genai model

app.py

@@ -19,6 +19,11 @@ from google.oauth2.service_account import Credentials
 import vertexai
 from vertexai.generative_models import GenerativeModel
 
+from google import genai
+from google.genai import types
+import base64
+import io
+
 # From other files
 from storage_service import GoogleCloudStorage
 from assignment_ui import create_assignment_ui
@@ -69,6 +74,13 @@ vertexai.init(
     credentials=google_creds,
 )
 
+GENAI_CLIENT = genai.Client(
+    vertexai=True,
+    project="junyiacademy",
+    location='us-central1',
+    credentials=google_creds
+)
+
 _AssignmentService = AssignmentService(GCS_SERVICE)
 _SubmissionService = SubmissionService(GCS_SERVICE)
 _DashboardService = DashboardService(_AssignmentService, _SubmissionService)
@@ -2272,7 +2284,69 @@ def download_content(content):
 # OCR
 def extract_text_from_image(image):
     """從上傳的圖片中提取文字"""
-    return "test"
+    
+    # 檢查是否有圖片
+    if image is None:
+        return ""
+        
+    # 將圖片轉換為 base64 編碼
+    buffered = io.BytesIO()
+    image.save(buffered, format="JPEG")
+    img_str = base64.b64encode(buffered.getvalue()).decode()
+
+    # 準備圖片輸入
+    image_part = types.Part.from_bytes(
+        data=base64.b64decode(img_str),
+        mime_type="image/jpeg",
+    )
+
+    # 設定模型和內容
+    model = "gemini-2.0-pro-exp-02-05"
+    contents = [
+        types.Content(
+            role="user",
+            parts=[
+                image_part,
+                types.Part.from_text(text="verify the text and return the text only. do not return any other text and check the text is correct or not.")
+            ]
+        )
+    ]
+
+    # 設定生成配置
+    generate_content_config = types.GenerateContentConfig(
+        temperature=1,
+        top_p=0.95,
+        max_output_tokens=8192,
+        response_modalities=["TEXT"],
+        safety_settings=[
+            types.SafetySetting(
+                category="HARM_CATEGORY_HATE_SPEECH",
+                threshold="OFF"
+            ),
+            types.SafetySetting(
+                category="HARM_CATEGORY_DANGEROUS_CONTENT",
+                threshold="OFF"
+            ),
+            types.SafetySetting(
+                category="HARM_CATEGORY_SEXUALLY_EXPLICIT",
+                threshold="OFF"
+            ),
+            types.SafetySetting(
+                category="HARM_CATEGORY_HARASSMENT",
+                threshold="OFF"
+            )
+        ],
+    )
+
+    # 生成內容
+    response = GENAI_CLIENT.models.generate_content(
+        model=model,
+        contents=contents,
+        config=generate_content_config,
+    )
+
+    # 返回識別的文字
+    return response.text
 
     # 使用 OCR 工具
 def extract_text_from_file(file):
@@ -3947,7 +4021,7 @@ with gr.Blocks(theme=THEME, css=CSS) as demo:
                     with gr.Column():
                         chinese_full_paragraph_input = gr.Textbox(label="輸入段落全文", lines=5)
                         with gr.Row():
-                            with gr.Column(scale=1):
+                            with gr.Column(scale=1, visible=False):
                                 # 修正文件類型設置
                                 chinese_file_upload = gr.File(
                                     label="上傳文件", 

requirements.txt

@@ -6,4 +6,5 @@ google-auth-httplib2
 google-auth-oauthlib
 google-cloud-storage
 google-cloud-bigquery
-vertexai
+vertexai
+google-genai