.001,
+            },
         }
 
         # Create reverse lookup: unit -> unit_type