update openai version to v1

demo/animal_story.py

@@ -5,8 +5,8 @@
 import os
 import time
 
-import openai
 import requests
+from openai import OpenAI
 from PIL import Image
 from requests.adapters import HTTPAdapter  # type: ignore
 from requests.packages.urllib3.util.retry import Retry  # type: ignore
@@ -31,19 +31,30 @@ def prompt_generation(prompt: str, style) -> str:
     '''generate prompt for chatgpt based on the input'''
     logger = TokenLogger()
     logger.reset()
-    openai.api_key = os.getenv("OPENAI_API_KEY")
-    messages = [{"role": "user", "content": prompt}]
-    response = openai.ChatCompletion.create(
+    client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+    response = client.chat.completions.create(
         model="gpt-3.5-turbo",
-        messages=messages,
+        messages=[{
+            "role": "user",
+            "content": prompt
+        }],
         temperature=1.0,
         max_tokens=3000
     )
-    res = str(
-        f'Drawing style: {style}, ' +
-        response['choices'][0]['message']['content'][:1000]
-    )
-    token_usage = response['usage']['total_tokens']
+
+    if response.choices[0].message.content is not None:
+        res = str(
+            f'Drawing style: {style}, ' +
+            response.choices[0].message.content[:1000]
+        )
+    else:
+        res = "No content found"
+
+    if response.usage is not None:
+        token_usage = response.usage.total_tokens
+    else:
+        token_usage = 0
+
     logger.log(token_usage)
     return res
 

demo/auto_prompt_bot.py

@@ -1,6 +1,6 @@
 import os
 
-import openai
+from openai import OpenAI
 
 from yival.logger.token_logger import TokenLogger
 from yival.schemas.experiment_config import MultimodalOutput
@@ -12,25 +12,27 @@ def reply(input: str, state: ExperimentState) -> MultimodalOutput:
     logger = TokenLogger()
     logger.reset()
     # Ensure you have your OpenAI API key set up
-    openai.api_key = os.getenv("OPENAI_API_KEY")
+    client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
 
-    # Create a chat message sequence
-    messages = [{
-        "role":
-        "user",
-        "content":
-        str(StringWrapper("", name="prompt", state=state)) + f'\n{input}'
-    }]
     # Use the chat-based completion
-    response = openai.ChatCompletion.create(
-        model="gpt-3.5-turbo", messages=messages
+    response = client.chat.completions.create(
+        model="gpt-3.5-turbo",
+        messages=[{
+            "role":
+            "user",
+            "content":
+            str(StringWrapper("", name="prompt", state=state)) + f'\n{input}'
+        }]
     )
 
     # Extract the assistant's message (translated text) from the response
     answer = MultimodalOutput(
-        text_output=response['choices'][0]['message']['content'],
+        text_output=response.choices[0].message.content,
     )
-    token_usage = response['usage']['total_tokens']
+    if response.usage is not None:
+        token_usage = response.usage.total_tokens
+    else:
+        token_usage = 0
     logger.log(token_usage)
 
     return answer

demo/auto_reply/reply.py

@@ -4,11 +4,11 @@
 import uuid
 
 import faiss
-import openai
 from langchain.docstore import InMemoryDocstore
 from langchain.embeddings.openai import OpenAIEmbeddings
 from langchain.schema import Document
 from langchain.vectorstores import FAISS
+from openai import OpenAI
 
 from yival.logger.token_logger import TokenLogger
 from yival.schemas.experiment_config import MultimodalOutput
@@ -95,8 +95,7 @@ def reply(self, weibo_post: str, user_input: str, state: ExperimentState):
         logger = TokenLogger()
         logger.reset()
         # Ensure you have your OpenAI API key set up
-        openai.api_key = os.getenv("OPENAI_API_KEY")
-
+        client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
         documents = self.retriever.get_relevant_documents(user_input)
         query_context = ""
         if documents:
@@ -140,18 +139,16 @@ def reply(self, weibo_post: str, user_input: str, state: ExperimentState):
             )
         ) + suffix
 
-        # Create a chat message sequence
-        messages = [{"role": "user", "content": message}]
         # Use the chat-based completion
-        response = openai.ChatCompletion.create(
-            model="gpt-3.5-turbo", messages=messages
+        response = client.completions.create(
+            model="gpt-3.5-turbo", prompt=prompt
         )
 
         # Extract the assistant's message (translated text) from the response
         answer = MultimodalOutput(
-            text_output=response['choices'][0]['message']['content'],
+            text_output=response.choices[0].message.content,
         )
-        token_usage = response['usage']['total_tokens']
+        token_usage = response.usage.total_tokens
         logger.log(token_usage)
 
         return answer

demo/building_design.py

@@ -0,0 +1,100 @@
+'''This script is used to generate image from a building design prompt'''
+
+import os
+
+from openai import OpenAI
+
+from yival.logger.token_logger import TokenLogger
+from yival.schemas.experiment_config import MultimodalOutput
+from yival.states.experiment_state import ExperimentState
+from yival.wrappers.string_wrapper import StringWrapper
+
+
+def prompt_generation(prompt: str) -> str:
+    '''generate prompt for chatgpt based on the input'''
+    logger = TokenLogger()
+    logger.reset()
+    client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+    response = client.chat.completions.create(
+        model="gpt-3.5-turbo",
+        messages=[{
+            "role": "user",
+            "content": prompt
+        }],
+        temperature=1.0,
+        max_tokens=3000
+    )
+    if response.choices[0].message.content is not None:
+        res = str(response.choices[0].message.content[:1000])
+    else:
+        res = "No content found"
+
+    if response.usage is not None:
+        token_usage = response.usage.total_tokens
+    else:
+        token_usage = 0
+    logger.log(token_usage)
+    return res
+
+
+# def load_image(response):
+#     '''load image from response'''
+#     print("[INFO] start load images")
+#     url = f"{BASE_URL}/getImage"
+#     image_urls = response['response']['imageUrls']
+#     image_list = []
+#     for image_url in image_urls:
+#         payload = json.dumps({"imgUrl": image_url})
+#         response = s.post(url, headers=HEADERS, data=payload)
+#         if response.status_code == 200:
+#             image_data = response.content
+#             image = Image.open(io.BytesIO(image_data))
+#             image_list.append(image)
+#         else:
+#             print(
+#                 f"[Error] Failed to load image from {image_url}. Response code: {response.status_code}"
+#             )
+#     print("[INFO] Successfully load images.")
+
+#     return image_list
+
+
+def building_design(location: str, function: str, state: ExperimentState):
+    # Ensure you have your OpenAI API key set up
+    client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+    prompt = prompt_generation(
+        str(
+            StringWrapper(
+                "Generate a building design for a building",
+                name="task",
+                variables={
+                    "location": location,
+                    "function": function,
+                },
+                state=state
+            )
+        )
+    )
+    print(f"prompt: {prompt}")
+    response = client.images.generate(
+        model="dall-e-3", prompt=prompt, n=1, size="1024x1024"
+    )
+    print(f"response: {response}")
+    # image_res = MultimodalOutput(
+    #     text_output=response['response']['content'],
+    #     image_output=response['response']['imageUrls'],
+    # )
+    # return image_res
+    return response
+
+
+def main():
+    '''main function'''
+    location = "New York"
+    function = "office"
+    state = ExperimentState()
+    print(building_design(location, function, state))
+
+
+if __name__ == "__main__":
+    main()

demo/complete_task.py

@@ -3,6 +3,7 @@
 import time
 
 import openai
+from openai import OpenAI
 from tenacity import retry, stop_after_attempt, wait_random
 
 from yival.logger.token_logger import TokenLogger
@@ -13,7 +14,10 @@
 
 @retry(wait=wait_random(min=1, max=20), stop=stop_after_attempt(5))
 def completion_with_backpff(**kwargs):
-    response = openai.ChatCompletion.create(**kwargs)
+    request_timeout = kwargs.pop("request_timeout")
+    openai.timeout = request_timeout
+    client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+    response = client.chat.completions.create(**kwargs)
     return response
 
 
@@ -55,10 +59,8 @@ def complete_task(
         max_tokens=1000,
         request_timeout=20
     )
-    res = MultimodalOutput(
-        text_output=response['choices'][0]['message']['content'],
-    )
-    token_usage = response['usage']['total_tokens']
+    res = MultimodalOutput(text_output=response.choices[0].message.content, )
+    token_usage = response.usage.total_tokens
     logger.log(token_usage)
     return res
 

demo/configs/animal_story.yml

@@ -6,19 +6,18 @@ dataset:
       chunk_size: 1000
       diversify: true
       prompt:
-          "Please provide a concrete and realistic test case as a dictionary for function invocation using the ** operator.
-          Only include parameters, excluding description and name.
-          Ensure it's succinct and well-structured.
-          **Only provide the dictionary.**"
+        "Please provide a concrete and realistic test case as a dictionary for function invocation using the ** operator.
+        Only include parameters, excluding description and name.
+        Ensure it's succinct and well-structured.
+        **Only provide the dictionary.**"
       input_function:
-        description:
-          Given the species of an animal and its character, generate a corresponding story
+        description: Given the species of an animal and its character, generate a corresponding story
         name: animal_story_generation
         parameters:
           species: str
           character: str
           drawing_style: str
-      number_of_examples: 2
+      number_of_examples: 1
       model_name: gpt-4
   source_type: machine_generated
 
@@ -27,11 +26,11 @@ variations:
     generator_name: openai_prompt_based_variation_generator
     generator_config:
       model_name: gpt-4
-      number_of_variations: 2
+      number_of_variations: 1
       diversify: false
       max_tokens: 2000
       variables: null
-      prompt: 
+      prompt:
         - content: |-
             Your object is to construct a concise instruction prompt for GPT-4. This prompt will instruct GPT-4 as a gentle, imaginative children's writer to write all kinds of cute, kid-friendly stories based on animal species and personalities for an audience of YOUNG ADULTS and TEENAGERS.
             Points to emphasize in your instruction:
@@ -87,7 +86,6 @@ human_rating_configs:
     instructions: Rate the quality of the generated image.
     scale: [1, 5]
 
-
 selection_strategy:
   ahp_selection:
     criteria:
@@ -107,12 +105,12 @@ selection_strategy:
 enhancer:
   name: "optimize_by_prompt_enhancer"
   model_name: "gpt-4"
-  max_iterations: 2
+  max_iterations: 1
   enhance_var: ["task"]
   head_meta_instruction: |
     Now you will help me generate a prompt which is used to generate a corresponding
     story according to the species of an animal which is [animal_species] and its character [animal_character]. 
     I already have some prompt and its evaluation results :
-    
+
   end_meta_instruction: |
-    Give me a new prompt that is different from all pairs above, and has a evaluation value higher than any of above.
+    Give me a new prompt that is different from all pairs above, and has a evaluation value higher than any of above.