Conversation Memory Buffer Class | Dataford Interview Questions

Implement a simple memory buffer class in Python that retains the last N turns of a conversation but prioritizes keeping system prompts and critical metadata intact.

A turn is a dictionary. Preserve turns whose role is system or whose type is metadata and critical is True; evict the oldest unprotected turns first. Implement MemoryBuffer(capacity), add(turn), and get_turns(), returning retained turns in original order. If protected turns exceed the capacity, retain all protected turns.

Constraints

0 <= capacity <= 10^4
0 <= len(turns) <= 10^4
Each turn is a dictionary
A protected turn has role equal to system, or type equal to metadata and critical equal to True
The relative order of retained turns must not change

Function Signature

def retain_turns(turns, capacity):

Your question is Conversation Memory Buffer Class. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Implement a simple memory buffer class in Python that retains the last N turns of a conversation but prioritizes keeping system prompts and critical metadata intact.

Constraints

0 <= capacity <= 10^4
0 <= len(turns) <= 10^4
Each turn is a dictionary
A protected turn has role equal to system, or type equal to metadata and critical equal to True
The relative order of retained turns must not change

Function Signature

def retain_turns(turns, capacity):

Your question is Conversation Memory Buffer Class. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.