From bfc3b4a3a88d3a5ad433343be37f97acfc2e769b Mon Sep 17 00:00:00 2001
From: Conrado Silva Miranda <csilvamiranda@cloudera.com>
Date: Thu, 21 Nov 2024 14:09:29 -0800
Subject: [PATCH] Use lipsum to generate bigger text file for test

---
 llm-service/app/tests/conftest.py |  6 +++++-
 llm-service/pdm.lock              | 11 ++++++++++-
 llm-service/pyproject.toml        |  1 +
 3 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/llm-service/app/tests/conftest.py b/llm-service/app/tests/conftest.py
index 092df5048..eec513e26 100644
--- a/llm-service/app/tests/conftest.py
+++ b/llm-service/app/tests/conftest.py
@@ -44,6 +44,7 @@
 from typing import Any, Dict, Sequence
 
 import boto3
+import lipsum
 import pytest
 from boto3.resources.base import ServiceResource
 from fastapi.testclient import TestClient
@@ -67,6 +68,7 @@
 from app.services.rag_qdrant_vector_store import RagQdrantVectorStore
 from app.services.utils import get_last_segment
 
+
 @dataclass
 class BotoObject:
     bucket_name: str
@@ -249,12 +251,14 @@ def s3_object(
     bucket_name = "test_bucket"
     key = "test/" + document_id
 
+    body = lipsum.generate_words(1000)
+
     bucket = s3_client.Bucket(bucket_name)
     bucket.create(CreateBucketConfiguration={"LocationConstraint": aws_region})
     bucket.put_object(
         Key=key,
         # TODO: fixturize file
-        Body=b"Some text to be summarized and indexed",
+        Body=body.encode("utf-8"),
         Metadata={"originalfilename": "test.txt"},
     )
     return BotoObject(bucket_name=bucket_name, key=key)
diff --git a/llm-service/pdm.lock b/llm-service/pdm.lock
index cddb261f3..a742bad18 100644
--- a/llm-service/pdm.lock
+++ b/llm-service/pdm.lock
@@ -5,7 +5,7 @@
 groups = ["default", "dev"]
 strategy = ["inherit_metadata"]
 lock_version = "4.5.0"
-content_hash = "sha256:2a9b3e86ee90d639241d72fdeee20e779c2e7c42e90ab2e43e335c18454e0858"
+content_hash = "sha256:8a759766434529522c347027893fec1b8c80f1d4e4b42c37a2f66f52ed1e6a6a"
 
 [[metadata.targets]]
 requires_python = "==3.10.*"
@@ -795,6 +795,15 @@ files = [
     {file = "jsonpath_python-1.0.6-py3-none-any.whl", hash = "sha256:1e3b78df579f5efc23565293612decee04214609208a2335884b3ee3f786b575"},
 ]
 
+[[package]]
+name = "lipsum"
+version = "0.1.2"
+summary = "A randomised Lorem Ipsum generator library for Python"
+groups = ["dev"]
+files = [
+    {file = "lipsum-0.1.2.tar.gz", hash = "sha256:ba5f46cef19104c07f889b14486a3772845fc25afa1eb5e2eee1f2d9badcb8ab"},
+]
+
 [[package]]
 name = "llama-index-core"
 version = "0.10.68"
diff --git a/llm-service/pyproject.toml b/llm-service/pyproject.toml
index d22188469..d48ea05f7 100644
--- a/llm-service/pyproject.toml
+++ b/llm-service/pyproject.toml
@@ -16,6 +16,7 @@ dev = [
     "pytest>=8.3.3",
     "ruff>=0.7.4",
     "mypy>=1.13.0",
+    "lipsum>=0.1.2",
 ]
 
 [tool.pytest.ini_options]