From f4b2769cafa54b902179fa64d468b196cbf57c37 Mon Sep 17 00:00:00 2001
From: Thibault Sottiaux <tibo@openai.com>
Date: Sat, 23 Aug 2025 10:26:57 -0700
Subject: [PATCH] feat: add conversation storage

---
 codirect/conversation_storage.py | 66 ++++++++++++++++++++++++++++++++
 1 file changed, 66 insertions(+)
 create mode 100644 codirect/conversation_storage.py

diff --git a/codirect/conversation_storage.py b/codirect/conversation_storage.py
new file mode 100644
index 0000000000..80ecbf1600
--- /dev/null
+++ b/codirect/conversation_storage.py
@@ -0,0 +1,66 @@
+"""Storage utilities for conversations using boostedblob."""
+
+from __future__ import annotations
+
+import json
+from datetime import datetime
+from typing import Any, List
+
+import boostedblob
+from pydantic import BaseModel, Field
+
+
+CONVERSATIONS_PREFIX = "az://whistlerdata/codirect/conversations"
+
+
+class ConversationMetadata(BaseModel):
+    """Metadata about a stored conversation.
+
+    This includes several dummy fields for demonstration purposes.
+    """
+
+    conversation_id: str
+    title: str
+    user_id: str
+    created_at: datetime
+    language: str = "en"
+    tags: List[str] = Field(default_factory=list)
+    is_favorite: bool = False
+    model_name: str | None = None
+
+
+def _conversation_path(conversation_id: str) -> str:
+    return f"{CONVERSATIONS_PREFIX}/{conversation_id}"
+
+
+def write_conversation(
+    conversation_id: str,
+    metadata: ConversationMetadata,
+    conversation: Any,
+) -> None:
+    """Write a conversation and its metadata to object storage.
+
+    Args:
+        conversation_id: Unique identifier for the conversation.
+        metadata: Metadata describing the conversation.
+        conversation: The conversation payload (must be JSON serializable).
+    """
+
+    base = _conversation_path(conversation_id)
+    with boostedblob.open(f"{base}/metadata.json", "w") as f:
+        f.write(metadata.model_dump_json(indent=2))
+
+    with boostedblob.open(f"{base}/conversation.json", "w") as f:
+        json.dump(conversation, f, indent=2)
+
+
+def list_conversations() -> List[str]:
+    """List all conversation IDs under the conversations prefix."""
+
+    entries = boostedblob.listdir(CONVERSATIONS_PREFIX)
+    ids = []
+    for entry in entries:
+        name = entry.rstrip("/").split("/")[-1]
+        if name:
+            ids.append(name)
+    return ids