Upload 2 files

gpt4v.py

@@ -1,30 +1,28 @@
 import base64
 from openai import OpenAI
 
-
 class GPT4Vision:
     def __init__(self):
         self.client = OpenAI()
 
-    def encode_image(self, image_path):
+    def encode_image(self, image_file):
         """
         Encode the image to base64 format.
 
-        :param image_path: Path to the image file.
+        :param image_file: File object of the image.
         :return: Base64 encoded string of the image.
         """
-        with open(image_path, "rb") as image_file:
-            return base64.b64encode(image_file.read()).decode('utf-8')
+        return base64.b64encode(image_file.read()).decode('utf-8')
 
-    def describe(self, image_path, user_message):
+    def describe(self, image_file, user_message):
         """
         Get a description of the image using OpenAI's GPT-4 Vision API.
 
-        :param image_path: Path to the image file.
+        :param image_file: File object of the image.
         :param user_message: Custom text message to send as user input.
         :return: The API response.
         """
-        base64_image = self.encode_image(image_path)
+        base64_image = self.encode_image(image_file)
 
         response = self.client.chat.completions.create(
             model="gpt-4-vision-preview",
@@ -44,4 +42,12 @@ class GPT4Vision:
             ],
             max_tokens=1000,
         )
-        return response.choices[0]
+        return response.choices[0].message.content
+
+# # Example usage
+# gpt4v = GPT4Vision()
+# image_path = "/path/to/image.png"
+# with open(image_path, "rb") as image_file:
+#     user_message = "Describe this image for me."
+#     description = gpt4v.describe(image_file, user_message)
+#     print(description)