Update app.py

app.py

@@ -13,6 +13,7 @@ import pymorphy2
 import re
 import string
 import io
+from datetime import datetime
 
 morph = pymorphy2.MorphAnalyzer()
 
@@ -59,6 +60,74 @@ for sheet_name, df in data.items():
 current_request_index = -1  # Изначально указывает на последний запрос
 
 
+def download_current_message_database():
+    # GitHub repositories and authors mapping
+    repos = {
+        'Storage_1': 'Редакторы',
+        'Storage_Ira': 'Ира',
+        'Storage_Kate': 'Катя',
+        'Storage_Sveta': 'Света'
+    }
+
+    # Base GitHub API URL
+    base_url = 'https://api.github.com/repos/fruitpicker01/{repo}/contents'
+
+    data_list = []
+
+    headers = {
+        "Authorization": f"token {token}",  # 'token' is from your environment variables
+        "Content-Type": "application/json"
+    }
+
+    for repo_name, author in repos.items():
+        url = base_url.format(repo=repo_name)
+        response = requests.get(url, headers=headers)
+        if response.status_code == 200:
+            files = response.json()
+            # Filter files starting with 'file'
+            json_files = [file for file in files if file['name'].startswith("file") and file['name'].endswith('.json')]
+            for file_info in json_files:
+                file_name = file_info['name']
+                file_url = file_info['download_url']
+
+                # Extract timestamp from filename
+                try:
+                    timestamp = file_name.split('_')[1].split('.')[0]
+                    save_date = datetime.utcfromtimestamp(int(timestamp)).strftime('%Y-%m-%d %H:%M:%S')
+                except:
+                    save_date = None  # or set to some default
+
+                # Download and decode file content
+                file_response = requests.get(file_url)
+                if file_response.status_code == 200:
+                    data = json.loads(file_response.text)
+                    normalized_data = pd.json_normalize(data)
+                    normalized_data['Дата сохранения'] = save_date
+                    normalized_data['Автор'] = author
+                    data_list.append(normalized_data)
+                else:
+                    print(f"Error downloading file {file_name} from repo {repo_name}: {file_response.status_code}")
+        else:
+            print(f"Error accessing repo {repo_name}: {response.status_code}")
+
+    if data_list:
+        df = pd.concat(data_list, ignore_index=True)
+        # Convert 'Дата сохранения' to datetime
+        df['Дата сохранения'] = pd.to_datetime(df['Дата сохранения'], format='%Y-%m-%d %H:%M:%S', errors='coerce')
+        # Drop duplicates
+        df.drop_duplicates(inplace=True)
+        # Sort by 'Дата сохранения' descending
+        df.sort_values(by='Дата сохранения', ascending=False, inplace=True)
+        # Create Excel file in memory
+        output = io.BytesIO()
+        df.to_excel(output, index=False)
+        output.seek(0)  # Reset pointer
+        # Return as a tuple with the filename
+        return ('messages.xlsx', output)
+    else:
+        return None
+
+
 def correct_dash_usage(text):
     # Step 1: Replace any dash with long dash if surrounded by spaces
     text = re.sub(r'\s[-–—]\s', ' — ', text)
@@ -1781,6 +1850,10 @@ with gr.Blocks() as demo:
 
     gr.Markdown("---")
 
+    with gr.Row():
+        download_btn = gr.Button("Выгрузить актуальную базу сообщений")
+        download_file = gr.File(label="Скачать базу сообщений")
+
     # Очистка всех полей кроме prompt_display
     description_input.change(
         fn=clear_on_change,  # Сначала вызываем функцию очистки полей
@@ -2442,5 +2515,10 @@ with gr.Blocks() as demo:
         outputs=[save_meta_llama_405b_btn]
     )
 
+    download_btn.click(
+        fn=download_current_message_database,
+        inputs=[],
+        outputs=download_file
+    )
 
     demo.launch()