seed database and get threads

server/cli.py

@@ -1,7 +1,6 @@
 """Flask CLI commands."""
 
 import datetime
-import json
 
 from flask import Blueprint
 
@@ -11,6 +10,7 @@
     increment_response_count,
     thread_emails_to_openai_messages,
 )
+from server.fake_data import generate_test_documents
 from server.models.document import Document
 from server.models.email import Email
 from server.models.response import Response
@@ -20,37 +20,6 @@
 
 seed = Blueprint("seed", __name__)
 
-FLASK_SEED_CORPUS = "server/nlp/corpus_flask_seed.json"
-
-
-def _generate_test_documents():
-    """Generate test documents."""
-    with open(FLASK_SEED_CORPUS) as f:
-        corpus = json.load(f)
-
-    documents = []
-    for doc in corpus:
-        document = Document(
-            question=doc["question"],
-            label=doc["question"],
-            source=doc["source"],
-            content=doc["content"],
-        )
-        db.session.add(document)
-        db.session.commit()
-        documents.append(document)
-
-    test_documents = [
-        {
-            "question": doc.question,
-            "source": doc.source,
-            "content": doc.content,
-            "sql_id": doc.id,
-        }
-        for doc in documents
-    ]
-    return test_documents
-
 
 def _embed_existing_documents(documents: list[Document]):
     """Embed existing documents."""
@@ -69,7 +38,7 @@ def _embed_existing_documents(documents: list[Document]):
 @seed.cli.command()
 def corpus():
     """Add test documents to the corpus."""
-    test_documents = _generate_test_documents()
+    test_documents = generate_test_documents()
     embed_corpus(test_documents)
 
 
@@ -82,7 +51,7 @@ def email():
         # responses, so the only way this command succeeds is if the corpus is
         # already populated
         print("No documents in the database. Generating test documents...")
-        test_documents = _generate_test_documents()
+        test_documents = generate_test_documents()
         embed_corpus(test_documents)
     else:
         print("Embedding existing documents...")

server/fake_data.py

@@ -0,0 +1,37 @@
+"""Fake data for seed cli and testing."""
+
+import json
+
+from server import db
+from server.models.document import Document
+
+FLASK_SEED_CORPUS = "server/nlp/corpus_flask_seed.json"
+
+
+def generate_test_documents():
+    """Generate test documents."""
+    with open(FLASK_SEED_CORPUS) as f:
+        corpus = json.load(f)
+
+    documents = []
+    for doc in corpus:
+        document = Document(
+            question=doc["question"],
+            label=doc["question"],
+            source=doc["source"],
+            content=doc["content"],
+        )
+        db.session.add(document)
+        db.session.commit()
+        documents.append(document)
+
+    test_documents = [
+        {
+            "question": doc.question,
+            "source": doc.source,
+            "content": doc.content,
+            "sql_id": doc.id,
+        }
+        for doc in documents
+    ]
+    return test_documents

server/nlp/responses.py

@@ -131,7 +131,6 @@ def generate_context(
     docs = {}
 
     results = query_all(3, questions)
-    print("results", results)
     message = "Here is some context to help you answer this email: \n"
     for result in results:
         confidence = 0

server_tests/conftest.py

@@ -19,6 +19,7 @@
 
 from server import create_app, db
 from server.config import LOCAL, VECTOR_DIMENSION
+from server_tests.utils import seed_database
 
 
 @pytest.fixture(scope="session")
@@ -75,6 +76,9 @@ def app(db_url: str, redis_host: str):
         }
     )
 
+    with app.app_context():
+        seed_database(db)
+
     yield app
 
 

server_tests/test_email.py

@@ -8,3 +8,22 @@ def test_get_threads(app: APIFlask, client: FlaskClient):
     """Test fetching threads."""
     response = client.get("/api/emails/get_threads")
     assert_status(response, 200)
+
+    threads = response.json
+    assert threads is not None
+    assert len(threads) == 1
+
+    thread = threads[0]
+    assert thread["id"] == 1
+    assert not thread["resolved"]  # resolved is false
+    assert len(thread["emailList"]) == 5
+
+    for email in thread["emailList"]:
+        # checking that .map() is intact
+        assert email["id"] is not None
+        assert email["body"] is not None
+        assert email["subject"] is not None
+        assert email["sender"] is not None
+        assert email["message_id"] is not None
+        assert email["is_reply"] is not None
+        assert email["thread_id"] is not None

server_tests/utils.py

@@ -1,9 +1,15 @@
 """Utils for testing."""
 
+import datetime
 import logging
 
 from werkzeug.test import TestResponse
 
+from server import ProperlyTypedSQLAlchemy
+from server.fake_data import generate_test_documents
+from server.models.email import Email
+from server.models.thread import Thread
+
 
 def assert_status(response: TestResponse, status: int):
     """Asserts a response's status code, logging the response if it fails."""
@@ -15,3 +21,31 @@ def assert_status(response: TestResponse, status: int):
             f"Response body: {response.data.decode()}"
         )
         raise
+
+def seed_database(db: ProperlyTypedSQLAlchemy):
+    """Seeds the database with some fake data."""
+
+    # add some documents to the database
+    generate_test_documents()
+
+    # create fake thread with 5 emails
+    thread = Thread()
+    db.session.add(thread)
+    db.session.commit()
+
+    emails = []
+    for i in range(5):
+        # every other email is a reply sent from pigeon
+        is_reply = i % 2 == 1
+        test_email = Email(
+            date=datetime.datetime.now(datetime.timezone.utc),
+            sender="[email protected]",
+            subject="Test Subject",
+            body="Test Body",
+            message_id=f"test-message-id-{i}",
+            is_reply=is_reply,
+            thread_id=thread.id
+        )
+        emails.append(test_email)
+        db.session.add(test_email)
+    db.session.commit()