Add new classmethod

meeteval/io/base.py

@@ -4,13 +4,14 @@
 import typing
 from pathlib import Path
 import contextlib
-from typing import Dict, List, NamedTuple
+from typing import Dict, List
 import dataclasses
 from dataclasses import dataclass
 from itertools import groupby
 
 if typing.TYPE_CHECKING:
     from typing import Self
+    from meeteval.io.seglst import SegLstSegment, SegLSTMixin
     from meeteval.io.uem import UEM, UEMLine
     from meeteval.io.stm import STM, STMLine
     from meeteval.io.ctm import CTM, CTMLine
@@ -119,16 +120,11 @@ def replace(self, **kwargs) -> 'Self':
         return dataclasses.replace(self, **kwargs)
 
 
-class Base:
+class Base(SegLSTMixin):
     lines: 'List[LineSubclasses]'
     line_cls = 'LineSubclasses'
 
-    def __init__(self, data, **defaults):
-        if isinstance(data, self.__class__):
-            self.lines = data.lines
-        elif hasattr(data, 'to_seglst'):
-            self.lines = [self.line_cls.from_seglst({**defaults, **segment}) for segment in data.to_seglst()]
-        else:
+    def __init__(self, data):
             self.lines = data
 
     @classmethod

meeteval/io/ctm.py

@@ -13,7 +13,7 @@
     'CTMGroup',
 ]
 
-from meeteval.io.seglst import SegLstSegment, SegLST
+from meeteval.io.seglst import SegLstSegment, SegLST, SegLSTMixin
 
 
 @dataclass(frozen=True)
@@ -128,7 +128,7 @@ def from_seglst(cls, s: 'SegLST', **defaults) -> 'Self':
 
 
 @dataclass(frozen=True)
-class CTMGroup:
+class CTMGroup(SegLSTMixin):
     ctms: 'Dict[str, CTM]'
 
     @classmethod

meeteval/io/pbjson.py

@@ -8,7 +8,7 @@
 import json
 from pathlib import Path
 
-from meeteval.io.seglst import SegLST
+from meeteval.io.seglst import SegLST, SegLSTMixin
 
 
 def _load_json(file):
@@ -42,7 +42,7 @@ def get_sample_rate(ex):
     return sample_rate
 
 
-class PBJson:
+class PBJson(SegLSTMixin):
     """
     The JSON format used at the NT department internally for storing databases.
 

meeteval/io/py.py

@@ -5,7 +5,7 @@
 
 if typing.TYPE_CHECKING:
     from typing import Self
-    from meeteval.io.seglst import SegLST
+    from meeteval.io.seglst import SegLST, SegLSTMixin
 
 
 def _convert_python_structure(structure, *, keys=(), final_key='words', final_types=str):
@@ -120,7 +120,7 @@ def _invert_python_structure(t: 'SegLST', types, keys):
 
 
 @dataclasses.dataclass(frozen=True)
-class NestedStructure:
+class NestedStructure(SegLSTMixin):
     """
     Wraps a Python structure where the structure levels represent keys.
 

meeteval/io/seglst.py

@@ -27,6 +27,20 @@ class SegLstSegment(TypedDict, total=False):
     confidence: float
 
 
+class SegLSTMixin:
+    @classmethod
+    def new(cls, d, **defaults):
+        return cls.from_seglst(asseglst(d, **defaults))
+
+    @classmethod
+    def from_seglst(cls, d: 'SegLST', **defaults):
+        raise NotImplementedError()
+
+    def to_seglst(self):
+        raise NotImplementedError()
+
+
+
 @dataclasses.dataclass(frozen=True)
 class SegLST:
     """