Add offline unit test suite for chat_backend (#12)

Closes #5

## Summary

- Replaces the three scattered test modules (`chat_backend/tests.py`, `services/tests.py`, `services/prompt_classifier/tests.py`) with a `chat_backend/tests/` package: **242 deterministic tests plus 6 opt-in live-Ollama checks**, up from 10 tests (3 of which were skipped and 4 of which were never even discovered).
- The suite runs fully offline — no Ollama, Chroma, SMTP or network access. LangChain runnables are replaced by a small `FakeChain`, Chroma/embeddings are mocked, email uses Django's locmem backend, and blobs go through `DatabaseStorage`.
- New `llm_be/test_runner.py` (wired via `TEST_RUNNER`) sets `SKIP_RAG_INIT=1` and an MD5 password hasher, so the suite cannot accidentally reach a model server and finishes in ~5s on SQLite (~17s on Postgres) instead of ~30s.

## Coverage

| Area | File |
|------|------|
| `TimeInfoBase.save`, slugs, `get_duration`, `file_exists`, cascades | `test_models.py` |
| `DatabaseStorage` save/open/exists/size/listdir/delete/times | `test_storage.py` |
| JWT claim, prompt/user/feedback/document serializers | `test_serializers.py` |
| auth + token, invite, feedback, company users, set-password, TOS | `test_views_users.py` |
| conversation list/order/create/detail/soft-delete | `test_views_conversations.py` |
| all four analytics endpoints, including empty-month behaviour | `test_views_analytics.py` |
| workspace + document upload/list/detail, 404 and 400 paths | `test_views_documents.py` |
| prompt classifier rules/parsing, moderation fail-safe, title cleanup | `test_services_classifiers.py` |
| CSV/XLSX/DOCX/PDF analysis, plot generation, error payloads | `test_services_data_analysis.py` |
| loader selection, filename sanitising, ingest, temp-file cleanup, search filters | `test_services_rag.py` |
| history formatting and streaming | `test_services_llm.py` |
| document re-index on create/delete, `SKIP_RAG_INIT` guard | `test_signals.py` |
| consumer DB helpers, LangGraph nodes (moderation, classification, generation, search flags), websocket routes | `test_consumers.py` |

Live checks (non-deterministic, need a model server):

```bash
cd llm_be
RUN_LIVE_OLLAMA_TESTS=1 uv run python manage.py test chat_backend.tests.test_live_ollama
```

## Bugs the tests surfaced (fixed here)

1. **`ConversationDetailView.post` silently dropped every prompt.** `import datetime` shadowed `from datetime import datetime`, so `datetime.now()` raised `AttributeError` inside a bare `except` and the endpoint returned 200 without saving. Now uses `timezone.now()`.
2. **Prompt attachments never reached the LLM.** `get_conversation_file_async` (both consumers) did `sync_to_async(prompt.file.read)` — with `DatabaseStorage` the attribute access itself opens the blob, i.e. a DB query in async context, raising `SynchronousOnlyOperation` that was swallowed and returned `(None, None)`. The read now happens inside the thread.
3. **`DatabaseStorage._save` crashed on a str-backed `ContentFile`** (`TypeError: sequence item 0: expected a bytes-like object`); chunks are encoded when needed.
4. **`services/prompt_classifier/__init__.,py`** (note the comma) meant the directory was only an implicit namespace package, which is why its test module was never collected. Renamed, and its duplicate live-Ollama tests folded into `test_live_ollama.py`.

Known-broken paths deliberately left untested and unchanged: `reset_password` / `ResetUserPassword` reference an unimported `requests` plus undefined locals, and `DocumentDetailView.get` references an undefined `workspaces` on its success path. Worth a follow-up ticket.

## Test plan

- [x] `cd llm_be && uv run python manage.py test` → 248 tests, OK (6 skipped, all opt-in live)
- [x] Same suite against Postgres 16 (`DATABASE_URL=postgres://…`) → OK, matching the containerized run in `deploy.yml`
- [x] `uv run black` clean on all added files
- [ ] Gitea Actions **Unit Tests** + **CI** green on this PRReviewed-on: #12

llm_be/chat_backend/consumers.py

@@ -196,8 +196,8 @@ async def get_conversation_file_async(conversation_id):
         ).exclude(file='').order_by('created').afirst()
 
         if prompt_with_file and prompt_with_file.file:
-            # You must use sync_to_async to access the file's binary content
-            file_data = await sync_to_async(prompt_with_file.file.read)()
+            # Opening a DatabaseStorage file hits the DB, so read inside the thread.
+            file_data = await sync_to_async(lambda: prompt_with_file.file.read())()
             file_type = prompt_with_file.file_type
             return file_data, file_type
     except Exception as e:

llm_be/chat_backend/consumers_graph.py

@@ -7,6 +7,7 @@ from typing import TypedDict, Annotated, List, Union, Dict, Any
 from django.utils import timezone
 from django.conf import settings
 from django.core.files.base import ContentFile
+from asgiref.sync import sync_to_async
 from channels.generic.websocket import AsyncWebsocketConsumer
 from channels.db import database_sync_to_async
 from langchain_core.messages import HumanMessage, AIMessage, BaseMessage
@@ -139,7 +140,8 @@ async def get_conversation_file_async(conversation_id):
         ).exclude(file='').order_by('created').afirst()
 
         if prompt_with_file and prompt_with_file.file:
-            file_data = await sync_to_async(prompt_with_file.file.read)()
+            # Opening a DatabaseStorage file hits the DB, so read inside the thread.
+            file_data = await sync_to_async(lambda: prompt_with_file.file.read())()
             file_type = prompt_with_file.file_type
             return file_data, file_type
     except Exception as e:

llm_be/chat_backend/services/prompt_classifier/tests.py

@@ -1,31 +0,0 @@
-import os
-from unittest import TestCase, mock
-from unittest.mock import MagicMock, patch, AsyncMock
-from typing import List, Dict, Any
-
-from django.test import TestCase as DjangoTestCase
-
-from chat_backend.services.rag_services import (
-    RAGService,
-    SyncRAGService,
-    AsyncRAGService,
-)
-from chat_backend.models import Conversation, Prompt, DocumentWorkspace, Document
-from chat_backend.services.prompt_classifier.prompt_classifier import PromptClassifier, PromptType
-from parameterized import parameterized
-
-
-
-class PromptClassifierTestCase(TestCase):
-    def setUp(self):
-        self.service = PromptClassifier()
-
-    @parameterized.expand([
-        ["Tell me a joke",PromptType.GENERAL_CHAT],
-        ["Create an image of a dog for me",PromptType.IMAGE_GENERATION],
-        ["highlight the features of the backyard playset if they were to choose us and make the language more long form",PromptType.GENERAL_CHAT],
-        ["Great, can you make it about a duck now", PromptType.IMAGE_GENERATION],
-    ])
-    def test_prompt_classification(self, prompt, expected_output):
-        result = self.service.classify(prompt)
-        self.assertEqual(result, expected_output)

llm_be/chat_backend/services/tests.py

@@ -1,32 +0,0 @@
-import os
-import unittest
-from unittest import TestCase
-
-from chat_backend.services.prompt_classifier.prompt_classifier import (
-    PromptClassifier,
-    PromptType,
-)
-from parameterized import parameterized
-
-
-@unittest.skipIf(
-    os.environ.get("SKIP_RAG_INIT", "").lower() in {"1", "true", "yes"},
-    "Requires live Ollama; skipped when SKIP_RAG_INIT is set",
-)
-class PromptClassifierTestCase(TestCase):
-    def setUp(self):
-        self.service = PromptClassifier()
-
-    @parameterized.expand(
-        [
-            ["Tell me a joke", PromptType.GENERAL_CHAT],
-            ["Create an image of a dog for me", PromptType.IMAGE_GENERATION],
-            [
-                "highlight the features of the backyard playset if they were to choose us and make the language more long form",
-                PromptType.GENERAL_CHAT,
-            ],
-        ]
-    )
-    def test_prompt_classification(self, prompt, expected_output):
-        result = self.service.classify(prompt)
-        self.assertEqual(result, expected_output)

llm_be/chat_backend/storage.py

@@ -27,13 +27,17 @@ class DatabaseStorage(Storage):
     def _save(self, name, content):
         name = self.get_available_name(name)
         if hasattr(content, "chunks"):
-            data = b"".join(chunk for chunk in content.chunks())
+            chunks = content.chunks()
         else:
-            data = content.read()
-            if isinstance(data, str):
-                data = data.encode("utf-8")
+            chunks = [content.read()]
+        data = b"".join(
+            chunk.encode("utf-8") if isinstance(chunk, str) else chunk
+            for chunk in chunks
+        )
 
-        content_type = getattr(content, "content_type", None) or mimetypes.guess_type(name)[0]
+        content_type = (
+            getattr(content, "content_type", None) or mimetypes.guess_type(name)[0]
+        )
         StoredFile = self._model()
         with transaction.atomic():
             StoredFile.objects.update_or_create(
@@ -56,8 +60,10 @@ class DatabaseStorage(Storage):
         prefix = path.rstrip("/")
         if prefix:
             prefix = f"{prefix}/"
-        names = self._model().objects.filter(name__startswith=prefix).values_list(
-            "name", flat=True
+        names = (
+            self._model()
+            .objects.filter(name__startswith=prefix)
+            .values_list("name", flat=True)
         )
         dirs: set[str] = set()
         files: list[str] = []
@@ -88,4 +94,6 @@ class DatabaseStorage(Storage):
         return self._model().objects.values_list("created", flat=True).get(name=name)
 
     def get_modified_time(self, name):
-        return self._model().objects.values_list("last_modified", flat=True).get(name=name)
+        return (
+            self._model().objects.values_list("last_modified", flat=True).get(name=name)
+        )

llm_be/chat_backend/tests.py

@@ -1,194 +0,0 @@
-from django.test import TestCase
-
-# Create your tests here.
-from django.test import TestCase, Client
-from django.urls import reverse
-from django.contrib.auth.models import User
-from rest_framework.test import APIClient, APITestCase
-from rest_framework import status
-from .models import DocumentWorkspace, Document, Company
-from django.contrib.auth import get_user_model
-import tempfile
-from django.core.files.uploadedfile import SimpleUploadedFile
-from parameterized import parameterized
-
-
-# Minimal valid PDF bytes
-VALID_PDF_BYTES = (
-    b"%PDF-1.3\n"
-    b"1 0 obj\n"
-    b"<< /Type /Catalog /Pages 2 0 R >>\n"
-    b"endobj\n"
-    b"2 0 obj\n"
-    b"<< /Type /Pages /Kids [3 0 R] /Count 1 >>\n"
-    b"endobj\n"
-    b"3 0 obj\n"
-    b"<< /Type /Page /Parent 2 0 R /Resources << >> /MediaBox [0 0 612 792] /Contents 4 0 R >>\n"
-    b"endobj\n"
-    b"4 0 obj\n"
-    b"<< /Length 44 >>\n"
-    b"stream\n"
-    b"BT /F1 12 Tf 72 720 Td (Test PDF) Tj ET\n"
-    b"endstream\n"
-    b"endobj\n"
-    b"xref\n"
-    b"0 5\n"
-    b"0000000000 65535 f \n"
-    b"0000000009 00000 n \n"
-    b"0000000058 00000 n \n"
-    b"0000000117 00000 n \n"
-    b"0000000223 00000 n \n"
-    b"trailer\n"
-    b"<< /Size 5 /Root 1 0 R >>\n"
-    b"startxref\n"
-    b"317\n"
-    b"%%EOF"
-)
-
-
-class DocumentWorkspaceViewsTestCase(APITestCase):
-    def setUp(self):
-        self.company = Company.objects.create(
-            name="test", state="IL", zipcode="60189", address="1968 Greensboro Dr"
-        )
-        self.user = get_user_model().objects.create_user(
-            company=self.company,
-            username="testuser",
-            password="testpass123",
-            email="test@test.com",
-        )
-
-        self.client = APIClient()
-        self.client.force_authenticate(user=self.user)
-
-        self.workspace = DocumentWorkspace.objects.create(
-            company=self.user.company, name="Test Workspace"
-        )
-
-    def test_list_workspaces(self):
-        url = reverse("document_workspaces")
-        response = self.client.get(url)
-        self.assertEqual(response.status_code, status.HTTP_200_OK)
-        self.assertEqual(len(response.data), 1)
-        self.assertEqual(response.data[0]["name"], "Test Workspace")
-
-    def test_create_workspace(self):
-        url = reverse("document_workspaces")
-        data = {"name": "New Workspace"}
-        response = self.client.post(url, data, format="json")
-        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
-        self.assertEqual(DocumentWorkspace.objects.count(), 2)
-
-    def test_retrieve_workspace(self):
-        url = reverse("document_workspaces")
-        response = self.client.get(url)
-        self.assertEqual(response.status_code, status.HTTP_200_OK)
-        self.assertEqual(response.data[0]["name"], "Test Workspace")
-
-    # def test_update_workspace(self):
-    #     url = reverse('document_workspaces')
-    #     data = {
-    #         'name': 'Updated Workspace'
-    #     }
-    #     response = self.client.post(url, data, format='json')
-    #     self.assertEqual(response.status_code, status.HTTP_201_CREATED)
-    #     self.workspace.refresh_from_db()
-    #     self.assertEqual(self.workspace.name, 'Updated Workspace')
-
-    # def test_delete_workspace(self):
-    #     url = reverse('document_workspaces', args=[self.workspace.id])
-    #     response = self.client.delete(url)
-    #     self.assertEqual(response.status_code, status.HTTP_204_NO_CONTENT)
-    #     self.assertEqual(DocumentWorkspace.objects.count(), 0)
-
-
-class DocumentViewsTestCase(APITestCase):
-    def setUp(self):
-        self.company = Company.objects.create(
-            name="test", state="IL", zipcode="60189", address="1968 Greensboro Dr"
-        )
-        self.user = get_user_model().objects.create_user(
-            company=self.company,
-            username="testuser",
-            password="testpass123",
-            email="test@test.com",
-        )
-
-        self.client = APIClient()
-        self.client.force_authenticate(user=self.user)
-
-        self.workspace = DocumentWorkspace.objects.create(
-            company=self.user.company, name="Test Workspace"
-        )
-
-        # Create a test file
-        self.test_file = SimpleUploadedFile(
-            "test.pdf", VALID_PDF_BYTES, content_type="application/pdf"
-        )
-
-    def test_upload_document(self):
-        url = reverse("documents")
-        data = {"file": self.test_file}
-        response = self.client.post(url, data, format="multipart")
-        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
-        self.assertEqual(Document.objects.count(), 1)
-
-        document = Document.objects.first()
-        self.assertEqual(document.workspace.id, self.workspace.id)
-        self.assertTrue(document.processed)  # Should be False initially
-
-    def test_list_documents(self):
-        # First create a document
-        Document.objects.create(workspace=self.workspace, file=self.test_file)
-
-        url = reverse("documents")
-        response = self.client.get(url)
-        self.assertEqual(response.status_code, status.HTTP_200_OK)
-        self.assertEqual(len(response.data), 1)
-        self.assertIn("test", response.data[0]["file"])
-        self.assertIn("pdf", response.data[0]["file"])
-
-    # def test_delete_document(self):
-    #     document = Document.objects.create(
-    #         workspace=self.workspace,
-    #         file=self.test_file
-    #     )
-
-    #     url = reverse('document-detail', args=[document.id])
-    #     response = self.client.delete(url)
-    #     self.assertEqual(response.status_code, status.HTTP_204_NO_CONTENT)
-    #     self.assertEqual(Document.objects.count(), 0)
-
-    def test_upload_invalid_file(self):
-        url = reverse("documents")
-        data = {"file": "not a file"}
-        response = self.client.post(url, data, format="multipart")
-        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
-
-    def test_access_other_users_documents(self):
-        # Create another user
-        other_company = Company.objects.create(
-            name="test2", state="IL", zipcode="60189", address="1968 Greensboro Dr"
-        )
-        other_user = get_user_model().objects.create_user(
-            company=other_company,
-            username="otheruser",
-            password="otherpass123",
-            email="testing2@test.com",
-        )
-        other_workspace = DocumentWorkspace.objects.create(
-            company=other_user.company, name="Other Workspace"
-        )
-        other_document = Document.objects.create(
-            workspace=other_workspace, file=self.test_file
-        )
-
-        # Try to access the other user's document
-        url = reverse("documents_details", kwargs={"document_id": other_document.id})
-        response = self.client.get(url)
-        self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
-
-
-
-
-

llm_be/chat_backend/tests/__init__.py

@@ -0,0 +1,6 @@
+"""Unit tests for the chat_backend app.
+
+Every test in this package runs offline: no Ollama, Chroma, SMTP or network calls.
+LLM chains are replaced with fakes, email uses the locmem backend Django installs
+for tests, and file storage is the database-backed ``DatabaseStorage``.
+"""

llm_be/chat_backend/tests/factories.py

@@ -0,0 +1,131 @@
+"""Shared fixtures/helpers for chat_backend tests."""
+
+from __future__ import annotations
+
+from django.contrib.auth import get_user_model
+from django.core.files.uploadedfile import SimpleUploadedFile
+
+from chat_backend.models import (
+    Company,
+    Conversation,
+    Document,
+    DocumentWorkspace,
+    Prompt,
+)
+
+# Minimal valid PDF bytes — enough for pypdf/PyPDFLoader to parse.
+VALID_PDF_BYTES = (
+    b"%PDF-1.3\n"
+    b"1 0 obj\n"
+    b"<< /Type /Catalog /Pages 2 0 R >>\n"
+    b"endobj\n"
+    b"2 0 obj\n"
+    b"<< /Type /Pages /Kids [3 0 R] /Count 1 >>\n"
+    b"endobj\n"
+    b"3 0 obj\n"
+    b"<< /Type /Page /Parent 2 0 R /Resources << >> /MediaBox [0 0 612 792] /Contents 4 0 R >>\n"
+    b"endobj\n"
+    b"4 0 obj\n"
+    b"<< /Length 44 >>\n"
+    b"stream\n"
+    b"BT /F1 12 Tf 72 720 Td (Test PDF) Tj ET\n"
+    b"endstream\n"
+    b"endobj\n"
+    b"xref\n"
+    b"0 5\n"
+    b"0000000000 65535 f \n"
+    b"0000000009 00000 n \n"
+    b"0000000058 00000 n \n"
+    b"0000000117 00000 n \n"
+    b"0000000223 00000 n \n"
+    b"trailer\n"
+    b"<< /Size 5 /Root 1 0 R >>\n"
+    b"startxref\n"
+    b"317\n"
+    b"%%EOF"
+)
+
+CSV_BYTES = b"name,sales,units\nalice,100,3\nbob,200,5\ncarol,300,9\n"
+
+
+def pdf_bytes_with_text(text: str = "Quarterly Report") -> bytes:
+    """A real, text-extractable PDF (matplotlib is already a dependency)."""
+    import io
+
+    import matplotlib
+
+    matplotlib.use("Agg")
+    import matplotlib.pyplot as plt
+
+    figure = plt.figure()
+    figure.text(0.1, 0.5, text)
+    buffer = io.BytesIO()
+    figure.savefig(buffer, format="pdf")
+    plt.close(figure)
+    return buffer.getvalue()
+
+
+def docx_bytes_with_text(*paragraphs: str) -> bytes:
+    import io
+
+    import docx
+
+    document = docx.Document()
+    for paragraph in paragraphs:
+        document.add_paragraph(paragraph)
+    buffer = io.BytesIO()
+    document.save(buffer)
+    return buffer.getvalue()
+
+
+def make_company(name: str = "Acme", **kwargs) -> Company:
+    defaults = {
+        "state": "IL",
+        "zipcode": "60189",
+        "address": "1968 Greensboro Dr",
+    }
+    defaults.update(kwargs)
+    return Company.objects.create(name=name, **defaults)
+
+
+def make_user(
+    email: str = "test@test.com",
+    password: str | None = "testpass123",
+    company: Company | None = None,
+    **kwargs,
+):
+    user_model = get_user_model()
+    return user_model.objects.create_user(
+        username=kwargs.pop("username", email),
+        email=email,
+        password=password,
+        company=company,
+        **kwargs,
+    )
+
+
+def make_conversation(user=None, title: str = "Test Conversation", **kwargs):
+    return Conversation.objects.create(user=user, title=title, **kwargs)
+
+
+def make_prompt(
+    conversation, message: str = "hello", user_created: bool = True, **kwargs
+):
+    return Prompt.objects.create(
+        conversation=conversation,
+        message=message,
+        user_created=user_created,
+        **kwargs,
+    )
+
+
+def make_workspace(company, name: str = "Test Workspace") -> DocumentWorkspace:
+    return DocumentWorkspace.objects.create(company=company, name=name)
+
+
+def pdf_upload(name: str = "test.pdf") -> SimpleUploadedFile:
+    return SimpleUploadedFile(name, VALID_PDF_BYTES, content_type="application/pdf")
+
+
+def make_document(workspace, name: str = "test.pdf") -> Document:
+    return Document.objects.create(workspace=workspace, file=pdf_upload(name))

llm_be/chat_backend/tests/fakes.py

@@ -0,0 +1,44 @@
+"""Stand-ins for LangChain runnables so tests never reach Ollama."""
+
+from __future__ import annotations
+
+
+class FakeChain:
+    """Mimics the subset of the Runnable API the services use.
+
+    Records every payload it was invoked with, so tests can assert on the
+    prompt variables a service builds.
+    """
+
+    def __init__(
+        self,
+        response: str = "",
+        chunks: list[str] | None = None,
+        error: Exception | None = None,
+    ):
+        self.response = response
+        self.chunks = chunks if chunks is not None else [response]
+        self.error = error
+        self.calls: list[dict] = []
+
+    def invoke(self, payload):
+        self.calls.append(payload)
+        if self.error:
+            raise self.error
+        return self.response
+
+    async def ainvoke(self, payload):
+        return self.invoke(payload)
+
+    def stream(self, payload):
+        self.calls.append(payload)
+        if self.error:
+            raise self.error
+        yield from self.chunks
+
+    async def astream(self, payload):
+        self.calls.append(payload)
+        if self.error:
+            raise self.error
+        for chunk in self.chunks:
+            yield chunk

llm_be/chat_backend/tests/test_consumers.py

@@ -0,0 +1,373 @@
+from unittest import mock
+
+from asgiref.sync import sync_to_async
+from channels.testing import WebsocketCommunicator
+from django.test import TestCase, TransactionTestCase, override_settings
+from langchain_core.messages import AIMessage, HumanMessage
+from parameterized import parameterized
+
+from chat_backend import consumers, consumers_graph
+from chat_backend.models import Conversation, Prompt, PromptMetric
+from chat_backend.routing import websocket_urlpatterns
+from chat_backend.services.moderation_classifier import ModerationLabel
+from chat_backend.services.prompt_classifier.prompt_classifier import PromptType
+from llm_be.asgi import application
+
+from .factories import make_company, make_conversation, make_user, make_workspace
+
+
+class DatabaseHelperTestCase(TransactionTestCase):
+    """The ``database_sync_to_async`` helpers both consumers share.
+
+    ``TransactionTestCase`` is required: ``database_sync_to_async`` closes old
+    connections, which a class-wide atomic block (plain ``TestCase``) cannot survive.
+    """
+
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.workspace = make_workspace(self.company)
+        self.conversation = make_conversation(user=self.user)
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_create_conversation(self, _name, module):
+        conversation_id = await module.create_conversation(
+            "first prompt", self.user.email, "Weather Inquiry"
+        )
+
+        stored = await sync_to_async(
+            lambda: list(
+                Conversation.objects.filter(
+                    id=conversation_id, user=self.user
+                ).values_list("title", flat=True)
+            )
+        )()
+        self.assertEqual(stored, ["Weather Inquiry"])
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_get_workspace(self, _name, module):
+        workspace = await module.get_workspace(self.conversation.id)
+
+        self.assertEqual(workspace.id, self.workspace.id)
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_get_messages_stores_prompt_and_returns_history(self, _name, module):
+        messages, prompt_instance = await module.get_messages(
+            self.conversation.id, "what is the weather"
+        )
+
+        self.assertEqual([type(message) for message in messages], [HumanMessage])
+        self.assertEqual(messages[0].content, "what is the weather")
+        self.assertEqual(prompt_instance.conversation_id, self.conversation.id)
+        self.assertTrue(prompt_instance.user_created)
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_get_messages_maps_assistant_turns(self, _name, module):
+        await sync_to_async(Prompt.objects.create)(
+            conversation=self.conversation, message="earlier answer", user_created=False
+        )
+
+        messages, _prompt = await module.get_messages(self.conversation.id, "follow up")
+
+        self.assertEqual(
+            [type(message) for message in messages], [AIMessage, HumanMessage]
+        )
+
+    async def test_get_messages_stores_attachment_and_inlines_csv(self):
+        messages, prompt_instance = await consumers.get_messages(
+            self.conversation.id, "analyze this", b"name,sales\nalice,10\n", "csv"
+        )
+
+        self.assertEqual(prompt_instance.file_type, "csv")
+        self.assertIn("The file type is csv", messages[-1].content)
+        self.assertIn("alice,10", messages[-1].content)
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_get_conversation_file_returns_stored_attachment(self, _name, module):
+        await consumers.get_messages(
+            self.conversation.id, "analyze this", b"name,sales\nalice,10\n", "csv"
+        )
+
+        file_data, file_type = await module.get_conversation_file_async(
+            self.conversation.id
+        )
+
+        self.assertEqual(file_data, b"name,sales\nalice,10\n")
+        self.assertEqual(file_type, "csv")
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_get_conversation_file_without_attachment(self, _name, module):
+        file_data, file_type = await module.get_conversation_file_async(
+            self.conversation.id
+        )
+
+        self.assertIsNone(file_data)
+        self.assertIsNone(file_type)
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_save_generated_message(self, _name, module):
+        await module.save_generated_message(self.conversation.id, "generated answer")
+
+        stored = await sync_to_async(
+            lambda: list(
+                Prompt.objects.filter(conversation=self.conversation).values(
+                    "message", "user_created"
+                )
+            )
+        )()
+        self.assertEqual(
+            stored, [{"message": "generated answer", "user_created": False}]
+        )
+
+    @parameterized.expand([("websocket", consumers), ("langgraph", consumers_graph)])
+    async def test_prompt_metric_lifecycle(self, _name, module):
+        metric = await module.create_prompt_metric(
+            prompt_id=7,
+            prompt="what is the weather",
+            has_file=False,
+            file_type="",
+            model_name="llama3.2",
+            conversation_id=self.conversation.id,
+        )
+
+        self.assertEqual(metric.prompt_length, len("what is the weather"))
+        self.assertEqual(metric.event, "CREATED")
+
+        await module.finish_prompt_metric(metric, 120)
+
+        refreshed = await sync_to_async(PromptMetric.objects.get)(id=metric.id)
+        self.assertEqual(refreshed.event, "FINISHED")
+        self.assertEqual(refreshed.reponse_length, 120)
+        self.assertIsNotNone(refreshed.end_time)
+
+
+class GraphNodeTestCase(TransactionTestCase):
+    """The LangGraph nodes behind ``ws/conditional_chat/``."""
+
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.workspace = make_workspace(self.company)
+        self.conversation = make_conversation(user=self.user)
+        self.prompt = Prompt.objects.create(
+            conversation=self.conversation,
+            message="what is our policy?",
+            user_created=True,
+        )
+
+    def _state(self, **overrides):
+        state = {
+            "message": "what is our policy?",
+            "conversation_id": self.conversation.id,
+            "decoded_file": None,
+            "file_type": None,
+            "messages": [HumanMessage(content="what is our policy?")],
+            "prompt_instance": self.prompt,
+            "moderation_label": ModerationLabel.FINE,
+            "prompt_type": PromptType.GENERAL_CHAT,
+            "response_generator": None,
+            "error": None,
+            "model_name": "Turbo",
+        }
+        state.update(overrides)
+        return state
+
+    async def test_moderation_node_records_label(self):
+        with mock.patch.object(
+            consumers_graph.moderation_classifier,
+            "classify_async",
+            mock.AsyncMock(return_value=ModerationLabel.NSFW),
+        ):
+            result = await consumers_graph.moderation_node(self._state())
+
+        self.assertEqual(result["moderation_label"], ModerationLabel.NSFW)
+
+    async def test_classification_node_skipped_for_nsfw_prompts(self):
+        with mock.patch.object(
+            consumers_graph.PROMPT_CLASSIFIER, "classify_async", mock.AsyncMock()
+        ) as classify:
+            result = await consumers_graph.classification_node(
+                self._state(moderation_label=ModerationLabel.NSFW)
+            )
+
+        self.assertIsNone(result["prompt_type"])
+        classify.assert_not_called()
+
+    async def test_classification_node_uses_classifier(self):
+        with mock.patch.object(
+            consumers_graph.PROMPT_CLASSIFIER,
+            "classify_async",
+            mock.AsyncMock(return_value=PromptType.RAG),
+        ):
+            result = await consumers_graph.classification_node(self._state())
+
+        self.assertEqual(result["prompt_type"], PromptType.RAG)
+
+    async def test_attached_file_forces_data_analysis(self):
+        with mock.patch.object(
+            consumers_graph.PROMPT_CLASSIFIER,
+            "classify_async",
+            mock.AsyncMock(return_value=PromptType.GENERAL_CHAT),
+        ):
+            result = await consumers_graph.classification_node(
+                self._state(message="analyze this file", decoded_file=b"a,b\n1,2\n")
+            )
+
+        self.assertEqual(result["prompt_type"], PromptType.DATA_ANALYSIS)
+
+    async def test_attached_file_with_unrelated_prompt_stays_general_chat(self):
+        with mock.patch.object(
+            consumers_graph.PROMPT_CLASSIFIER,
+            "classify_async",
+            mock.AsyncMock(return_value=PromptType.SEARCH),
+        ):
+            result = await consumers_graph.classification_node(
+                self._state(message="tell me a joke", decoded_file=b"a,b\n1,2\n")
+            )
+
+        self.assertEqual(result["prompt_type"], PromptType.GENERAL_CHAT)
+
+    async def test_generation_node_refuses_nsfw_prompts(self):
+        result = await consumers_graph.generation_node(
+            self._state(moderation_label=ModerationLabel.NSFW)
+        )
+
+        payload = result["response_generator"]
+        self.assertEqual(payload["type"], "error")
+        self.assertIn("NSFW", payload["content"])
+
+    @override_settings(ALLOW_IMAGE_GENERATION=False)
+    async def test_generation_node_reports_disabled_image_generation(self):
+        result = await consumers_graph.generation_node(
+            self._state(prompt_type=PromptType.IMAGE_GENERATION)
+        )
+
+        self.assertEqual(
+            result["response_generator"],
+            {"type": "text", "content": "Image Generation is disabled."},
+        )
+
+    @override_settings(ALLOW_IMAGE_GENERATION=True)
+    async def test_generation_node_reports_unimplemented_image_generation(self):
+        result = await consumers_graph.generation_node(
+            self._state(prompt_type=PromptType.IMAGE_GENERATION)
+        )
+
+        self.assertIn(
+            "not supported at this time", result["response_generator"]["content"]
+        )
+
+    async def test_generation_node_requires_a_file_for_data_analysis(self):
+        result = await consumers_graph.generation_node(
+            self._state(prompt_type=PromptType.DATA_ANALYSIS)
+        )
+
+        self.assertEqual(
+            result["response_generator"],
+            {
+                "type": "text",
+                "content": "Please upload a file to perform data analysis.",
+            },
+        )
+
+    async def test_generation_node_streams_data_analysis(self):
+        with mock.patch.object(consumers_graph, "AsyncDataAnalysisService") as service:
+            service.return_value.generate_response.return_value = "generator"
+
+            result = await consumers_graph.generation_node(
+                self._state(
+                    prompt_type=PromptType.DATA_ANALYSIS,
+                    decoded_file=b"a,b\n1,2\n",
+                    file_type="csv",
+                )
+            )
+
+        self.assertEqual(result["response_generator"], "generator")
+        service.return_value.generate_response.assert_called_once_with(
+            self.prompt.message, b"a,b\n1,2\n", "csv"
+        )
+
+    async def test_generation_node_routes_rag_prompts_with_workspace(self):
+        with mock.patch.object(consumers_graph, "AsyncRAGService") as service:
+            service.return_value.generate_response.return_value = "generator"
+
+            result = await consumers_graph.generation_node(
+                self._state(prompt_type=PromptType.RAG)
+            )
+
+        self.assertEqual(result["response_generator"], "generator")
+        _args, _kwargs = service.return_value.generate_response.call_args
+        self.assertEqual(_args[2].id, self.workspace.id)
+
+    async def test_generation_node_defaults_to_general_chat(self):
+        with mock.patch.object(consumers_graph, "AsyncLLMService") as service:
+            service.return_value.generate_response.return_value = "generator"
+
+            result = await consumers_graph.generation_node(self._state())
+
+        self.assertEqual(result["response_generator"], "generator")
+        service.return_value.generate_response.assert_called_once()
+
+    @override_settings(ALLOW_INTERNET_ACCESS=True)
+    async def test_search_prompts_append_web_results(self):
+        state = self._state(prompt_type=PromptType.SEARCH)
+
+        with mock.patch.object(consumers_graph, "DuckDuckGoSearchRun") as search:
+            search.return_value.run.return_value = "top result"
+            with mock.patch.object(consumers_graph, "AsyncLLMService"):
+                await consumers_graph.generation_node(state)
+
+        self.assertIn("Search Results: top result", state["messages"][-1].content)
+
+    @override_settings(ALLOW_INTERNET_ACCESS=True)
+    async def test_fast_model_skips_web_search(self):
+        state = self._state(prompt_type=PromptType.SEARCH, model_name="FAST")
+
+        with mock.patch.object(consumers_graph, "DuckDuckGoSearchRun") as search:
+            with mock.patch.object(consumers_graph, "AsyncLLMService"):
+                await consumers_graph.generation_node(state)
+
+        search.assert_not_called()
+        self.assertEqual(len(state["messages"]), 1)
+
+    @override_settings(ALLOW_INTERNET_ACCESS=False)
+    async def test_search_is_skipped_when_internet_access_is_disabled(self):
+        state = self._state(prompt_type=PromptType.SEARCH)
+
+        with mock.patch.object(consumers_graph, "DuckDuckGoSearchRun") as search:
+            with mock.patch.object(consumers_graph, "AsyncLLMService"):
+                await consumers_graph.generation_node(state)
+
+        search.assert_not_called()
+
+    @override_settings(ALLOW_INTERNET_ACCESS=True)
+    async def test_search_failures_fall_back_to_plain_chat(self):
+        state = self._state(prompt_type=PromptType.SEARCH)
+
+        with mock.patch.object(consumers_graph, "DuckDuckGoSearchRun") as search:
+            search.return_value.run.side_effect = RuntimeError("ddg unreachable")
+            with mock.patch.object(consumers_graph, "AsyncLLMService") as service:
+                service.return_value.generate_response.return_value = "generator"
+                result = await consumers_graph.generation_node(state)
+
+        self.assertEqual(result["response_generator"], "generator")
+        self.assertEqual(len(state["messages"]), 1)
+
+
+class WebSocketRoutingTestCase(TransactionTestCase):
+    @parameterized.expand(
+        [("chat", "/ws/chat_again/"), ("conditional_chat", "/ws/conditional_chat/")]
+    )
+    async def test_route_accepts_connections(self, _name, path):
+        communicator = WebsocketCommunicator(application, path)
+
+        connected, _subprotocol = await communicator.connect()
+        self.assertTrue(connected)
+
+        await communicator.disconnect()
+
+    def test_only_the_two_chat_routes_are_registered(self):
+        self.assertEqual(
+            [str(route.pattern) for route in websocket_urlpatterns],
+            ["ws/chat_again/$", "ws/conditional_chat/$"],
+        )

llm_be/chat_backend/tests/test_live_ollama.py

@@ -0,0 +1,72 @@
+"""Opt-in checks that talk to a real Ollama instance.
+
+These are excluded from CI because they need a model server and their answers are
+not deterministic. Run them against a reachable ``OLLAMA_BASE_URL`` with::
+
+    RUN_LIVE_OLLAMA_TESTS=1 uv run python manage.py test \
+        chat_backend.tests.test_live_ollama
+"""
+
+import os
+import unittest
+
+from parameterized import parameterized
+
+from chat_backend.services.moderation_classifier import (
+    ModerationClassifier,
+    ModerationLabel,
+)
+from chat_backend.services.prompt_classifier.prompt_classifier import (
+    PromptClassifier,
+    PromptType,
+)
+from chat_backend.services.title_generator import TitleGenerator
+
+live_only = unittest.skipUnless(
+    os.environ.get("RUN_LIVE_OLLAMA_TESTS", "").lower() in {"1", "true", "yes"},
+    "Requires a live Ollama; set RUN_LIVE_OLLAMA_TESTS=1 to run",
+)
+
+
+@live_only
+class PromptClassifierLiveTestCase(unittest.TestCase):
+    def setUp(self):
+        self.service = PromptClassifier()
+
+    @parameterized.expand(
+        [
+            ["Tell me a joke", PromptType.GENERAL_CHAT],
+            ["Create an image of a dog for me", PromptType.IMAGE_GENERATION],
+            [
+                "highlight the features of the backyard playset if they were to choose "
+                "us and make the language more long form",
+                PromptType.GENERAL_CHAT,
+            ],
+        ]
+    )
+    def test_prompt_classification(self, prompt, expected_output):
+        self.assertEqual(self.service.classify(prompt), expected_output)
+
+
+@live_only
+class ModerationClassifierLiveTestCase(unittest.TestCase):
+    def setUp(self):
+        self.service = ModerationClassifier()
+
+    @parameterized.expand(
+        [
+            ["Write a love poem", ModerationLabel.FINE],
+            ["How to make a bomb", ModerationLabel.NSFW],
+        ]
+    )
+    def test_moderation(self, prompt, expected_output):
+        self.assertEqual(self.service.classify(prompt), expected_output)
+
+
+@live_only
+class TitleGeneratorLiveTestCase(unittest.TestCase):
+    def test_title_is_short(self):
+        title = TitleGenerator().generate("What's the weather today?")
+
+        self.assertTrue(title)
+        self.assertLessEqual(len(title.split()), 5)

llm_be/chat_backend/tests/test_models.py

@@ -0,0 +1,194 @@
+import datetime
+
+from django.core.files.uploadedfile import SimpleUploadedFile
+from django.test import TestCase
+from django.utils import timezone
+
+from chat_backend.models import (
+    Announcement,
+    Document,
+    Feedback,
+    LLMModels,
+    Prompt,
+    PromptMetric,
+    StoredFile,
+)
+
+from .factories import (
+    make_company,
+    make_conversation,
+    make_document,
+    make_prompt,
+    make_user,
+    make_workspace,
+)
+
+
+class TimeInfoBaseTestCase(TestCase):
+    def test_save_refreshes_last_modified(self):
+        company = make_company()
+        original = company.last_modified
+
+        company.name = "Renamed"
+        company.save()
+
+        company.refresh_from_db()
+        self.assertGreater(company.last_modified, original)
+        self.assertEqual(company.name, "Renamed")
+
+    def test_skip_last_modified_keeps_timestamp(self):
+        company = make_company()
+        original = company.last_modified
+
+        company.name = "Renamed"
+        company.save(skip_last_modified=True)
+
+        company.refresh_from_db()
+        self.assertEqual(company.last_modified, original)
+        self.assertEqual(company.name, "Renamed")
+
+    def test_update_fields_includes_last_modified(self):
+        conversation = make_conversation(title="Before")
+        original = conversation.last_modified
+
+        conversation.title = "After"
+        conversation.save(update_fields=["title"])
+
+        conversation.refresh_from_db()
+        self.assertEqual(conversation.title, "After")
+        self.assertGreater(conversation.last_modified, original)
+
+
+class CompanyAndUserTestCase(TestCase):
+    def test_company_str_is_name(self):
+        self.assertEqual(str(make_company("Globex")), "Globex")
+
+    def test_user_slug_is_generated_from_email(self):
+        user = make_user(email="person@example.com")
+        self.assertTrue(user.slug)
+        self.assertNotIn("@", user.slug)
+
+    def test_set_password_url_contains_slug(self):
+        user = make_user(email="person@example.com")
+        self.assertEqual(
+            user.get_set_password_url(),
+            f"https://chat.aimloperations.com/set_password?slug={user.slug}",
+        )
+
+    def test_user_defaults(self):
+        user = make_user()
+        self.assertFalse(user.is_company_manager)
+        self.assertFalse(user.deleted)
+        self.assertFalse(user.has_signed_tos)
+        self.assertTrue(user.conversation_order)
+
+
+class ConversationAndPromptTestCase(TestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.conversation = make_conversation(user=self.user, title="Weather Inquiry")
+
+    def test_conversation_str_and_user_email(self):
+        self.assertEqual(str(self.conversation), "Weather Inquiry")
+        self.assertEqual(self.conversation.get_user_email(), self.user.email)
+
+    def test_conversation_without_user_returns_empty_email(self):
+        self.assertEqual(make_conversation().get_user_email(), "")
+
+    def test_prompt_conversation_title(self):
+        prompt = make_prompt(self.conversation)
+        self.assertEqual(prompt.get_conversation_title(), "Weather Inquiry")
+
+    def test_prompt_without_conversation_returns_empty_title(self):
+        prompt = Prompt.objects.create(message="orphan", user_created=True)
+        self.assertEqual(prompt.get_conversation_title(), "")
+
+    def test_file_exists_false_without_file(self):
+        self.assertFalse(make_prompt(self.conversation).file_exists())
+
+    def test_file_exists_tracks_stored_blob(self):
+        prompt = make_prompt(self.conversation)
+        prompt.file.save("data.csv", SimpleUploadedFile("data.csv", b"a,b\n1,2\n"))
+
+        self.assertTrue(prompt.file_exists())
+        self.assertTrue(StoredFile.objects.filter(name=prompt.file.name).exists())
+
+        StoredFile.objects.filter(name=prompt.file.name).delete()
+        self.assertFalse(prompt.file_exists())
+
+
+class PromptMetricTestCase(TestCase):
+    def _metric(self, **kwargs):
+        defaults = {
+            "prompt_id": 1,
+            "conversation_id": 1,
+            "model_name": "llama3.2",
+            "start_time": timezone.now(),
+            "prompt_length": 10,
+            "has_file": False,
+        }
+        defaults.update(kwargs)
+        return PromptMetric.objects.create(**defaults)
+
+    def test_duration_is_zero_until_finished(self):
+        self.assertEqual(self._metric().get_duration(), 0)
+
+    def test_duration_in_seconds(self):
+        start = timezone.now()
+        metric = self._metric(
+            start_time=start, end_time=start + datetime.timedelta(seconds=42)
+        )
+        self.assertEqual(metric.get_duration(), 42)
+
+    def test_default_event_is_created(self):
+        self.assertEqual(self._metric().event, "CREATED")
+
+
+class FeedbackTestCase(TestCase):
+    def test_get_user_email(self):
+        user = make_user()
+        feedback = Feedback.objects.create(title="t", text="body", user=user)
+        self.assertEqual(feedback.get_user_email(), user.email)
+
+    def test_get_user_email_without_user(self):
+        feedback = Feedback.objects.create(title="t", text="body")
+        self.assertEqual(feedback.get_user_email(), "")
+
+    def test_defaults(self):
+        feedback = Feedback.objects.create(text="body")
+        self.assertEqual(feedback.status, "SUBMITTED")
+        self.assertEqual(feedback.category, "NOT_DEFINED")
+
+
+class DocumentModelTestCase(TestCase):
+    def setUp(self):
+        self.workspace = make_workspace(make_company())
+
+    def test_document_defaults_and_storage(self):
+        document = make_document(self.workspace)
+
+        self.assertFalse(document.processed)
+        self.assertFalse(document.active)
+        self.assertIsNotNone(document.uploaded_at)
+        self.assertTrue(document.file.name.startswith("documents/"))
+        self.assertTrue(StoredFile.objects.filter(name=document.file.name).exists())
+
+    def test_deleting_workspace_cascades_to_documents(self):
+        make_document(self.workspace)
+        self.workspace.delete()
+        self.assertEqual(Document.objects.count(), 0)
+
+
+class MiscModelTestCase(TestCase):
+    def test_stored_file_str_is_name(self):
+        stored = StoredFile.objects.create(name="documents/a.txt", content=b"x", size=1)
+        self.assertEqual(str(stored), "documents/a.txt")
+
+    def test_announcement_default_status(self):
+        announcement = Announcement.objects.create(message="hello")
+        self.assertEqual(announcement.status, Announcement.Status.default)
+
+    def test_llm_model_fields(self):
+        model = LLMModels.objects.create(name="llama3.2", port=11434, description="d")
+        self.assertEqual(model.port, 11434)

llm_be/chat_backend/tests/test_serializers.py

@@ -0,0 +1,188 @@
+from django.test import TestCase
+
+from chat_backend.models import Prompt
+from chat_backend.serializers import (
+    BasicUserSerializer,
+    ConversationSerializer,
+    CustomUserSerializer,
+    DocumentSerializer,
+    DocumentWorkspaceSerializer,
+    FeedbackSerializer,
+    MyTokenObtainPairSerializer,
+    PromptSerializer,
+)
+
+from .factories import (
+    make_company,
+    make_conversation,
+    make_document,
+    make_user,
+    make_workspace,
+    pdf_upload,
+)
+
+
+class TokenSerializerTestCase(TestCase):
+    def test_token_carries_company_claim(self):
+        user = make_user(company=make_company())
+
+        token = MyTokenObtainPairSerializer.get_token(user)
+
+        self.assertEqual(token["company"], "something here")
+        self.assertEqual(str(token["user_id"]), str(user.id))
+
+
+class ConversationSerializerTestCase(TestCase):
+    def test_exposes_only_whitelisted_fields(self):
+        conversation = make_conversation(user=make_user(), title="Weather Inquiry")
+
+        data = ConversationSerializer(conversation).data
+
+        self.assertEqual(set(data.keys()), {"title", "created", "last_modified", "id"})
+        self.assertEqual(data["title"], "Weather Inquiry")
+
+
+class PromptSerializerTestCase(TestCase):
+    def test_valid_payload_creates_prompt(self):
+        serializer = PromptSerializer(data={"message": "hi", "user_created": True})
+
+        self.assertTrue(serializer.is_valid(), serializer.errors)
+        prompt = serializer.save()
+
+        self.assertEqual(Prompt.objects.count(), 1)
+        self.assertTrue(prompt.user_created)
+
+    def test_message_is_required(self):
+        serializer = PromptSerializer(data={"user_created": True})
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("message", serializer.errors)
+
+    def test_user_created_is_required(self):
+        serializer = PromptSerializer(data={"message": "hi"})
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("user_created", serializer.errors)
+
+    def test_file_is_not_serialized(self):
+        prompt = Prompt.objects.create(message="hi", user_created=True)
+        self.assertNotIn("file", PromptSerializer(prompt).data)
+
+
+class UserSerializerTestCase(TestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company("Globex"), first_name="Ada")
+
+    def test_custom_user_serializer_hides_password(self):
+        data = CustomUserSerializer(self.user).data
+
+        self.assertEqual(data["email"], self.user.email)
+        self.assertEqual(data["company"]["name"], "Globex")
+        self.assertTrue(data["has_usable_password"])
+        self.assertNotIn("password", data)
+
+    def test_custom_user_serializer_rejects_short_password(self):
+        serializer = CustomUserSerializer(
+            data={
+                "username": "someone",
+                "email": "someone@example.com",
+                "password": "short",
+                "has_usable_password": True,
+                "company": {
+                    "name": "Acme",
+                    "state": "IL",
+                    "zipcode": "60189",
+                    "address": "1 Main St",
+                },
+            }
+        )
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("password", serializer.errors)
+
+    def test_custom_user_serializer_rejects_invalid_email(self):
+        serializer = CustomUserSerializer(
+            data={
+                "username": "someone",
+                "email": "not-an-email",
+                "password": "longenough",
+            }
+        )
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("email", serializer.errors)
+
+    def test_basic_user_serializer_field_set(self):
+        data = BasicUserSerializer(self.user).data
+
+        self.assertEqual(
+            set(data.keys()),
+            {
+                "email",
+                "first_name",
+                "last_name",
+                "is_active",
+                "has_usable_password",
+                "is_company_manager",
+                "has_signed_tos",
+            },
+        )
+        self.assertEqual(data["first_name"], "Ada")
+
+
+class FeedbackSerializerTestCase(TestCase):
+    def test_text_is_required(self):
+        serializer = FeedbackSerializer(data={"title": "no body"})
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("text", serializer.errors)
+
+    def test_defaults_applied_on_save(self):
+        serializer = FeedbackSerializer(data={"title": "t", "text": "body"})
+
+        self.assertTrue(serializer.is_valid(), serializer.errors)
+        feedback = serializer.save()
+
+        self.assertEqual(feedback.status, "SUBMITTED")
+        self.assertEqual(feedback.category, "NOT_DEFINED")
+
+
+class DocumentSerializerTestCase(TestCase):
+    def setUp(self):
+        self.workspace = make_workspace(make_company())
+
+    def test_workspace_serializer_read_only_fields(self):
+        serializer = DocumentWorkspaceSerializer(
+            data={"id": 999, "name": "New Workspace"}
+        )
+
+        self.assertTrue(serializer.is_valid(), serializer.errors)
+        self.assertEqual(set(serializer.validated_data.keys()), {"name"})
+
+    def test_workspace_name_is_required(self):
+        serializer = DocumentWorkspaceSerializer(data={})
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("name", serializer.errors)
+
+    def test_document_serializer_reports_stored_file_url(self):
+        document = make_document(self.workspace)
+
+        data = DocumentSerializer(document).data
+
+        self.assertEqual(data["workspace"], self.workspace.id)
+        self.assertIn("test", data["file"])
+        self.assertIn("pdf", data["file"])
+        self.assertFalse(data["processed"])
+
+    def test_document_serializer_ignores_read_only_processed(self):
+        serializer = DocumentSerializer(
+            data={
+                "workspace": self.workspace.id,
+                "file": pdf_upload(),
+                "processed": True,
+            }
+        )
+
+        self.assertTrue(serializer.is_valid(), serializer.errors)
+        self.assertNotIn("processed", serializer.validated_data)

llm_be/chat_backend/tests/test_services_classifiers.py

@@ -0,0 +1,221 @@
+from django.test import SimpleTestCase, override_settings
+from parameterized import parameterized
+
+from chat_backend.services.moderation_classifier import (
+    ModerationClassifier,
+    ModerationLabel,
+)
+from chat_backend.services.prompt_classifier.prompt_classifier import (
+    PromptClassifier,
+    PromptType,
+)
+from chat_backend.services.title_generator import TitleGenerator
+
+from .fakes import FakeChain
+
+
+class PromptClassifierQuickCheckTestCase(SimpleTestCase):
+    def setUp(self):
+        self.classifier = PromptClassifier()
+
+    @parameterized.expand(
+        [
+            (
+                "data_analysis",
+                "Please read this document and make me an index",
+                PromptType.DATA_ANALYSIS,
+            ),
+            (
+                "csv",
+                "Based on this csv, who bought the most?",
+                PromptType.DATA_ANALYSIS,
+            ),
+            ("rag", "What does the uploaded pdf say about revenue?", PromptType.RAG),
+            (
+                "search_news",
+                "What is the latest news on the merger?",
+                PromptType.SEARCH,
+            ),
+            ("search_sports", "Who won the game last night?", PromptType.SEARCH),
+        ]
+    )
+    def test_rule_based_hits(self, _name, prompt, expected):
+        self.assertEqual(self.classifier._quick_check(prompt), expected)
+
+    def test_plain_chat_falls_through_to_the_llm(self):
+        self.assertIsNone(self.classifier._quick_check("Tell me a joke"))
+
+    @override_settings(ALLOW_IMAGE_GENERATION=True)
+    def test_image_requests_when_feature_enabled(self):
+        self.assertEqual(
+            self.classifier._quick_check("Generate an image of a dog"),
+            PromptType.IMAGE_GENERATION,
+        )
+
+    @override_settings(ALLOW_IMAGE_GENERATION=False)
+    def test_image_requests_downgrade_when_feature_disabled(self):
+        self.assertEqual(
+            self.classifier._quick_check("Generate an image of a dog"),
+            PromptType.GENERAL_CHAT,
+        )
+
+
+class PromptClassifierParseResponseTestCase(SimpleTestCase):
+    def setUp(self):
+        self.classifier = PromptClassifier()
+
+    @parameterized.expand(
+        [
+            ("exact", "RAG", PromptType.RAG),
+            ("lowercase", "general_chat", PromptType.GENERAL_CHAT),
+            ("missing_underscore", "GENERALCHAT", PromptType.GENERAL_CHAT),
+            ("padded", "  SEARCH  ", PromptType.SEARCH),
+            ("verbose", "The answer is DATA_ANALYSIS", PromptType.DATA_ANALYSIS),
+            ("nonsense", "banana", PromptType.UNKNOWN),
+        ]
+    )
+    def test_parses_llm_labels(self, _name, response, expected):
+        self.assertEqual(self.classifier._parse_response(response), expected)
+
+    @override_settings(ALLOW_IMAGE_GENERATION=False)
+    def test_image_generation_is_downgraded_when_disabled(self):
+        self.assertEqual(
+            self.classifier._parse_response("IMAGE_GENERATION"), PromptType.GENERAL_CHAT
+        )
+
+    @override_settings(ALLOW_IMAGE_GENERATION=True)
+    def test_image_generation_survives_when_enabled(self):
+        self.assertEqual(
+            self.classifier._parse_response("IMAGE_GENERATION"),
+            PromptType.IMAGE_GENERATION,
+        )
+
+
+class PromptClassifierClassifyTestCase(SimpleTestCase):
+    def setUp(self):
+        self.classifier = PromptClassifier()
+
+    def test_classify_uses_llm_when_no_rule_matches(self):
+        self.classifier.chain = FakeChain("GENERAL_CHAT")
+
+        result = self.classifier.classify("Tell me a joke")
+
+        self.assertEqual(result, PromptType.GENERAL_CHAT)
+        self.assertEqual(self.classifier.chain.calls, [{"prompt": "Tell me a joke"}])
+
+    def test_classify_skips_llm_when_rule_matches(self):
+        self.classifier.chain = FakeChain("GENERAL_CHAT")
+
+        result = self.classifier.classify("What is the latest news?")
+
+        self.assertEqual(result, PromptType.SEARCH)
+        self.assertEqual(self.classifier.chain.calls, [])
+
+    def test_classify_returns_unknown_when_llm_fails(self):
+        self.classifier.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(self.classifier.classify("Tell me a joke"), PromptType.UNKNOWN)
+
+    async def test_classify_async_uses_llm(self):
+        self.classifier.chain = FakeChain("RAG")
+
+        self.assertEqual(
+            await self.classifier.classify_async("Tell me a joke"), PromptType.RAG
+        )
+
+    async def test_classify_async_returns_unknown_when_llm_fails(self):
+        self.classifier.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(
+            await self.classifier.classify_async("Tell me a joke"), PromptType.UNKNOWN
+        )
+
+
+class ModerationClassifierTestCase(SimpleTestCase):
+    def setUp(self):
+        self.classifier = ModerationClassifier()
+
+    @parameterized.expand(
+        [
+            ("nsfw", "NSFW", ModerationLabel.NSFW),
+            ("fine", "FINE", ModerationLabel.FINE),
+            ("verbose_nsfw", "This prompt is NSFW because...", ModerationLabel.NSFW),
+            ("unclear_defaults_to_fine", "not sure", ModerationLabel.FINE),
+        ]
+    )
+    def test_parses_labels(self, _name, response, expected):
+        self.assertEqual(self.classifier._parse_response(response), expected)
+
+    def test_classify_normalizes_llm_output(self):
+        self.classifier.chain = FakeChain("  fine \n")
+
+        self.assertEqual(
+            self.classifier.classify("Python tutorial"), ModerationLabel.FINE
+        )
+
+    def test_classify_fails_safe_to_nsfw(self):
+        self.classifier.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(
+            self.classifier.classify("Python tutorial"), ModerationLabel.NSFW
+        )
+
+    async def test_classify_async_fails_safe_to_nsfw(self):
+        self.classifier.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(
+            await self.classifier.classify_async("Python tutorial"),
+            ModerationLabel.NSFW,
+        )
+
+
+class TitleGeneratorTestCase(SimpleTestCase):
+    def setUp(self):
+        self.generator = TitleGenerator()
+
+    @parameterized.expand(
+        [
+            ("strips_quotes", '"Weather Inquiry"', "Weather Inquiry"),
+            ("strips_punctuation", "weather inquiry!", "Weather Inquiry"),
+            (
+                "title_cases",
+                "quantum computing explanation",
+                "Quantum Computing Explanation",
+            ),
+        ]
+    )
+    def test_cleans_llm_output(self, _name, response, expected):
+        self.assertEqual(self.generator._clean_response(response), expected)
+
+    def test_titles_are_capped_at_fifty_characters(self):
+        cleaned = self.generator._clean_response("word " * 40)
+
+        self.assertEqual(len(cleaned), 50)
+
+    def test_generate_returns_cleaned_title(self):
+        self.generator.chain = FakeChain('"dragon image generation"')
+
+        self.assertEqual(
+            self.generator.generate("Generate an image of a dragon"),
+            "Dragon Image Generation",
+        )
+
+    def test_generate_falls_back_on_error(self):
+        self.generator.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(self.generator.generate("anything"), "Conversation")
+
+    async def test_generate_async_falls_back_on_error(self):
+        self.generator.chain = FakeChain(error=RuntimeError("ollama down"))
+
+        self.assertEqual(
+            await self.generator.generate_async("anything"), "Conversation"
+        )
+
+    async def test_generate_async_returns_cleaned_title(self):
+        self.generator.chain = FakeChain("weather inquiry")
+
+        self.assertEqual(
+            await self.generator.generate_async("What's the weather?"),
+            "Weather Inquiry",
+        )

llm_be/chat_backend/tests/test_services_data_analysis.py

@@ -0,0 +1,173 @@
+import base64
+import json
+
+import pandas as pd
+from django.test import SimpleTestCase
+
+from chat_backend.services.data_analysis_service import AsyncDataAnalysisService
+
+from .factories import CSV_BYTES, docx_bytes_with_text, pdf_bytes_with_text
+from .fakes import FakeChain
+
+PNG_MAGIC = b"\x89PNG\r\n\x1a\n"
+
+
+def frame():
+    return pd.DataFrame(
+        {"name": ["a", "b", "c"], "sales": [1, 2, 3], "units": [4, 5, 6]}
+    )
+
+
+async def collect(generator):
+    return [chunk async for chunk in generator]
+
+
+class DataFrameSummaryTestCase(SimpleTestCase):
+    def setUp(self):
+        self.service = AsyncDataAnalysisService()
+
+    def test_summary_reports_shape_columns_and_sample(self):
+        summary = self.service._get_dataframe_summary(frame())
+
+        self.assertIn("DataFrame has 3 rows and 3 columns.", summary)
+        self.assertIn("Descriptive Statistics", summary)
+        self.assertIn("sales", summary)
+        self.assertIn("units", summary)
+
+
+class DocumentReadersTestCase(SimpleTestCase):
+    def setUp(self):
+        self.service = AsyncDataAnalysisService()
+
+    def test_reads_docx_paragraphs(self):
+        text = self.service._read_docx(
+            docx_bytes_with_text("first line", "second line")
+        )
+
+        self.assertEqual(text, "first line\nsecond line")
+
+    def test_reads_pdf_text(self):
+        text = self.service._read_pdf(pdf_bytes_with_text("Quarterly Report"))
+
+        self.assertIn("Quarterly Report", text)
+
+
+class PlotGenerationTestCase(SimpleTestCase):
+    def setUp(self):
+        self.service = AsyncDataAnalysisService()
+
+    def test_uses_columns_named_in_the_query(self):
+        encoded = self.service._generate_plot("plot sales vs units", frame())
+
+        self.assertTrue(base64.b64decode(encoded).startswith(PNG_MAGIC))
+
+    def test_falls_back_to_first_two_numeric_columns(self):
+        encoded = self.service._generate_plot("graph the data please", frame())
+
+        self.assertTrue(base64.b64decode(encoded).startswith(PNG_MAGIC))
+
+    def test_requires_two_numeric_columns(self):
+        single_numeric = pd.DataFrame({"name": ["a", "b"], "sales": [1, 2]})
+
+        with self.assertRaises(ValueError):
+            self.service._generate_plot("plot it", single_numeric)
+
+
+class GenerateResponseTestCase(SimpleTestCase):
+    def setUp(self):
+        self.service = AsyncDataAnalysisService()
+        self.service.analysis_chain = FakeChain(chunks=["Sales ", "are up."])
+
+    async def test_csv_question_is_answered_from_the_summary(self):
+        chunks = await collect(
+            self.service.generate_response(
+                "What are total sales?", CSV_BYTES, "text/csv"
+            )
+        )
+
+        self.assertEqual("".join(chunks), "Sales are up.")
+        payload = self.service.analysis_chain.calls[0]
+        self.assertEqual(payload["query"], "What are total sales?")
+        self.assertIn("DataFrame has 3 rows", payload["data_summary"])
+
+    async def test_plot_request_returns_png_payload(self):
+        chunks = await collect(
+            self.service.generate_response("plot sales vs units", CSV_BYTES, "text/csv")
+        )
+
+        self.assertEqual(len(chunks), 1)
+        payload = json.loads(chunks[0])
+        self.assertEqual(payload["type"], "plot")
+        self.assertEqual(payload["format"], "png")
+        self.assertTrue(base64.b64decode(payload["image"]).startswith(PNG_MAGIC))
+        self.assertEqual(self.service.analysis_chain.calls, [])
+
+    async def test_plot_request_without_numeric_columns_reports_error(self):
+        csv = b"name,city\nalice,chicago\nbob,denver\n"
+
+        chunks = await collect(
+            self.service.generate_response("plot it", csv, "text/csv")
+        )
+
+        payload = json.loads(chunks[0])
+        self.assertEqual(payload["type"], "error")
+        self.assertIn("two numerical columns", payload["content"])
+
+    async def test_docx_question_uses_document_text(self):
+        docx = docx_bytes_with_text("Revenue grew by 10%")
+
+        await collect(
+            self.service.generate_response(
+                "Summarize this",
+                docx,
+                "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+            )
+        )
+
+        self.assertIn(
+            "Revenue grew by 10%", self.service.analysis_chain.calls[0]["data_summary"]
+        )
+
+    async def test_pdf_question_uses_document_text(self):
+        await collect(
+            self.service.generate_response(
+                "Summarize this",
+                pdf_bytes_with_text("Quarterly Report"),
+                "application/pdf",
+            )
+        )
+
+        self.assertIn(
+            "Quarterly Report", self.service.analysis_chain.calls[0]["data_summary"]
+        )
+
+    async def test_unsupported_file_type_is_reported(self):
+        chunks = await collect(
+            self.service.generate_response("What is this?", b"\x00\x01", "image/png")
+        )
+
+        payload = json.loads(chunks[0])
+        self.assertEqual(payload["type"], "error")
+        self.assertIn("Unsupported file type", payload["content"])
+
+    async def test_unreadable_file_is_reported(self):
+        chunks = await collect(
+            self.service.generate_response("What is this?", b"", "text/csv")
+        )
+
+        payload = json.loads(chunks[0])
+        self.assertEqual(payload["type"], "error")
+        self.assertIn("An error occurred", payload["content"])
+
+    async def test_llm_failure_is_reported(self):
+        self.service.analysis_chain = FakeChain(error=RuntimeError("ollama down"))
+
+        chunks = await collect(
+            self.service.generate_response(
+                "What are total sales?", CSV_BYTES, "text/csv"
+            )
+        )
+
+        payload = json.loads(chunks[0])
+        self.assertEqual(payload["type"], "error")
+        self.assertIn("ollama down", payload["content"])

llm_be/chat_backend/tests/test_services_llm.py

@@ -0,0 +1,64 @@
+from django.test import SimpleTestCase
+from langchain_core.messages import AIMessage, HumanMessage
+
+from chat_backend.services.llm_service import AsyncLLMService, SyncLLMService
+
+from .fakes import FakeChain
+
+
+def conversation(length: int):
+    messages = []
+    for index in range(length):
+        messages.append(HumanMessage(content=f"question {index}"))
+        messages.append(AIMessage(content=f"answer {index}"))
+    return messages
+
+
+class AsyncLLMServiceTestCase(SimpleTestCase):
+    def setUp(self):
+        self.service = AsyncLLMService()
+
+    async def test_format_history_labels_speakers(self):
+        history = await self.service._format_history(
+            [HumanMessage(content="hello"), AIMessage(content="hi")]
+        )
+
+        self.assertEqual(history, "User: hello\nAI: hi")
+
+    async def test_generate_response_streams_chunks(self):
+        self.service.conversation_chain = FakeChain(chunks=["Hel", "lo!"])
+
+        chunks = [
+            chunk
+            async for chunk in self.service.generate_response(
+                conversation(1), "hello", conversation_id=1
+            )
+        ]
+
+        self.assertEqual("".join(chunks), "Hello!")
+
+    async def test_generate_response_sends_full_and_recent_history(self):
+        self.service.conversation_chain = FakeChain(chunks=["ok"])
+        messages = conversation(4)  # 8 messages
+
+        async for _ in self.service.generate_response(messages, "latest", 1):
+            pass
+
+        payload = self.service.conversation_chain.calls[0]
+        self.assertEqual(payload["query"], "latest")
+        self.assertEqual(len(payload["conversation"].splitlines()), 8)
+        self.assertEqual(len(payload["recent_conversation"].splitlines()), 6)
+        self.assertTrue(payload["recent_conversation"].endswith("AI: answer 3"))
+
+
+class SyncLLMServiceTestCase(SimpleTestCase):
+    def test_generate_response_streams_chunks(self):
+        service = SyncLLMService()
+        service.conversation_chain = FakeChain(chunks=["one ", "two"])
+
+        chunks = list(service.generate_response(conversation=None, query="hello"))
+
+        self.assertEqual("".join(chunks), "one two")
+        self.assertEqual(
+            service.conversation_chain.calls, [{"query": "hello", "conversation": None}]
+        )

llm_be/chat_backend/tests/test_services_rag.py

@@ -0,0 +1,279 @@
+import os
+import shutil
+import tempfile
+from unittest import mock
+
+from asgiref.sync import sync_to_async
+from django.core.files.uploadedfile import SimpleUploadedFile
+from django.test import TransactionTestCase
+from langchain_community.document_loaders import (
+    Docx2txtLoader,
+    PyPDFLoader,
+    TextLoader,
+    UnstructuredFileLoader,
+)
+from langchain_core.documents import Document as LangDocument
+from langchain_core.messages import AIMessage, HumanMessage
+from parameterized import parameterized
+
+from chat_backend.models import Document
+from chat_backend.services.rag_services import (
+    AsyncRAGService,
+    RAGService,
+    SyncRAGService,
+    get_documents,
+)
+
+from .factories import make_company, make_workspace
+from .fakes import FakeChain
+
+
+def reset_singletons():
+    for service in (RAGService, SyncRAGService, AsyncRAGService):
+        service._instance = None
+
+
+class RAGServiceTestCase(TransactionTestCase):
+    """Exercises the RAG plumbing with Chroma and Ollama replaced by mocks.
+
+    ``TransactionTestCase`` is required: the async helpers close old connections,
+    which a class-wide atomic block (plain ``TestCase``) cannot survive.
+    """
+
+    def setUp(self):
+        self.persist_dir = tempfile.mkdtemp()
+        self.addCleanup(shutil.rmtree, self.persist_dir, True)
+
+        self.chroma = self._patch("chat_backend.services.rag_services.Chroma")
+        # A fresh mock per call so re-initialisation is observable.
+        self.chroma.side_effect = lambda *args, **kwargs: mock.MagicMock(
+            name="vector_store"
+        )
+        self.embeddings = self._patch(
+            "chat_backend.services.rag_services.OllamaEmbeddings"
+        )
+        self._patch("chat_backend.services.base_service.OllamaLLM")
+
+        overrides = self.settings(CHROMA_PERSIST_DIRECTORY=self.persist_dir)
+        overrides.enable()
+        self.addCleanup(overrides.disable)
+
+        reset_singletons()
+        self.addCleanup(reset_singletons)
+        self.service = AsyncRAGService()
+
+        self.workspace = make_workspace(make_company())
+
+    def _patch(self, target):
+        patcher = mock.patch(target)
+        self.addCleanup(patcher.stop)
+        return patcher.start()
+
+    def _text_document(self, name="notes.txt", body=b"hello world"):
+        return Document.objects.create(
+            workspace=self.workspace, file=SimpleUploadedFile(name, body)
+        )
+
+    def test_vector_store_uses_configured_persist_directory(self):
+        self.assertTrue(os.path.isdir(self.persist_dir))
+        self.chroma.assert_called_with(
+            embedding_function=self.embeddings.return_value,
+            persist_directory=self.persist_dir,
+        )
+
+    @parameterized.expand(
+        [
+            ("pdf", "/tmp/a.pdf", PyPDFLoader),
+            ("docx", "/tmp/a.docx", Docx2txtLoader),
+            ("txt_uppercase", "/tmp/A.TXT", TextLoader),
+            ("unknown", "/tmp/a.pptx", UnstructuredFileLoader),
+        ]
+    )
+    def test_loader_selection(self, _name, path, expected_loader):
+        self.assertIs(self.service._get_file_loader(path), expected_loader)
+
+    @parameterized.expand(
+        [
+            ("spaces_kept", "quarterly report.pdf", "quarterly report.pdf"),
+            ("parens", "report (final).pdf", "report _final_.pdf"),
+            ("path_separators", "../../etc/passwd", ".._.._etc_passwd"),
+        ]
+    )
+    def test_sanitize_filename(self, _name, given, expected):
+        self.assertEqual(self.service._sanitize_filename(given), expected)
+
+    def test_materialize_file_field_writes_a_temp_copy(self):
+        document = self._text_document(body=b"stored in postgres")
+
+        path = self.service._materialize_file_field(document.file)
+        try:
+            self.assertTrue(path.endswith(".txt"))
+            with open(path, "rb") as handle:
+                self.assertEqual(handle.read(), b"stored in postgres")
+        finally:
+            os.unlink(path)
+
+    def test_load_and_split_documents_merges_metadata(self):
+        path = self._temp_text_file(b"sentence one. sentence two.")
+
+        chunks = self.service._load_and_split_documents(path, {"workspace_id": 7})
+
+        self.assertTrue(chunks)
+        self.assertEqual(chunks[0].metadata["workspace_id"], 7)
+        self.assertIn("sentence one", chunks[0].page_content)
+
+    def _temp_text_file(self, body: bytes) -> str:
+        with tempfile.NamedTemporaryFile(suffix=".txt", delete=False) as handle:
+            handle.write(body)
+        self.addCleanup(lambda: os.path.exists(handle.name) and os.unlink(handle.name))
+        return handle.name
+
+    def test_ingest_documents_adds_chunks_with_metadata(self):
+        document = self._text_document(body=b"ingest me")
+
+        self.service.ingest_documents()
+
+        self.service.vector_store.add_documents.assert_called_once()
+        added = self.service.vector_store.add_documents.call_args[0][0]
+        self.assertIn("ingest me", added[0].page_content)
+        self.assertEqual(added[0].metadata["workspace_id"], self.workspace.id)
+        self.assertEqual(added[0].metadata["document_id"], document.id)
+
+    def test_ingest_documents_deletes_the_materialized_temp_file(self):
+        self._text_document(body=b"ingest me")
+        materialized = []
+        original = self.service._materialize_file_field
+
+        def spy(file_field):
+            path = original(file_field)
+            materialized.append(path)
+            return path
+
+        with mock.patch.object(
+            self.service, "_materialize_file_field", side_effect=spy
+        ):
+            self.service.ingest_documents()
+
+        self.assertTrue(materialized)
+        self.assertFalse([path for path in materialized if os.path.exists(path)])
+
+    def test_ingest_documents_can_be_scoped_to_a_workspace(self):
+        other_workspace = make_workspace(make_company("Other"), name="Other")
+        Document.objects.create(
+            workspace=other_workspace, file=SimpleUploadedFile("other.txt", b"other")
+        )
+        self._text_document(body=b"mine")
+
+        self.service.ingest_documents(workspace=self.workspace)
+
+        self.assertEqual(self.service.vector_store.add_documents.call_count, 1)
+        added = self.service.vector_store.add_documents.call_args[0][0]
+        self.assertIn("mine", added[0].page_content)
+
+    def test_add_files_to_store_reports_processed_files(self):
+        path = self._temp_text_file(b"upload body")
+
+        results = self.service.add_files_to_store(
+            [(path, "upload.txt", self.workspace.id)], workspace_id=self.workspace.id
+        )
+
+        self.assertEqual(results["failed_files"], [])
+        self.assertEqual(results["total_added"], 1)
+        self.assertEqual(
+            results["processed_files"],
+            [{"filename": "upload.txt", "document_count": 1}],
+        )
+        added = self.service.vector_store.add_documents.call_args[0][0]
+        self.assertEqual(added[0].metadata["source"], "upload.txt")
+        self.assertEqual(added[0].metadata["workspace_id"], self.workspace.id)
+
+    def test_add_files_to_store_accepts_database_backed_file_fields(self):
+        document = self._text_document(body=b"from the database")
+
+        results = self.service.add_files_to_store(
+            [(document.file, document.file.name, self.workspace.id)],
+            workspace_id=self.workspace.id,
+        )
+
+        self.assertEqual(results["total_added"], 1)
+        self.assertEqual(results["failed_files"], [])
+
+    def test_add_files_to_store_records_failures(self):
+        results = self.service.add_files_to_store(
+            [("/tmp/does-not-exist.txt", "missing.txt", self.workspace.id)],
+            workspace_id=self.workspace.id,
+        )
+
+        self.assertEqual(results["total_added"], 0)
+        self.assertEqual(results["failed_files"][0]["filename"], "missing.txt")
+
+    def test_clear_vector_store_recreates_the_collection(self):
+        first_store = self.service.vector_store
+
+        self.service.clear_vector_store()
+
+        first_store.delete_collection.assert_called_once()
+        self.assertIsNot(self.service.vector_store, first_store)
+
+    async def test_search_documents_filters_by_workspace(self):
+        retriever = self.service.vector_store.as_retriever.return_value
+        retriever.aget_relevant_documents = mock.AsyncMock(
+            return_value=[LangDocument(page_content="chunk")]
+        )
+
+        docs = await self.service.search_documents("revenue", self.workspace, k=2)
+
+        self.assertEqual([doc.page_content for doc in docs], ["chunk"])
+        self.service.vector_store.as_retriever.assert_called_with(
+            search_type="mmr",
+            search_kwargs={"k": 2, "filter": {"workspace_id": self.workspace.id}},
+        )
+
+    async def test_search_documents_without_workspace_has_no_filter(self):
+        retriever = self.service.vector_store.as_retriever.return_value
+        retriever.aget_relevant_documents = mock.AsyncMock(return_value=[])
+
+        await self.service.search_documents("revenue")
+
+        self.service.vector_store.as_retriever.assert_called_with(
+            search_type="mmr", search_kwargs={"k": 4, "filter": None}
+        )
+
+    async def test_format_history_labels_speakers(self):
+        history = await self.service._format_history(
+            [
+                HumanMessage(content="what is our policy?"),
+                AIMessage(content="here it is"),
+            ]
+        )
+
+        self.assertEqual(history, "User: what is our policy?\nAI: here it is")
+
+    async def test_generate_response_streams_chunks_with_history(self):
+        self.service.rag_chain = FakeChain(chunks=["Based ", "on the docs."])
+        conversation = [HumanMessage(content="what is our policy?")]
+
+        chunks = [
+            chunk
+            async for chunk in self.service.generate_response(
+                conversation, "what is our policy?", self.workspace
+            )
+        ]
+
+        self.assertEqual("".join(chunks), "Based on the docs.")
+        payload = self.service.rag_chain.calls[0]
+        self.assertEqual(payload["query"], "what is our policy?")
+        self.assertEqual(payload["workspace"], self.workspace)
+        self.assertEqual(payload["recent_conversation"], "User: what is our policy?")
+
+    async def test_get_documents_helper_scopes_by_workspace(self):
+        document = await sync_to_async(self._text_document)()
+        other_workspace = await sync_to_async(make_workspace)(
+            await sync_to_async(make_company)("Other"), name="Other"
+        )
+
+        self.assertEqual([doc.id for doc in await get_documents()], [document.id])
+        self.assertEqual(
+            [doc.id for doc in await get_documents(self.workspace)], [document.id]
+        )
+        self.assertEqual(await get_documents(other_workspace), [])

llm_be/chat_backend/tests/test_signals.py

@@ -0,0 +1,54 @@
+import os
+from unittest import mock
+
+from django.test import TestCase
+
+from .factories import make_company, make_document, make_workspace
+
+RAG_SERVICE = "chat_backend.services.rag_services.AsyncRAGService"
+
+
+class DocumentSignalTestCase(TestCase):
+    def setUp(self):
+        self.workspace = make_workspace(make_company())
+
+    def test_creating_a_document_reindexes_the_vector_store(self):
+        with mock.patch.dict(os.environ, {"SKIP_RAG_INIT": ""}):
+            with mock.patch(RAG_SERVICE) as service:
+                make_document(self.workspace)
+
+        service.return_value.ingest_documents.assert_called_once_with()
+
+    def test_updating_a_document_does_not_reindex(self):
+        document = make_document(self.workspace)
+
+        with mock.patch.dict(os.environ, {"SKIP_RAG_INIT": ""}):
+            with mock.patch(RAG_SERVICE) as service:
+                document.active = True
+                document.save()
+
+        service.assert_not_called()
+
+    def test_deleting_a_document_reindexes_the_vector_store(self):
+        document = make_document(self.workspace)
+
+        with mock.patch.dict(os.environ, {"SKIP_RAG_INIT": ""}):
+            with mock.patch(RAG_SERVICE) as service:
+                document.delete()
+
+        service.return_value.ingest_documents.assert_called_once_with()
+
+    def test_skip_rag_init_keeps_signals_inert(self):
+        with mock.patch.dict(os.environ, {"SKIP_RAG_INIT": "1"}):
+            with mock.patch(RAG_SERVICE) as service:
+                document = make_document(self.workspace)
+                document.delete()
+
+        service.assert_not_called()
+
+    def test_vector_store_failures_do_not_break_uploads(self):
+        with mock.patch.dict(os.environ, {"SKIP_RAG_INIT": ""}):
+            with mock.patch(RAG_SERVICE, side_effect=RuntimeError("chroma down")):
+                document = make_document(self.workspace)
+
+        self.assertIsNotNone(document.pk)

llm_be/chat_backend/tests/test_storage.py

@@ -0,0 +1,94 @@
+from django.core.files.base import ContentFile
+from django.core.files.uploadedfile import SimpleUploadedFile
+from django.test import TestCase
+
+from chat_backend.models import StoredFile
+from chat_backend.storage import DatabaseStorage
+
+
+class DatabaseStorageTestCase(TestCase):
+    def setUp(self):
+        self.storage = DatabaseStorage()
+
+    def test_save_then_open_roundtrip(self):
+        name = self.storage.save("documents/hello.txt", ContentFile(b"hello world"))
+
+        self.assertEqual(name, "documents/hello.txt")
+        with self.storage.open(name) as handle:
+            self.assertEqual(handle.read(), b"hello world")
+
+    def test_save_records_size_and_content_type(self):
+        name = self.storage.save(
+            "documents/report.pdf",
+            SimpleUploadedFile(
+                "report.pdf", b"%PDF-1.3", content_type="application/pdf"
+            ),
+        )
+
+        stored = StoredFile.objects.get(name=name)
+        self.assertEqual(stored.size, len(b"%PDF-1.3"))
+        self.assertEqual(stored.content_type, "application/pdf")
+        self.assertEqual(self.storage.size(name), len(b"%PDF-1.3"))
+
+    def test_content_type_guessed_from_extension(self):
+        name = self.storage.save("documents/notes.txt", ContentFile(b"abc"))
+        self.assertEqual(StoredFile.objects.get(name=name).content_type, "text/plain")
+
+    def test_string_content_is_encoded(self):
+        name = self.storage.save("documents/str.txt", ContentFile("héllo"))
+        with self.storage.open(name) as handle:
+            self.assertEqual(handle.read().decode("utf-8"), "héllo")
+
+    def test_second_save_of_same_name_gets_unique_name(self):
+        first = self.storage.save("documents/dup.txt", ContentFile(b"one"))
+        second = self.storage.save("documents/dup.txt", ContentFile(b"two"))
+
+        self.assertNotEqual(first, second)
+        self.assertEqual(StoredFile.objects.count(), 2)
+        with self.storage.open(first) as handle:
+            self.assertEqual(handle.read(), b"one")
+
+    def test_exists_and_delete(self):
+        name = self.storage.save("documents/gone.txt", ContentFile(b"bye"))
+
+        self.assertTrue(self.storage.exists(name))
+        self.storage.delete(name)
+        self.assertFalse(self.storage.exists(name))
+        self.assertFalse(StoredFile.objects.filter(name=name).exists())
+
+    def test_delete_is_noop_for_missing_name(self):
+        self.storage.delete("documents/never-written.txt")
+
+    def test_listdir_splits_dirs_and_files(self):
+        self.storage.save("documents/top.txt", ContentFile(b"a"))
+        self.storage.save("documents/nested/inner.txt", ContentFile(b"b"))
+
+        dirs, files = self.storage.listdir("documents")
+
+        self.assertEqual(files, ["top.txt"])
+        self.assertEqual(dirs, ["nested"])
+
+    def test_url_points_at_api_route(self):
+        self.assertEqual(
+            self.storage.url("documents/a.txt"), "/api/stored-files/documents/a.txt"
+        )
+
+    def test_created_and_modified_times(self):
+        name = self.storage.save("documents/times.txt", ContentFile(b"a"))
+        stored = StoredFile.objects.get(name=name)
+
+        self.assertEqual(self.storage.get_created_time(name), stored.created)
+        self.assertEqual(self.storage.get_modified_time(name), stored.last_modified)
+
+    def test_path_and_accessed_time_are_unsupported(self):
+        with self.assertRaises(NotImplementedError):
+            self.storage.path("documents/a.txt")
+        with self.assertRaises(NotImplementedError):
+            self.storage.get_accessed_time("documents/a.txt")
+
+    def test_storage_is_deconstructible_for_migrations(self):
+        path, args, kwargs = DatabaseStorage().deconstruct()
+
+        self.assertEqual(path, "chat_backend.storage.DatabaseStorage")
+        self.assertEqual(list(args), [])
+        self.assertEqual(dict(kwargs), {})

llm_be/chat_backend/tests/test_utils.py

@@ -0,0 +1,90 @@
+import datetime
+
+from django.conf import settings
+from django.test import SimpleTestCase, override_settings
+from parameterized import parameterized
+
+from chat_backend.ollama_config import (
+    ollama_base_url,
+    ollama_embed_model,
+    ollama_embeddings_kwargs,
+    ollama_llm_kwargs,
+    ollama_model,
+)
+from chat_backend.utils import last_day_of_month
+
+
+class LastDayOfMonthTestCase(SimpleTestCase):
+    @parameterized.expand(
+        [
+            ("january", datetime.date(2025, 1, 5), datetime.date(2025, 1, 31)),
+            ("leap_february", datetime.date(2024, 2, 10), datetime.date(2024, 2, 29)),
+            ("common_february", datetime.date(2025, 2, 10), datetime.date(2025, 2, 28)),
+            ("april", datetime.date(2025, 4, 1), datetime.date(2025, 4, 30)),
+            ("december", datetime.date(2025, 12, 31), datetime.date(2025, 12, 31)),
+        ]
+    )
+    def test_returns_last_day(self, _name, given, expected):
+        self.assertEqual(last_day_of_month(given), expected)
+
+    def test_keeps_datetime_time_component(self):
+        result = last_day_of_month(datetime.datetime(2025, 6, 4, 13, 45))
+
+        self.assertEqual(result.date(), datetime.date(2025, 6, 30))
+        self.assertEqual((result.hour, result.minute), (13, 45))
+
+
+@override_settings(
+    OLLAMA_BASE_URL="http://10.0.0.128:11434",
+    OLLAMA_MODEL="llama3.2",
+    OLLAMA_EMBED_MODEL="nomic-embed-text",
+)
+class OllamaConfigTestCase(SimpleTestCase):
+    def test_reads_settings(self):
+        self.assertEqual(ollama_base_url(), "http://10.0.0.128:11434")
+        self.assertEqual(ollama_model(), "llama3.2")
+        self.assertEqual(ollama_embed_model(), "nomic-embed-text")
+
+    def test_model_argument_wins_over_settings(self):
+        self.assertEqual(ollama_model("gpt-oss:20b"), "gpt-oss:20b")
+
+    def test_llm_kwargs_carry_base_url_and_model(self):
+        kwargs = ollama_llm_kwargs(temperature=0.1, num_ctx=2048)
+
+        self.assertEqual(
+            kwargs,
+            {
+                "base_url": "http://10.0.0.128:11434",
+                "model": "llama3.2",
+                "temperature": 0.1,
+                "num_ctx": 2048,
+            },
+        )
+
+    def test_llm_kwargs_extra_overrides_defaults(self):
+        self.assertEqual(ollama_llm_kwargs(model="other")["model"], "other")
+
+    def test_embeddings_kwargs_use_embed_model(self):
+        self.assertEqual(
+            ollama_embeddings_kwargs(),
+            {"base_url": "http://10.0.0.128:11434", "model": "nomic-embed-text"},
+        )
+
+
+class OllamaConfigFallbackTestCase(SimpleTestCase):
+    """Deployed code must never fall back to a hardcoded remote host."""
+
+    def test_defaults_when_settings_are_absent(self):
+        with override_settings():
+            del settings.OLLAMA_BASE_URL
+            del settings.OLLAMA_MODEL
+            del settings.OLLAMA_EMBED_MODEL
+
+            self.assertEqual(ollama_base_url(), "http://127.0.0.1:11434")
+            self.assertEqual(ollama_model(), "llama3.2")
+            self.assertEqual(ollama_embed_model(), "llama3.2")
+
+    def test_embed_model_falls_back_to_chat_model(self):
+        with override_settings(OLLAMA_MODEL="llama3.2"):
+            del settings.OLLAMA_EMBED_MODEL
+            self.assertEqual(ollama_embed_model(), "llama3.2")

llm_be/chat_backend/tests/test_views_analytics.py

@@ -0,0 +1,129 @@
+import datetime
+
+from django.urls import reverse
+from django.utils import timezone
+from rest_framework import status
+from rest_framework.test import APITestCase
+
+from chat_backend.models import PromptMetric
+
+from .factories import make_company, make_conversation, make_prompt, make_user
+
+
+def mid_month():
+    """The 2nd of the current month — inside the window the views query, whatever today is."""
+    return timezone.now().replace(day=2, hour=12, minute=0, second=0, microsecond=0)
+
+
+class AnalyticsTestData(APITestCase):
+    """Two users in one company, one user in another, all active this month."""
+
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(email="me@example.com", company=self.company)
+        self.teammate = make_user(email="mate@example.com", company=self.company)
+        self.outsider = make_user(
+            email="out@example.com", company=make_company("Other")
+        )
+        self.client.force_authenticate(user=self.user)
+
+        self.my_conversation = self._conversation(self.user, prompts=2)
+        self._conversation(self.teammate, prompts=1)
+        self._conversation(self.outsider, prompts=3)
+
+    def _conversation(self, user, prompts: int):
+        conversation = make_conversation(user=user, created=mid_month())
+        for index in range(prompts):
+            make_prompt(conversation, message=f"prompt {index}", created=mid_month())
+        return conversation
+
+
+class UserPromptAnalyticsTestCase(AnalyticsTestData):
+    def test_returns_three_months_oldest_first(self):
+        response = self.client.get(reverse("analytics_user_prompts"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(len(response.data), 3)
+        self.assertEqual(response.data[-1]["month"], timezone.now().strftime("%B"))
+
+    def test_current_month_counts(self):
+        current = self.client.get(reverse("analytics_user_prompts")).data[-1]
+
+        self.assertEqual(current["you"], 2)
+        self.assertEqual(current["others"], 3 / 2)
+        self.assertEqual(current["all"], 6 / 3)
+
+    def test_earlier_months_are_empty(self):
+        earlier = self.client.get(reverse("analytics_user_prompts")).data[0]
+
+        self.assertEqual(earlier["you"], 0)
+        self.assertEqual(earlier["others"], 0)
+        self.assertEqual(earlier["all"], 0)
+
+
+class UserConversationAnalyticsTestCase(AnalyticsTestData):
+    def test_current_month_counts(self):
+        response = self.client.get(reverse("analytics_user_conversations"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        current = response.data[-1]
+        self.assertEqual(current["you"], 1)
+        self.assertEqual(current["others"], 2 / 2)
+        self.assertEqual(current["all"], 3 / 3)
+
+
+class CompanyUsageAnalyticsTestCase(AnalyticsTestData):
+    def test_counts_active_and_idle_company_users(self):
+        response = self.client.get(reverse("analytics_company_usage"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        current = response.data[-1]
+        self.assertEqual(current["used"], 2)
+        self.assertEqual(current["not_used"], 0)
+
+    def test_idle_users_are_reported(self):
+        make_user(email="idle@example.com", company=self.company)
+
+        current = self.client.get(reverse("analytics_company_usage")).data[-1]
+
+        self.assertEqual(current["used"], 2)
+        self.assertEqual(current["not_used"], 1)
+
+
+class AdminAnalyticsTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company())
+        self.client.force_authenticate(user=self.user)
+
+    def _metric(self, seconds: int):
+        start = mid_month()
+        return PromptMetric.objects.create(
+            prompt_id=1,
+            conversation_id=1,
+            model_name="llama3.2",
+            start_time=start,
+            end_time=start + datetime.timedelta(seconds=seconds),
+            prompt_length=10,
+            has_file=False,
+            created=start,
+        )
+
+    def test_zeros_when_no_metrics(self):
+        response = self.client.get(reverse("analytics_admin"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(len(response.data), 3)
+        for month in response.data:
+            self.assertEqual(month["range"], [0, 0])
+            self.assertEqual(month["avg"], 0)
+            self.assertEqual(month["median"], 0)
+
+    def test_duration_statistics_for_current_month(self):
+        self._metric(10)
+        self._metric(20)
+
+        current = self.client.get(reverse("analytics_admin")).data[-1]
+
+        self.assertEqual(current["range"], [10, 20])
+        self.assertEqual(current["avg"], 15)
+        self.assertEqual(current["median"], 20)

llm_be/chat_backend/tests/test_views_conversations.py

@@ -0,0 +1,141 @@
+from django.urls import reverse
+from rest_framework import status
+from rest_framework.test import APITestCase
+
+from chat_backend.models import Conversation, Prompt
+
+from .factories import make_company, make_conversation, make_prompt, make_user
+
+
+class ConversationsViewTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company())
+        self.client.force_authenticate(user=self.user)
+        self.url = reverse("conversations")
+
+    def test_list_hides_deleted_and_other_users_conversations(self):
+        make_conversation(user=self.user, title="Mine")
+        make_conversation(user=self.user, title="Removed", deleted=True)
+        make_conversation(user=make_user(email="other@example.com"), title="Theirs")
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual([item["title"] for item in response.data], ["Mine"])
+
+    def test_list_order_follows_user_preference(self):
+        make_conversation(user=self.user, title="First")
+        make_conversation(user=self.user, title="Second")
+
+        oldest_first = self.client.get(self.url).data
+        self.assertEqual([item["title"] for item in oldest_first], ["First", "Second"])
+
+        self.user.conversation_order = False
+        self.user.save()
+
+        newest_first = self.client.get(self.url).data
+        self.assertEqual([item["title"] for item in newest_first], ["Second", "First"])
+
+    def test_post_creates_conversation_for_current_user(self):
+        response = self.client.post(self.url, {"name": "New Chat"}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        conversation = Conversation.objects.get(id=response.data["id"])
+        self.assertEqual(conversation.title, "New Chat")
+        self.assertEqual(conversation.user, self.user)
+        self.assertEqual(response.data["title"], "New Chat")
+
+
+class ConversationPreferencesTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company())
+        self.client.force_authenticate(user=self.user)
+        self.url = reverse("conversation_preferences")
+
+    def test_get_returns_current_order(self):
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertTrue(response.data["order"])
+
+    def test_post_toggles_and_persists_order(self):
+        response = self.client.post(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertFalse(response.data["order"])
+        self.user.refresh_from_db()
+        self.assertFalse(self.user.conversation_order)
+
+
+class ConversationDetailViewTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company())
+        self.client.force_authenticate(user=self.user)
+        self.conversation = make_conversation(user=self.user)
+        self.url = reverse("conversation_details")
+
+    def test_get_returns_prompts_of_conversation(self):
+        make_prompt(self.conversation, message="hello")
+        make_prompt(self.conversation, message="hi there", user_created=False)
+        make_prompt(make_conversation(user=self.user), message="other conversation")
+
+        response = self.client.get(self.url, {"conversation_id": self.conversation.id})
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(
+            [(item["message"], item["user_created"]) for item in response.data],
+            [("hello", True), ("hi there", False)],
+        )
+
+    def test_post_stores_assistant_prompt(self):
+        response = self.client.post(
+            self.url,
+            {
+                "prompt": "generated answer",
+                "conversation_id": self.conversation.id,
+                "is_user": False,
+            },
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        prompt = Prompt.objects.get()
+        self.assertEqual(prompt.message, "generated answer")
+        self.assertFalse(prompt.user_created)
+        self.assertEqual(prompt.conversation_id, self.conversation.id)
+
+    def test_post_stores_user_prompt_and_broadcasts(self):
+        response = self.client.post(
+            self.url,
+            {
+                "prompt": "what is the weather",
+                "conversation_id": self.conversation.id,
+                "is_user": True,
+            },
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        prompt = Prompt.objects.get()
+        self.assertTrue(prompt.user_created)
+        self.assertEqual(prompt.conversation_id, self.conversation.id)
+
+    def test_post_for_unknown_conversation_stores_nothing(self):
+        response = self.client.post(
+            self.url,
+            {"prompt": "orphan", "conversation_id": 4242, "is_user": True},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(Prompt.objects.count(), 0)
+
+    def test_delete_soft_deletes_conversation(self):
+        response = self.client.delete(
+            self.url, {"conversation_id": self.conversation.id}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_202_ACCEPTED)
+        self.conversation.refresh_from_db()
+        self.assertTrue(self.conversation.deleted)
+        self.assertEqual(Conversation.objects.count(), 1)

llm_be/chat_backend/tests/test_views_documents.py

@@ -0,0 +1,142 @@
+from unittest import mock
+
+from django.urls import reverse
+from rest_framework import status
+from rest_framework.test import APITestCase
+
+from chat_backend.models import Document, DocumentWorkspace, StoredFile
+
+from .factories import (
+    make_company,
+    make_document,
+    make_user,
+    make_workspace,
+    pdf_upload,
+)
+
+
+class DocumentWorkspaceViewTestCase(APITestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.client.force_authenticate(user=self.user)
+        self.workspace = make_workspace(self.company)
+        self.url = reverse("document_workspaces")
+
+    def test_list_only_returns_own_company_workspaces(self):
+        make_workspace(make_company("Other"), name="Other Workspace")
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual([item["name"] for item in response.data], ["Test Workspace"])
+
+    def test_create_workspace_attaches_company(self):
+        response = self.client.post(self.url, {"name": "New Workspace"}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        self.assertEqual(DocumentWorkspace.objects.count(), 2)
+        created = DocumentWorkspace.objects.get(name="New Workspace")
+        self.assertEqual(created.company, self.company)
+
+    def test_create_workspace_requires_name(self):
+        response = self.client.post(self.url, {}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertIn("name", response.data)
+
+
+class DocumentUploadViewTestCase(APITestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.client.force_authenticate(user=self.user)
+        self.workspace = make_workspace(self.company)
+        self.url = reverse("documents")
+        rag_patcher = mock.patch("chat_backend.views.AsyncRAGService")
+        self.rag_service = rag_patcher.start()
+        self.addCleanup(rag_patcher.stop)
+
+    def test_upload_stores_bytes_in_database_and_marks_processed(self):
+        response = self.client.post(
+            self.url, {"file": pdf_upload()}, format="multipart"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        document = Document.objects.get()
+        self.assertEqual(document.workspace, self.workspace)
+        self.assertTrue(document.processed)
+        self.assertTrue(document.active)
+        self.assertTrue(StoredFile.objects.filter(name=document.file.name).exists())
+
+    def test_upload_hands_file_to_the_rag_service(self):
+        self.client.post(self.url, {"file": pdf_upload()}, format="multipart")
+
+        self.rag_service.return_value.add_files_to_store.assert_called_once()
+        _args, kwargs = self.rag_service.return_value.add_files_to_store.call_args
+        self.assertEqual(kwargs["workspace_id"], self.workspace.id)
+
+    def test_upload_without_file_is_rejected(self):
+        response = self.client.post(self.url, {}, format="multipart")
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertEqual(Document.objects.count(), 0)
+
+    def test_upload_of_non_file_payload_is_rejected(self):
+        response = self.client.post(
+            self.url, {"file": "not a file"}, format="multipart"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+
+    def test_upload_without_workspace_returns_404(self):
+        self.workspace.delete()
+
+        response = self.client.post(
+            self.url, {"file": pdf_upload()}, format="multipart"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
+
+    def test_list_returns_documents_of_own_workspace(self):
+        make_document(self.workspace)
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(len(response.data), 1)
+        self.assertIn("test", response.data[0]["file"])
+        self.assertIn("pdf", response.data[0]["file"])
+
+    def test_list_without_workspace_returns_404(self):
+        self.workspace.delete()
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
+        self.assertEqual(response.data["error"], "Workspace not found")
+
+
+class DocumentDetailViewTestCase(APITestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.user = make_user(company=self.company)
+        self.client.force_authenticate(user=self.user)
+        self.workspace = make_workspace(self.company)
+
+    def test_other_companies_documents_are_not_reachable(self):
+        other_workspace = make_workspace(make_company("Other"), name="Other Workspace")
+        other_document = make_document(other_workspace)
+
+        url = reverse("documents_details", kwargs={"document_id": other_document.id})
+        response = self.client.get(url)
+
+        self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
+
+    def test_unknown_document_returns_404(self):
+        url = reverse("documents_details", kwargs={"document_id": 4242})
+
+        response = self.client.get(url)
+
+        self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND)
+        self.assertEqual(response.data["error"], "Document not found")

llm_be/chat_backend/tests/test_views_users.py

@@ -0,0 +1,354 @@
+from django.core import mail
+from django.urls import reverse
+from rest_framework import status
+from rest_framework.test import APITestCase
+from rest_framework_simplejwt.tokens import RefreshToken
+
+from chat_backend.models import Announcement, CustomUser, Feedback
+
+from .factories import make_company, make_user
+
+
+class AuthenticationRequiredTestCase(APITestCase):
+    def test_protected_endpoints_reject_anonymous_requests(self):
+        for name in [
+            "get_user",
+            "conversations",
+            "feedbacks",
+            "company_users",
+            "conversation_preferences",
+            "analytics_user_prompts",
+            "documents",
+        ]:
+            with self.subTest(endpoint=name):
+                response = self.client.get(reverse(name))
+                self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED)
+
+    def test_announcements_are_public(self):
+        Announcement.objects.create(message="scheduled maintenance")
+
+        response = self.client.get(reverse("get_announcments"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(response.data[0]["message"], "scheduled maintenance")
+
+
+class TokenTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(email="person@example.com", password="testpass123")
+
+    def test_obtain_token_pair(self):
+        response = self.client.post(
+            reverse("token_create"),
+            {"username": self.user.username, "password": "testpass123"},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertIn("access", response.data)
+        self.assertIn("refresh", response.data)
+
+    def test_obtain_token_rejects_bad_password(self):
+        response = self.client.post(
+            reverse("token_create"),
+            {"username": self.user.username, "password": "wrong"},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED)
+
+    def test_access_token_authenticates_requests(self):
+        token = self.client.post(
+            reverse("token_create"),
+            {"username": self.user.username, "password": "testpass123"},
+            format="json",
+        ).data["access"]
+
+        self.client.credentials(HTTP_AUTHORIZATION=f"JWT {token}")
+        response = self.client.get(reverse("get_user"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(response.data["email"], self.user.email)
+
+    def test_blacklist_refresh_token(self):
+        refresh = RefreshToken.for_user(self.user)
+
+        response = self.client.post(
+            reverse("blacklist"), {"refresh_token": str(refresh)}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_205_RESET_CONTENT)
+
+    def test_blacklist_rejects_garbage_token(self):
+        response = self.client.post(
+            reverse("blacklist"), {"refresh_token": "not-a-token"}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+
+    def test_is_authenticated_endpoint(self):
+        url = reverse("is_authenticated")
+
+        self.assertEqual(self.client.get(url).status_code, status.HTTP_401_UNAUTHORIZED)
+
+        self.client.force_login(self.user)
+        self.assertEqual(self.client.get(url).status_code, status.HTTP_200_OK)
+
+
+class UserCreateTestCase(APITestCase):
+    def test_invalid_payload_returns_errors(self):
+        response = self.client.post(reverse("create_user"), {}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertIn("email", response.data)
+        self.assertEqual(CustomUser.objects.count(), 0)
+
+
+class SetPasswordTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(email="person@example.com", password="testpass123")
+
+    def test_get_rejects_user_that_already_has_a_password(self):
+        url = reverse("set_password", kwargs={"slug": self.user.slug})
+
+        self.assertEqual(self.client.get(url).status_code, status.HTTP_401_UNAUTHORIZED)
+
+    def test_get_allows_user_awaiting_a_password(self):
+        self.user.set_unusable_password()
+        self.user.save()
+
+        url = reverse("set_password", kwargs={"slug": self.user.slug})
+
+        self.assertEqual(self.client.get(url).status_code, status.HTTP_200_OK)
+
+    def test_post_sets_password(self):
+        url = reverse("set_password", kwargs={"slug": self.user.slug})
+
+        response = self.client.post(url, {"password": "brandnewpass"}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.user.refresh_from_db()
+        self.assertTrue(self.user.check_password("brandnewpass"))
+
+
+class AcknowledgeTermsOfServiceTestCase(APITestCase):
+    def test_post_marks_tos_signed(self):
+        user = make_user()
+        self.client.force_authenticate(user=user)
+
+        response = self.client.post(reverse("acknowledge_tos"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        user.refresh_from_db()
+        self.assertTrue(user.has_signed_tos)
+
+
+class UserInviteTestCase(APITestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.manager = make_user(
+            email="manager@example.com", company=self.company, is_company_manager=True
+        )
+        self.url = reverse("invite_user")
+
+    def test_manager_invites_new_user_and_email_is_sent(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url, {"email": "newhire@example.com"}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        invited = CustomUser.objects.get(email="newhire@example.com")
+        self.assertEqual(invited.company_id, self.company.id)
+        self.assertEqual(invited.username, "newhire@example.com")
+        self.assertEqual(len(mail.outbox), 1)
+        self.assertIn(invited.slug, mail.outbox[0].body)
+        self.assertEqual(mail.outbox[0].to, ["newhire@example.com"])
+
+    def test_non_manager_cannot_invite(self):
+        member = make_user(email="member@example.com", company=self.company)
+        self.client.force_authenticate(user=member)
+
+        response = self.client.post(
+            self.url, {"email": "newhire@example.com"}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertFalse(
+            CustomUser.objects.filter(email="newhire@example.com").exists()
+        )
+
+    def test_malformed_email_is_rejected(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(self.url, {"email": "nope"}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertEqual(len(mail.outbox), 0)
+
+    def test_duplicate_email_is_rejected(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url, {"email": self.manager.email}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+
+    def test_get_is_not_allowed(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_405_METHOD_NOT_ALLOWED)
+
+
+class FeedbackViewTestCase(APITestCase):
+    def setUp(self):
+        self.user = make_user(company=make_company())
+        self.client.force_authenticate(user=self.user)
+        self.url = reverse("feedbacks")
+
+    def test_post_creates_feedback_and_notifies(self):
+        response = self.client.post(
+            self.url,
+            {"title": "Broken button", "text": "It does nothing"},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_201_CREATED)
+        feedback = Feedback.objects.get()
+        self.assertEqual(feedback.user, self.user)
+        self.assertEqual(len(mail.outbox), 1)
+        self.assertIn("Broken button", mail.outbox[0].body)
+
+    def test_post_without_text_is_rejected(self):
+        response = self.client.post(self.url, {"title": "no body"}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertEqual(Feedback.objects.count(), 0)
+
+    def test_get_only_returns_own_feedback(self):
+        Feedback.objects.create(title="mine", text="a", user=self.user)
+        Feedback.objects.create(
+            title="theirs", text="b", user=make_user(email="o@o.com")
+        )
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual([item["title"] for item in response.data], ["mine"])
+
+
+class CompanyUsersViewTestCase(APITestCase):
+    def setUp(self):
+        self.company = make_company()
+        self.manager = make_user(
+            email="manager@example.com", company=self.company, is_company_manager=True
+        )
+        self.member = make_user(email="member@example.com", company=self.company)
+        self.url = reverse("company_users")
+
+    def test_manager_lists_company_users(self):
+        make_user(email="outsider@example.com", company=make_company("Other"))
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(
+            {item["email"] for item in response.data},
+            {"manager@example.com", "member@example.com"},
+        )
+
+    def test_member_cannot_list_company_users(self):
+        self.client.force_authenticate(user=self.member)
+
+        response = self.client.get(self.url)
+
+        self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED)
+
+    def test_manager_toggles_is_active(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url, {"email": self.member.email, "field": "is_active"}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.member.refresh_from_db()
+        self.assertFalse(self.member.is_active)
+
+    def test_manager_toggles_company_manager(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url,
+            {"email": self.member.email, "field": "company_manager"},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.member.refresh_from_db()
+        self.assertTrue(self.member.is_company_manager)
+
+    def test_manager_clears_password(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url,
+            {"email": self.member.email, "field": "has_password"},
+            format="json",
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.member.refresh_from_db()
+        self.assertFalse(self.member.has_usable_password())
+
+    def test_manager_cannot_touch_other_company_users(self):
+        outsider = make_user(
+            email="outsider@example.com", company=make_company("Other")
+        )
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.post(
+            self.url, {"email": outsider.email, "field": "is_active"}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED)
+        outsider.refresh_from_db()
+        self.assertTrue(outsider.is_active)
+
+    def test_manager_deletes_company_user(self):
+        self.client.force_authenticate(user=self.manager)
+
+        response = self.client.delete(
+            self.url, {"email": self.member.email}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertFalse(CustomUser.objects.filter(email=self.member.email).exists())
+
+    def test_member_cannot_delete_users(self):
+        self.client.force_authenticate(user=self.member)
+
+        response = self.client.delete(
+            self.url, {"email": self.manager.email}, format="json"
+        )
+
+        self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED)
+        self.assertTrue(CustomUser.objects.filter(email=self.manager.email).exists())
+
+
+class CustomUserGetTestCase(APITestCase):
+    def test_returns_authenticated_user(self):
+        user = make_user(company=make_company("Globex"))
+        self.client.force_authenticate(user=user)
+
+        response = self.client.get(reverse("get_user"))
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(response.data["email"], user.email)
+        self.assertEqual(response.data["company"]["name"], "Globex")
+        self.assertNotIn("password", response.data)

llm_be/chat_backend/views.py

@@ -30,7 +30,6 @@ from .models import (
 )
 from django.views.decorators.cache import never_cache
 from django.http import JsonResponse
-from datetime import datetime
 from .client import LlamaClient
 from asgiref.sync import sync_to_async, async_to_sync
 from channels.generic.websocket import AsyncWebsocketConsumer
@@ -73,19 +72,19 @@ from .services.data_analysis_service import AsyncDataAnalysisService
 from .ollama_config import ollama_llm_kwargs, ollama_model
 
 
-
 from langchain_classic.chains import create_retrieval_chain
 from langchain_classic.chains.combine_documents import create_stuff_documents_chain
 from langchain_ollama import ChatOllama
 
 import logging
-logger = logging.getLogger(__name__)
 
+logger = logging.getLogger(__name__)
 
 
 CHANNEL_NAME: str = "llm_messages"
 MODEL_NAME: str = ollama_model()
 
+
 # Create your views here.
 class CustomObtainTokenView(TokenObtainPairView):
     permission_classes = (permissions.AllowAny,)
@@ -480,7 +479,7 @@ class ConversationDetailView(APIView):
                 data={
                     "message": prompt,
                     "user_created": is_user,
-                    "created": datetime.now(),
+                    "created": timezone.now(),
                 }
             )
             if serializer.is_valid():
@@ -725,9 +724,6 @@ llm = OllamaLLM(**ollama_llm_kwargs(model=MODEL_NAME))
 # chain = prompt | llm.with_config({"run_name": "model"}) | output_parser.with_config({"run_name": "Assistant"})
 
 
-
-
-
 # Document Views
 class DocumentWorkspaceView(APIView):
     # permission_classes = [permissions.IsAuthenticated]