iil-chat-logging 0.2.0

Persistent chat conversation logging & QM for Django apps

chat-logging

Persistent chat conversation logging & quality management for Django apps.

Part of the platform ecosystem (ADR-037). Works with any app using chat-agent.

Features

• ChatConversation + ChatMessage — Full conversation persistence
• UseCaseCandidate — Track unmet user needs
• EvaluationScore — Store evaluation metrics (DeepEval, LangFuse)
• LoggingSessionBackend — Drop-in wrapper for any SessionBackend
• Django Admin — Filter, search, review, export conversations

Installation

pip install chat-logging

Add to INSTALLED_APPS:

INSTALLED_APPS = [
    ...
    "chat_logging",
]

Run migrations:

python manage.py migrate chat_logging

Usage

from chat_agent import ChatAgent, InMemorySessionBackend
from chat_logging import LoggingSessionBackend

backend = LoggingSessionBackend(
    wrapped=InMemorySessionBackend(),
    app_name="drifttales",
    user=request.user,
)

agent = ChatAgent(
    toolkit=my_toolkit,
    completion=my_backend,
    session_backend=backend,  # conversations are now logged!
    system_prompt="...",
)

Export

from chat_logging.exporters import (
    export_conversations_csv,
    export_conversations_jsonl,
)
from chat_logging.models import ChatConversation

qs = ChatConversation.objects.filter(app_name="drifttales")
jsonl = export_conversations_jsonl(qs)
csv_data = export_conversations_csv(qs)