step fixes, tests

verified_cogen/runners/chain_of_thought/step.py

@@ -6,6 +6,14 @@ def __init__(self, question: str, answer: str):
         self.question = question
         self.answer = answer
 
+    def __repr__(self) -> str:
+        return f"Substep(question={self.question}, answer={self.answer})"
+
+    def __eq__(self, value: object, /) -> bool:
+        if not isinstance(value, Substep):
+            return False
+        return self.question == value.question and self.answer == value.answer
+
 
 class Step:
     substeps: list[Substep] = []
@@ -26,10 +34,15 @@ def __init__(self, data: str):
             elif line.startswith("===== ANSWER ====="):
                 appending_to = current_answer
             elif line.startswith("===== QUESTION ====="):
+                self.substeps.append(
+                    Substep("\n".join(current_question), "\n".join(current_answer))
+                )
+
                 self.question = "\n".join(data_lines[i + 1 :])
                 break
             else:
                 appending_to.append(line)
+        self.substeps = self.substeps[1:]
 
     def __repr__(self) -> str:
         return f"Step(question={self.question}, substeps={self.substeps})"