Initial pass at backend changes to support agents

- Add a db model for Agents, attaching them to conversations
- When an agent is added to a conversation, override the system prompt to tweak the instructions
- Agents can be configured with prompt modification, model specification, a profile picture, and other things
- Admin-configured models will not be editable by individual users
- Add unit tests to verify agent behavior. Unit tests demonstrate imperfect adherence to prompt specifications

tests/conftest.py

@@ -12,6 +12,7 @@ from khoj.configure import (
     configure_search_types,
 )
 from khoj.database.models import (
+    Agent,
     GithubConfig,
     GithubRepoConfig,
     KhojApiUser,
@@ -181,6 +182,28 @@ def api_user4(default_user4):
     )
 
 
+@pytest.mark.django_db
+@pytest.fixture
+def offline_agent():
+    chat_model = ChatModelOptionsFactory()
+    return Agent.objects.create(
+        name="Accountant",
+        chat_model=chat_model,
+        tuning="You are a certified CPA. You are able to tell me how much I've spent based on my notes. Regardless of what I ask, you should always respond with the total amount I've spent. ALWAYS RESPOND WITH A SUMMARY TOTAL OF HOW MUCH MONEY I HAVE SPENT.",
+    )
+
+
+@pytest.mark.django_db
+@pytest.fixture
+def openai_agent():
+    chat_model = ChatModelOptionsFactory(chat_model="gpt-3.5-turbo", model_type="openai")
+    return Agent.objects.create(
+        name="Accountant",
+        chat_model=chat_model,
+        tuning="You are a certified CPA. You are able to tell me how much I've spent based on my notes. Regardless of what I ask, you should always respond with the total amount I've spent.",
+    )
+
+
 @pytest.fixture(scope="session")
 def search_models(search_config: SearchConfig):
     search_models = SearchModels()