chinese_full_paragraph_evaluate_output_text_history = content_json["chinese_full_paragraph_evaluate_output_text"] if "chinese_full_paragraph_evaluate_output_text" in content_json else ""

app.py

@@ -1,8 +1,9 @@
-import os
-from docx import Document
+
 import gradio as gr
 from openai import OpenAI
 
+import os
+from docx import Document
 import random
 import uuid
 import json
@@ -12,12 +13,10 @@ import json
 import tempfile
 import urllib.parse
 import pandas as pd
-
+import time
 
 from storage_service import GoogleCloudStorage
 
-
-
 is_env_local = os.getenv("IS_ENV_LOCAL", "false") == "true"
 print(f"is_env_local: {is_env_local}")
 
@@ -28,11 +27,16 @@ if is_env_local:
         IS_ENV_PROD = "False"
         OPEN_AI_KEY = config["OPEN_AI_KEY"]
         GCS_KEY = json.dumps(config["GOOGLE_APPLICATION_CREDENTIALS_JSON"])
+        CUTOR_OPEN_AI_KEY = config["CUTOR_OPEN_AI_KEY"]
+        CUTOR_OPEN_AI_ASSISTANT_ID = config["CUTOR_OPEN_AI_ASSISTANT_ID"]
 else:
     OPEN_AI_KEY = os.getenv("OPEN_AI_KEY")
     GCS_KEY = os.getenv("GOOGLE_APPLICATION_CREDENTIALS_JSON")
+    CUTOR_OPEN_AI_KEY = os.getenv("CUTOR_OPEN_AI_KEY")
+    CUTOR_OPEN_AI_ASSISTANT_ID = os.getenv("CUTOR_ACUTOR_OPEN_AI_ASSISTANT_IDSSISTANT_ID")
 
 OPEN_AI_CLIENT = OpenAI(api_key=OPEN_AI_KEY)
+CUTOR_OPEN_AI_CLIENT = OpenAI(api_key=CUTOR_OPEN_AI_KEY)
 
 # 设置 Google Cloud Storage 客户端
 GCS_SERVICE = GoogleCloudStorage(GCS_KEY)
@@ -845,46 +849,46 @@ def get_past_exam_practice_log_session_content(file_name):
             past_exam_save_output
 
 # === Chinese ===
-def generate_chinese_evaluation_table(model, sys_content, user_prompt, text):
-    # https://www.ceec.edu.tw/files/file_pool/1/0j052575870800204600/1216%E5%9C%8B%E6%96%87%E4%BD%9C%E6%96%87%E5%88%86%E9%A0%85%E5%BC%8F%E8%A9%95%E5%88%86%E6%8C%87%E6%A8%99.pdf
-
-    user_content = f"""
-        本篇作文：{text}
-        ---
-        {user_prompt}
-    """
-    messages = [
-        {"role": "system", "content": sys_content},
-        {"role": "user", "content": user_content}
-    ]
-
-    response_format = { "type": "json_object" }
-
-    request_payload = {
-        "model": model,
-        "messages": messages,
-        "max_tokens": 2000,
-        "response_format": response_format
-    }
-
-    response = OPEN_AI_CLIENT.chat.completions.create(**request_payload)
-    content = response.choices[0].message.content
-
-    print(f"====generate_chinese_evaluation_table====")
-    print(content)
-
-    data = json.loads(content)["results"]
-    headers = ["架構", "評分", "解釋"]
-    table_data = [
-        ["主題與內容", data['主題與內容']['level'], data['主題與內容']['explanation']],
-        ["段落結構", data['段落結構']['level'], data['段落結構']['explanation']],
-        ["遣詞造句", data['遣詞造句']['level'], data['遣詞造句']['explanation']],
-        ["錯別字與標點符號", data['錯別字與標點符號']['level'], data['錯別字與標點符號']['explanation']]
-    ]
+# def generate_chinese_evaluation_table(model, sys_content, user_prompt, text):
+#     # https://www.ceec.edu.tw/files/file_pool/1/0j052575870800204600/1216%E5%9C%8B%E6%96%87%E4%BD%9C%E6%96%87%E5%88%86%E9%A0%85%E5%BC%8F%E8%A9%95%E5%88%86%E6%8C%87%E6%A8%99.pdf
+
+#     user_content = f"""
+#         本篇作文：{text}
+#         ---
+#         {user_prompt}
+#     """
+#     messages = [
+#         {"role": "system", "content": sys_content},
+#         {"role": "user", "content": user_content}
+#     ]
+
+#     response_format = { "type": "json_object" }
+
+#     request_payload = {
+#         "model": model,
+#         "messages": messages,
+#         "max_tokens": 2000,
+#         "response_format": response_format
+#     }
+
+#     response = OPEN_AI_CLIENT.chat.completions.create(**request_payload)
+#     content = response.choices[0].message.content
+
+#     print(f"====generate_chinese_evaluation_table====")
+#     print(content)
+
+#     data = json.loads(content)["results"]
+#     headers = ["架構", "評分", "解釋"]
+#     table_data = [
+#         ["主題與內容", data['主題與內容']['level'], data['主題與內容']['explanation']],
+#         ["段落結構", data['段落結構']['level'], data['段落結構']['explanation']],
+#         ["遣詞造句", data['遣詞造句']['level'], data['遣詞造句']['explanation']],
+#         ["錯別字與標點符號", data['錯別字與標點符號']['level'], data['錯別字與標點符號']['explanation']]
+#     ]
     
-    gr_update = gr.update(value=table_data, headers=headers)
+#     gr_update = gr.update(value=table_data, headers=headers)
 
-    return gr_update
+#     return gr_update
 
 def load_exam_data():
     with open("exams.json", "r") as file:
@@ -902,6 +906,7 @@ def generate_chinese_paragraph_practice_history(
     session_timestamp,
     request_origin,
     chinese_full_paragraph_input,
+    chinese_full_paragraph_evaluate_output_text,
     chinese_full_paragraph_evaluate_output,
     chinese_full_paragraph_correct_grammatical_spelling_errors_input,
     chinese_full_paragraph_correct_grammatical_spelling_errors_output_table,
@@ -918,6 +923,7 @@ def generate_chinese_paragraph_practice_history(
             "session_timestamp": session_timestamp,
             "request_origin": request_origin,
             "full_paragraph_input": chinese_full_paragraph_input,
+            "chinese_full_paragraph_evaluate_output_text": chinese_full_paragraph_evaluate_output_text,
             "full_paragraph_evaluate_output": chinese_full_paragraph_evaluate_output.to_dict(orient='records'),
             "full_paragraph_correct_grammatical_spelling_errors_input": chinese_full_paragraph_correct_grammatical_spelling_errors_input,
             "full_paragraph_correct_grammatical_spelling_errors_output_table": chinese_full_paragraph_correct_grammatical_spelling_errors_output_table.to_dict(orient='records'),
@@ -929,6 +935,7 @@ def generate_chinese_paragraph_practice_history(
         GCS_SERVICE.upload_json_string("jutor_logs", file_name, json.dumps(content))
 
     return chinese_full_paragraph_input, \
+            chinese_full_paragraph_evaluate_output_text, \
             chinese_full_paragraph_evaluate_output, \
             chinese_full_paragraph_correct_grammatical_spelling_errors_input, \
             chinese_full_paragraph_correct_grammatical_spelling_errors_output_table, \
@@ -937,14 +944,13 @@ def generate_chinese_paragraph_practice_history(
             chinese_full_paragraph_refine_output, \
             chinese_full_paragraph_save_output
 
-
-
 def get_chinese_paragraph_practice_log_session_content(file_name):
     if file_name:
         content = GCS_SERVICE.download_as_string("jutor_logs", file_name)
         print(f"content: {content}")
         content_json = json.loads(content)
         chinese_full_paragraph_input_history = content_json["full_paragraph_input"]
+        chinese_full_paragraph_evaluate_output_text_history = content_json["chinese_full_paragraph_evaluate_output_text"] if "chinese_full_paragraph_evaluate_output_text" in content_json else ""
         chinese_full_paragraph_evaluate_output_history = pd.DataFrame(content_json["full_paragraph_evaluate_output"])
         chinese_full_paragraph_correct_grammatical_spelling_errors_input_history = content_json["full_paragraph_correct_grammatical_spelling_errors_input"]
         chinese_full_paragraph_correct_grammatical_spelling_errors_output_table_history = pd.DataFrame(content_json["full_paragraph_correct_grammatical_spelling_errors_output_table"])
@@ -954,6 +960,7 @@ def get_chinese_paragraph_practice_log_session_content(file_name):
         chinese_full_paragraph_save_output_history = content_json["full_paragraph_save_output"]
     else:
         chinese_full_paragraph_input_history = ""
+        chinese_full_paragraph_evaluate_output_text_history = ""
         chinese_full_paragraph_evaluate_output_history = pd.DataFrame()
         chinese_full_paragraph_correct_grammatical_spelling_errors_input_history = ""
         chinese_full_paragraph_correct_grammatical_spelling_errors_output_table_history = pd.DataFrame()
@@ -963,6 +970,7 @@ def get_chinese_paragraph_practice_log_session_content(file_name):
         chinese_full_paragraph_save_output_history = ""
 
     return chinese_full_paragraph_input_history, \
+            chinese_full_paragraph_evaluate_output_text_history, \
             chinese_full_paragraph_evaluate_output_history, \
             chinese_full_paragraph_correct_grammatical_spelling_errors_input_history, \
             chinese_full_paragraph_correct_grammatical_spelling_errors_output_table_history, \
@@ -971,6 +979,131 @@ def get_chinese_paragraph_practice_log_session_content(file_name):
             chinese_full_paragraph_refine_output_history, \
             chinese_full_paragraph_save_output_history
 
+
+# === OpenAI Assistant ===
+def get_chinese_paragraph_evaluate_content(model, sys_content, paragraph, user_generate_paragraph_evaluate_prompt):
+    user_content = f"""
+        sys_content: {sys_content}
+        ---
+        paragraph is: {paragraph}
+        ---
+        {user_generate_paragraph_evaluate_prompt}
+    """
+    content = generate_content_by_open_ai_assistant(user_content, thread_id=None, model_name=model)
+    print(f"====generate_paragraph_evaluate====")
+    print(content)
+
+    content_list = content.split("```json")
+    content_text = content_list[0]
+    content_json = content_list[1].split("```")[0].replace("```", "")
+
+    data = json.loads(content_json)["results"]
+    headers = ["架構", "評分", "解釋"]
+    table_data = [
+        ["主題與內容", data['主題與內容']['level'], data['主題與內容']['explanation']],
+        ["段落結構", data['段落結構']['level'], data['段落結構']['explanation']],
+        ["遣詞造句", data['遣詞造句']['level'], data['遣詞造句']['explanation']],
+        ["錯別字與標點符號", data['錯別字與標點符號']['level'], data['錯別字與標點符號']['explanation']]
+    ]
+    
+    gr_update = gr.update(value=table_data, headers=headers, visible=True)
+
+    return content_text, gr_update
+
+
+def generate_content_by_open_ai_assistant(user_content, thread_id=None, model_name=None):
+    client = CUTOR_OPEN_AI_CLIENT
+    assistant_id = CUTOR_OPEN_AI_ASSISTANT_ID
+
+    try:
+        if not thread_id:
+            thread = client.beta.threads.create()
+            thread_id = thread.id
+        else:
+            thread = client.beta.threads.retrieve(thread_id)
+
+        print(f"Thread ID: {thread.id}")
+
+        # if metadata:
+        #     client.beta.threads.update(thread_id=thread.id, metadata=metadata)
+
+        # Send the user message to the thread
+        print("==============Send the user message to the thread====================")
+        client.beta.threads.messages.create(thread_id=thread.id, role="user", content=user_content)
+
+        # Run the assistant
+        print("==============Run the assistant====================")
+        run = client.beta.threads.runs.create_and_poll(thread_id=thread.id, assistant_id=assistant_id)
+        
+        # Wait for the response
+        # run_status = poll_run_status(run.id, thread.id, timeout=30)
+        # print(f"Run status: {run_status}")
+        
+        if run.status == "completed":
+            print("==============completed====================")
+            print(f"Thread ID: {thread.id}")
+            messages = client.beta.threads.messages.list(thread_id=thread.id)
+            print(f"Messages: {messages}")
+            response = messages
+            response_text = messages.data[0].content[0].text.value
+            print(f"Response: {response_text}")
+    except Exception as e:
+        print(f"Error: {e}")
+        response_text = " 出現錯誤，請稍後再試！"
+
+    return response_text
+
+def poll_run_status(run_id, thread_id, timeout=600, poll_interval=5):
+    """
+    Polls the status of a Run and handles different statuses appropriately.
+
+    :param run_id: The ID of the Run to poll.
+    :param thread_id: The ID of the Thread associated with the Run.
+    :param timeout: Maximum time to wait for the Run to complete, in seconds.
+    :param poll_interval: Time to wait between each poll, in seconds.
+    """
+    client = OPEN_AI_CLIENT
+    start_time = time.time()
+    while time.time() - start_time < timeout:
+        run = client.beta.threads.runs.retrieve(thread_id=thread_id, run_id=run_id)
+
+        if run.status in ["completed", "cancelled", "failed"]:
+            print(f"Run completed with status: {run.status}")
+            break
+        elif run.status == "requires_action":
+            print("Run requires action. Performing required action...")
+            # Here, you would perform the required action, e.g., running functions
+            # and then submitting the outputs. This is simplified for this example.
+            # After performing the required action, you'd complete the action:
+            # OPEN_AI_CLIENT.beta.threads.runs.complete_required_action(...)
+        elif run.status == "expired":
+            print("Run expired. Exiting...")
+            break
+        else:
+            print(f"Run status is {run.status}. Waiting for updates...")
+        
+        time.sleep(poll_interval)
+    else:
+        print("Timeout reached. Run did not complete in the expected time.")
+
+    # Once the Run is completed, handle the result accordingly
+    if run.status == "completed":
+        # Retrieve and handle messages or run steps as needed
+        messages = client.beta.threads.messages.list(thread_id=thread_id)
+        for message in messages.data:
+            if message.role == "assistant":
+                print(f"Assistant response: {message.content}")
+    elif run.status in ["cancelled", "failed"]:
+        # Handle cancellation or failure
+        print(f"Run ended with status: {run.status}")
+    elif run.status == "expired":
+        # Handle expired run
+        print("Run expired without completion.")
+    
+    return run.status
+
+
+
 # 小工具
 def show_elements():
     return gr.update(visible=True)
@@ -2674,26 +2807,50 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
 
                         6. 批改回饋的最後請引用原文給予至少三個詞語或是句型改寫的範例，列出原文和修改的版本，修改的案例應該和前面四個面向的批改內容前後呼應。
 
+                        7. 將評分標準與回饋的內容以JSON格式輸出，並且請使用繁體中文（ZH-TW）來評分段落並輸出，用 ```json ..... ``` 包裹:
+
                         please use Chinese language (ZH-TW) to evaluate the paragraph and output use JSON format:
                         EXAMPLE:
-                        "results": {{
-                            "主題與內容": {{
-                                "level": "A+",
-                                "explanation": "#中文解釋 ZH-TW"
-                            }},
-                            "段落結構": {{
-                                "level": "B+",
-                                "explanation": "#中文解釋 ZH-TW"
-                            }},
-                            "遣詞造句": {{
-                                "level": "C",
-                                "explanation": "#中文解釋 ZH-TW"
-                            }},
-                            "錯別字與標點符號": {{
-                                "level": "C-",
-                                "explanation": "#中文解釋 ZH-TW"
+                        很可惜，你的作文有點短，你可以參考下面的建議，讓自己的字數不少於200字。
+
+                        # 評分標準與回饋
+                        主題與內容：B+ 你的主題很明確，講述了CSS在渲染空間上的問題及解決方案，這是一個重要而實用的話題。然而，內容相對較少，缺乏足夠的細節與實例來支撐你的觀點。建議你可以添加一些具體情境或例子，讓讀者更容易理解CSS的應用情況，例如提到常見的渲染問題以及具體的解決方法。
+
+                        段落結構：B 你的段落結構基本清晰，但目前只有一段，這使得整體文章顯得有些單薄。建議你可以將內容分成幾個小段落，每個段落著重於不同的要點，例如一段說明問題，另一段探討解決方案，這樣整體更具條理性。
+
+                        遣詞造句：A 你的遣詞造句大致良好，用詞得體且通順。不過可以嘗試加一些更具體的技術詞彙或示例，使文章更專業化。
+
+                        錯別字與標點符號：A+ 你的文章沒有明顯的錯別字，並且標點符號的使用也相當正確，這點做得很好。
+
+                        # 修改範例
+                        原文：內容雖然簡短，但主題明確。 修改：雖然內容相對簡短，但主題表達得非常明確。
+
+                        原文：缺乏實例和具體情境來支持內容。 修改：目前缺少具體的實例及情境來支持文章的內容與主張。
+
+                        原文：可以進一步擴展。 修改：可以進一步擴展來豐富內容，讓讀者更有共鳴。
+
+                        ```json
+                        {{
+                            "results": {{
+                                "主題與內容": {{
+                                    "level": "A+",
+                                    "explanation": "#中文解釋 ZH-TW"
+                                }},
+                                "段落結構": {{
+                                    "level": "B+",
+                                    "explanation": "#中文解釋 ZH-TW"
+                                }},
+                                "遣詞造句": {{
+                                    "level": "C",
+                                    "explanation": "#中文解釋 ZH-TW"
+                                }},
+                                "錯別字與標點符號": {{
+                                    "level": "C-",
+                                    "explanation": "#中文解釋 ZH-TW"
+                                }}
                             }}
                         }}
+                        ```
 
                         Restrictions:
                         - ALL the content should be in Traditional Chinese (zh-TW), it's very important.
@@ -2708,10 +2865,10 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                         with gr.Row():
                             chinese_full_paragraph_evaluate_button = gr.Button("段落全文分析", variant="primary")
                         with gr.Row():
-                            chinese_full_paragraph_evaluate_output = gr.Dataframe(label="段落全文分析", wrap=True, column_widths=[20, 15, 65], interactive=False)
+                            chinese_full_paragraph_evaluate_output_text = gr.Markdown(label="段落全文分析")
+                        with gr.Row():
+                            chinese_full_paragraph_evaluate_output = gr.Dataframe(label="段落全文分析表格", wrap=True, column_widths=[20, 15, 65], interactive=False)
 
-                            
-                
                 # JUTOR 段落批改與整體建議
                 with gr.Row():
                     gr.Markdown("# JUTOR 段落批改與整體建議")
@@ -2757,6 +2914,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                         gr.Markdown("<span style='color:#4e80ee'>輸入段落全文</span>")
                         chinese_full_paragraph_input_history = gr.Markdown()
                         gr.Markdown("<span style='color:#4e80ee'>段落全文分析</span>")
+                        chinese_full_paragraph_evaluate_output_text_history = gr.Markdown()
                         chinese_full_paragraph_evaluate_output_history = gr.Dataframe(wrap=True, column_widths=[35, 15, 50], interactive=False)
 
                         gr.Markdown("<span style='color:#4e80ee'>修訂文法與拼字錯誤 輸入</span>")
@@ -2781,6 +2939,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                             session_timestamp,
                             request_origin,
                             chinese_full_paragraph_input,
+                            chinese_full_paragraph_evaluate_output_text,
                             chinese_full_paragraph_evaluate_output,
                             chinese_full_paragraph_correct_grammatical_spelling_errors_input,
                             chinese_full_paragraph_correct_grammatical_spelling_errors_output_table,
@@ -2791,6 +2950,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                         ],
                         outputs=[
                             chinese_full_paragraph_input_history,
+                            chinese_full_paragraph_evaluate_output_text_history,
                             chinese_full_paragraph_evaluate_output_history,
                             chinese_full_paragraph_correct_grammatical_spelling_errors_input_history,
                             chinese_full_paragraph_correct_grammatical_spelling_errors_output_table_history,
@@ -2805,10 +2965,19 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                         outputs=[chinese_grapragh_practice_history_accordion]
                     )
 
+                # chinese_full_paragraph_evaluate_button.click(
+                #     fn=generate_chinese_evaluation_table,
+                #     inputs=[model, chinese_full_paragraph_sys_content_input, user_generate_chinese_full_paragraph_evaluate_prompt, chinese_full_paragraph_input],
+                #     outputs=chinese_full_paragraph_evaluate_output
+                # ).then(
+                #     fn=update_paragraph_correct_grammatical_spelling_errors_input,
+                #     inputs=[chinese_full_paragraph_input],
+                #     outputs=chinese_full_paragraph_correct_grammatical_spelling_errors_input
+                # )
                 chinese_full_paragraph_evaluate_button.click(
-                    fn=generate_chinese_evaluation_table,
-                    inputs=[model, chinese_full_paragraph_sys_content_input, user_generate_chinese_full_paragraph_evaluate_prompt, chinese_full_paragraph_input],
-                    outputs=chinese_full_paragraph_evaluate_output
+                    fn=get_chinese_paragraph_evaluate_content,
+                    inputs=[model, chinese_full_paragraph_sys_content_input, chinese_full_paragraph_input, user_generate_chinese_full_paragraph_evaluate_prompt],
+                    outputs=[chinese_full_paragraph_evaluate_output_text, chinese_full_paragraph_evaluate_output]
                 ).then(
                     fn=update_paragraph_correct_grammatical_spelling_errors_input,
                     inputs=[chinese_full_paragraph_input],
@@ -2859,6 +3028,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                             gr.Markdown("<span style='color:#4e80ee'>輸入段落全文</span>")
                             chinese_full_paragraph_input_history_log = gr.Markdown()
                             gr.Markdown("<span style='color:#4e80ee'>段落全文分析</span>")
+                            chinese_full_paragraph_evaluate_output_text_history_log = gr.Markdown()
                             chinese_full_paragraph_evaluate_output_history_log = gr.Dataframe(wrap=True, column_widths=[35, 15, 50], interactive=False)
 
                             gr.Markdown("<span style='color:#4e80ee'>修訂文法與拼字錯誤 輸入</span>")
@@ -2887,6 +3057,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                                 inputs=[chinese_paragraph_practice_logs_session_list],
                                 outputs=[
                                     chinese_full_paragraph_input_history_log,
+                                    chinese_full_paragraph_evaluate_output_text_history_log,
                                     chinese_full_paragraph_evaluate_output_history_log,
                                     chinese_full_paragraph_correct_grammatical_spelling_errors_input_history_log,
                                     chinese_full_paragraph_correct_grammatical_spelling_errors_output_table_history_log,