Migrate ChatGPT function to openai v1.0 (#1368)

Migrate ChatGPT function to openai v1.0.

The test is skipped in circleCI because we must supply the
`OPENAI_API_KEY`. The test passes on local machine.

- [x] Upgrade ChatGPT function.
- [x] Upgrade Dall-e function.
- [x] Update unit test cases.
- [x] Verify that notebooks work correctly.

Author: pchunduri6

evadb/functions/chatgpt.py

@@ -115,19 +115,21 @@ def setup(
     )
     def forward(self, text_df):
         try_to_import_openai()
-        import openai
+        from openai import OpenAI
 
-        @retry(tries=6, delay=20)
-        def completion_with_backoff(**kwargs):
-            return openai.ChatCompletion.create(**kwargs)
-
-        openai.api_key = self.openai_api_key
-        if len(openai.api_key) == 0:
-            openai.api_key = os.environ.get("OPENAI_API_KEY", "")
+        api_key = self.openai_api_key
+        if len(self.openai_api_key) == 0:
+            api_key = os.environ.get("OPENAI_API_KEY", "")
         assert (
-            len(openai.api_key) != 0
+            len(api_key) != 0
         ), "Please set your OpenAI API key using SET OPENAI_API_KEY = 'sk-' or environment variable (OPENAI_API_KEY)"
 
+        client = OpenAI(api_key=api_key)
+
+        @retry(tries=6, delay=20)
+        def completion_with_backoff(**kwargs):
+            return client.chat.completions.create(**kwargs)
+
         queries = text_df[text_df.columns[0]]
         content = text_df[text_df.columns[0]]
         if len(text_df.columns) > 1:

evadb/functions/dalle.py

@@ -56,24 +56,25 @@ def setup(self, openai_api_key="") -> None:
     )
     def forward(self, text_df):
         try_to_import_openai()
-        import openai
+        from openai import OpenAI
 
-        openai.api_key = self.openai_api_key
-        # If not found, try OS Environment Variable
-        if len(openai.api_key) == 0:
-            openai.api_key = os.environ.get("OPENAI_API_KEY", "")
+        api_key = self.openai_api_key
+        if len(self.openai_api_key) == 0:
+            api_key = os.environ.get("OPENAI_API_KEY", "")
         assert (
-            len(openai.api_key) != 0
-        ), "Please set your OpenAI API key using SET OPENAI_API_KEY = 'sk-'  or environment variable (OPENAI_API_KEY)"
+            len(api_key) != 0
+        ), "Please set your OpenAI API key using SET OPENAI_API_KEY = 'sk-' or environment variable (OPENAI_API_KEY)"
+
+        client = OpenAI(api_key=api_key)
 
         def generate_image(text_df: PandasDataframe):
             results = []
             queries = text_df[text_df.columns[0]]
             for query in queries:
-                response = openai.Image.create(prompt=query, n=1, size="1024x1024")
+                response = client.images.generate(prompt=query, n=1, size="1024x1024")
 
                 # Download the image from the link
-                image_response = requests.get(response["data"][0]["url"])
+                image_response = requests.get(response.data[0].url)
                 image = Image.open(BytesIO(image_response.content))
 
                 # Convert the image to an array format suitable for the DataFrame

setup.py

@@ -80,7 +80,7 @@ def read(path, encoding="utf-8"):
     "sentence-transformers",
     "protobuf",
     "bs4",
-    "openai==0.28",  # CHATGPT
+    "openai>=1.0",  # CHATGPT
     "gpt4all",  # PRIVATE GPT
     "sentencepiece",  # TRANSFORMERS
 ]

test/integration_tests/long/functions/test_chatgpt.py

@@ -13,6 +13,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import os
 import unittest
 from test.markers import chatgpt_skip_marker
 from test.util import get_evadb_for_testing
@@ -22,9 +23,8 @@
 from evadb.server.command_handler import execute_query_fetch_all
 
 
-def create_dummy_csv_file(config) -> str:
-    tmp_dir_from_config = config.get_value("storage", "tmp_dir")
-
+def create_dummy_csv_file(catalog) -> str:
+    tmp_dir_from_config = catalog.get_configuration_catalog_value("tmp_dir")
     df_dict = [
         {
             "prompt": "summarize",
@@ -49,17 +49,18 @@ def setUp(self) -> None:
             );"""
         execute_query_fetch_all(self.evadb, create_table_query)
 
-        self.csv_file_path = create_dummy_csv_file(self.evadb.config)
+        self.csv_file_path = create_dummy_csv_file(self.evadb.catalog())
 
         csv_query = f"""LOAD CSV '{self.csv_file_path}' INTO MyTextCSV;"""
         execute_query_fetch_all(self.evadb, csv_query)
+        os.environ["OPENAI_API_KEY"] = "sk-..."
 
     def tearDown(self) -> None:
         execute_query_fetch_all(self.evadb, "DROP TABLE IF EXISTS MyTextCSV;")
 
     @chatgpt_skip_marker
     def test_openai_chat_completion_function(self):
-        function_name = "OpenAIChatCompletion"
+        function_name = "ChatGPT"
         execute_query_fetch_all(self.evadb, f"DROP FUNCTION IF EXISTS {function_name};")
 
         create_function_query = f"""CREATE FUNCTION IF NOT EXISTS{function_name}
@@ -69,4 +70,4 @@ def test_openai_chat_completion_function(self):
 
         gpt_query = f"SELECT {function_name}('summarize', content) FROM MyTextCSV;"
         output_batch = execute_query_fetch_all(self.evadb, gpt_query)
-        self.assertEqual(output_batch.columns, ["openaichatcompletion.response"])
+        self.assertEqual(output_batch.columns, ["chatgpt.response"])

test/unit_tests/test_dalle.py

@@ -16,13 +16,26 @@
 import unittest
 from io import BytesIO
 from test.util import get_evadb_for_testing
+from typing import List, Optional
 from unittest.mock import MagicMock, patch
 
-from PIL import Image
+from PIL import Image as PILImage
+from pydantic import AnyUrl, BaseModel
 
 from evadb.server.command_handler import execute_query_fetch_all
 
 
+class Image(BaseModel):
+    b64_json: Optional[str]  # Replace with the actual type if different
+    revised_prompt: Optional[str]  # Replace with the actual type if different
+    url: AnyUrl
+
+
+class ImagesResponse(BaseModel):
+    created: Optional[int]  # Replace with the actual type if different
+    data: List[Image]
+
+
 class DallEFunctionTest(unittest.TestCase):
     def setUp(self) -> None:
         self.evadb = get_evadb_for_testing()
@@ -43,10 +56,10 @@ def tearDown(self) -> None:
 
     @patch.dict("os.environ", {"OPENAI_API_KEY": "mocked_openai_key"})
     @patch("requests.get")
-    @patch("openai.Image.create", return_value={"data": [{"url": "mocked_url"}]})
-    def test_dalle_image_generation(self, mock_openai_create, mock_requests_get):
+    @patch("openai.OpenAI")
+    def test_dalle_image_generation(self, mock_openai, mock_requests_get):
         # Generate a 1x1 white pixel PNG image in memory
-        img = Image.new("RGB", (1, 1), color="white")
+        img = PILImage.new("RGB", (1, 1), color="white")
         img_byte_array = BytesIO()
         img.save(img_byte_array, format="PNG")
         mock_image_content = img_byte_array.getvalue()
@@ -55,6 +68,18 @@ def test_dalle_image_generation(self, mock_openai_create, mock_requests_get):
         mock_response.content = mock_image_content
         mock_requests_get.return_value = mock_response
 
+        # Set up the mock for OpenAI instance
+        mock_openai_instance = mock_openai.return_value
+        mock_openai_instance.images.generate.return_value = ImagesResponse(
+            data=[
+                Image(
+                    b64_json=None,
+                    revised_prompt=None,
+                    url="https://images.openai.com/1234.png",
+                )
+            ]
+        )
+
         function_name = "DallE"
 
         execute_query_fetch_all(self.evadb, f"DROP FUNCTION IF EXISTS {function_name};")
@@ -67,6 +92,6 @@ def test_dalle_image_generation(self, mock_openai_create, mock_requests_get):
         gpt_query = f"SELECT {function_name}(prompt) FROM ImageGen;"
         execute_query_fetch_all(self.evadb, gpt_query)
 
-        mock_openai_create.assert_called_once_with(
+        mock_openai_instance.images.generate.assert_called_once_with(
             prompt="a surreal painting of a cat", n=1, size="1024x1024"
         )