Update app.py
Browse files
app.py
CHANGED
@@ -10,49 +10,54 @@ from langchain.vectorstores import Chroma
|
|
10 |
from langchain.embeddings.openai import OpenAIEmbeddings
|
11 |
from langchain.chat_models import ChatOpenAI
|
12 |
import shutil # 用於文件複製
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
# 獲取 OpenAI API 密鑰
|
15 |
api_key = os.getenv("OPENAI_API_KEY")
|
16 |
if not api_key:
|
17 |
raise ValueError("未能獲取 OPENAI_API_KEY。請在 Hugging Face Spaces 的 Secrets 中設置它。")
|
18 |
openai.api_key = api_key
|
19 |
-
|
20 |
|
21 |
# 確保向量資料庫目錄存在且有寫入權限
|
22 |
VECTORDB_DIR = os.path.abspath("./data")
|
23 |
os.makedirs(VECTORDB_DIR, exist_ok=True)
|
24 |
os.chmod(VECTORDB_DIR, 0o755) # 設置適當的權限
|
25 |
-
|
26 |
|
27 |
# 定義測試 PDF 加載器的函數
|
28 |
def test_pdf_loader(file_path, loader_type='PyMuPDFLoader'):
|
29 |
-
|
30 |
try:
|
31 |
if loader_type == 'PyMuPDFLoader':
|
32 |
loader = PyMuPDFLoader(file_path)
|
33 |
elif loader_type == 'PyPDFLoader':
|
34 |
loader = PyPDFLoader(file_path)
|
35 |
else:
|
36 |
-
|
37 |
return
|
38 |
loaded_docs = loader.load()
|
39 |
if loaded_docs:
|
40 |
-
|
41 |
-
|
42 |
else:
|
43 |
-
|
44 |
except Exception as e:
|
45 |
-
|
46 |
|
47 |
# 定義載入和處理 PDF 文件的函數
|
48 |
def load_and_process_documents(file_paths, loader_type='PyMuPDFLoader'):
|
49 |
documents = []
|
50 |
-
|
51 |
|
52 |
for file_path in file_paths:
|
53 |
-
|
54 |
if not os.path.exists(file_path):
|
55 |
-
|
56 |
continue
|
57 |
try:
|
58 |
if loader_type == 'PyMuPDFLoader':
|
@@ -60,28 +65,28 @@ def load_and_process_documents(file_paths, loader_type='PyMuPDFLoader'):
|
|
60 |
elif loader_type == 'PyPDFLoader':
|
61 |
loader = PyPDFLoader(file_path)
|
62 |
else:
|
63 |
-
|
64 |
continue
|
65 |
loaded_docs = loader.load()
|
66 |
if loaded_docs:
|
67 |
-
|
68 |
# 打印第一個文檔的部分內容以確認
|
69 |
-
|
70 |
documents.extend(loaded_docs)
|
71 |
else:
|
72 |
-
|
73 |
except Exception as e:
|
74 |
-
|
75 |
|
76 |
if not documents:
|
77 |
raise ValueError("沒有找到任何 PDF 文件或 PDF 文件無法載入。")
|
78 |
else:
|
79 |
-
|
80 |
|
81 |
# 分割長文本
|
82 |
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=50)
|
83 |
documents = text_splitter.split_documents(documents)
|
84 |
-
|
85 |
|
86 |
if not documents:
|
87 |
raise ValueError("分割後的文檔列表為空。請檢查 PDF 文件內容。")
|
@@ -89,7 +94,7 @@ def load_and_process_documents(file_paths, loader_type='PyMuPDFLoader'):
|
|
89 |
# 初始化向量資料庫
|
90 |
try:
|
91 |
embeddings = OpenAIEmbeddings(openai_api_key=api_key) # 直接傳遞 API 密鑰
|
92 |
-
|
93 |
except Exception as e:
|
94 |
raise ValueError(f"初始化 OpenAIEmbeddings 時出現錯誤: {e}")
|
95 |
|
@@ -99,7 +104,7 @@ def load_and_process_documents(file_paths, loader_type='PyMuPDFLoader'):
|
|
99 |
embedding=embeddings,
|
100 |
persist_directory=VECTORDB_DIR
|
101 |
)
|
102 |
-
|
103 |
except Exception as e:
|
104 |
raise ValueError(f"初始化 Chroma 向量資料庫時出現錯誤: {e}")
|
105 |
|
@@ -136,36 +141,48 @@ def handle_query(user_message, chat_history, vectordb):
|
|
136 |
return chat_history
|
137 |
|
138 |
except Exception as e:
|
|
|
139 |
return chat_history + [("系統", f"出現錯誤: {str(e)}")]
|
140 |
|
141 |
# 定義 Gradio 的處理函數
|
142 |
def process_files(files, state):
|
143 |
-
|
144 |
if files:
|
145 |
try:
|
146 |
-
|
147 |
saved_file_paths = []
|
148 |
for file in files:
|
149 |
-
|
150 |
save_path = os.path.join(VECTORDB_DIR, file.name)
|
151 |
with open(save_path, "wb") as f:
|
152 |
f.write(file.read())
|
153 |
-
#
|
154 |
if os.path.exists(save_path):
|
155 |
-
|
|
|
|
|
|
|
|
|
|
|
156 |
else:
|
157 |
-
|
|
|
158 |
saved_file_paths.append(save_path)
|
159 |
-
# 測試 PDF
|
160 |
-
|
|
|
|
|
|
|
|
|
|
|
161 |
# 列出 VECTORDB_DIR 中的所有文件
|
162 |
saved_files = os.listdir(VECTORDB_DIR)
|
163 |
-
|
164 |
vectordb = load_and_process_documents(saved_file_paths, loader_type='PyMuPDFLoader')
|
165 |
state['vectordb'] = vectordb
|
166 |
return "PDF 文件已成功上傳並處理。您現在可以開始提問。", state
|
167 |
except Exception as e:
|
168 |
-
|
169 |
return f"處理文件時出現錯誤: {e}", state
|
170 |
else:
|
171 |
return "請上傳至少一個 PDF 文件。", state
|
|
|
10 |
from langchain.embeddings.openai import OpenAIEmbeddings
|
11 |
from langchain.chat_models import ChatOpenAI
|
12 |
import shutil # 用於文件複製
|
13 |
+
import logging
|
14 |
+
|
15 |
+
# 設置日誌配置
|
16 |
+
logging.basicConfig(level=logging.INFO)
|
17 |
+
logger = logging.getLogger(__name__)
|
18 |
|
19 |
# 獲取 OpenAI API 密鑰
|
20 |
api_key = os.getenv("OPENAI_API_KEY")
|
21 |
if not api_key:
|
22 |
raise ValueError("未能獲取 OPENAI_API_KEY。請在 Hugging Face Spaces 的 Secrets 中設置它。")
|
23 |
openai.api_key = api_key
|
24 |
+
logger.info("OpenAI API 密鑰已設置。")
|
25 |
|
26 |
# 確保向量資料庫目錄存在且有寫入權限
|
27 |
VECTORDB_DIR = os.path.abspath("./data")
|
28 |
os.makedirs(VECTORDB_DIR, exist_ok=True)
|
29 |
os.chmod(VECTORDB_DIR, 0o755) # 設置適當的權限
|
30 |
+
logger.info(f"VECTORDB_DIR set to: {VECTORDB_DIR}")
|
31 |
|
32 |
# 定義測試 PDF 加載器的函數
|
33 |
def test_pdf_loader(file_path, loader_type='PyMuPDFLoader'):
|
34 |
+
logger.info(f"Testing PDF loader ({loader_type}) with file: {file_path}")
|
35 |
try:
|
36 |
if loader_type == 'PyMuPDFLoader':
|
37 |
loader = PyMuPDFLoader(file_path)
|
38 |
elif loader_type == 'PyPDFLoader':
|
39 |
loader = PyPDFLoader(file_path)
|
40 |
else:
|
41 |
+
logger.error(f"Unknown loader type: {loader_type}")
|
42 |
return
|
43 |
loaded_docs = loader.load()
|
44 |
if loaded_docs:
|
45 |
+
logger.info(f"Successfully loaded {file_path} with {len(loaded_docs)} documents.")
|
46 |
+
logger.info(f"Document content (first 500 chars): {loaded_docs[0].page_content[:500]}")
|
47 |
else:
|
48 |
+
logger.error(f"No documents loaded from {file_path}.")
|
49 |
except Exception as e:
|
50 |
+
logger.error(f"Error loading {file_path} with {loader_type}: {e}")
|
51 |
|
52 |
# 定義載入和處理 PDF 文件的函數
|
53 |
def load_and_process_documents(file_paths, loader_type='PyMuPDFLoader'):
|
54 |
documents = []
|
55 |
+
logger.info("開始載入上傳的 PDF 文件。")
|
56 |
|
57 |
for file_path in file_paths:
|
58 |
+
logger.info(f"載入 PDF 文件: {file_path}")
|
59 |
if not os.path.exists(file_path):
|
60 |
+
logger.error(f"文件不存在: {file_path}")
|
61 |
continue
|
62 |
try:
|
63 |
if loader_type == 'PyMuPDFLoader':
|
|
|
65 |
elif loader_type == 'PyPDFLoader':
|
66 |
loader = PyPDFLoader(file_path)
|
67 |
else:
|
68 |
+
logger.error(f"Unknown loader type: {loader_type}")
|
69 |
continue
|
70 |
loaded_docs = loader.load()
|
71 |
if loaded_docs:
|
72 |
+
logger.info(f"載入 {file_path} 成功,包含 {len(loaded_docs)} 個文檔。")
|
73 |
# 打印第一個文檔的部分內容以確認
|
74 |
+
logger.info(f"第一個文檔內容: {loaded_docs[0].page_content[:500]}")
|
75 |
documents.extend(loaded_docs)
|
76 |
else:
|
77 |
+
logger.error(f"載入 {file_path} 但未找到任何文檔。")
|
78 |
except Exception as e:
|
79 |
+
logger.error(f"載入 {file_path} 時出現錯誤: {e}")
|
80 |
|
81 |
if not documents:
|
82 |
raise ValueError("沒有找到任何 PDF 文件或 PDF 文件無法載入。")
|
83 |
else:
|
84 |
+
logger.info(f"總共載入了 {len(documents)} 個文檔。")
|
85 |
|
86 |
# 分割長文本
|
87 |
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=50)
|
88 |
documents = text_splitter.split_documents(documents)
|
89 |
+
logger.info(f"分割後的文檔數量: {len(documents)}")
|
90 |
|
91 |
if not documents:
|
92 |
raise ValueError("分割後的文檔列表為空。請檢查 PDF 文件內容。")
|
|
|
94 |
# 初始化向量資料庫
|
95 |
try:
|
96 |
embeddings = OpenAIEmbeddings(openai_api_key=api_key) # 直接傳遞 API 密鑰
|
97 |
+
logger.info("初始化 OpenAIEmbeddings 成功。")
|
98 |
except Exception as e:
|
99 |
raise ValueError(f"初始化 OpenAIEmbeddings 時出現錯誤: {e}")
|
100 |
|
|
|
104 |
embedding=embeddings,
|
105 |
persist_directory=VECTORDB_DIR
|
106 |
)
|
107 |
+
logger.info("初始化 Chroma 向量資料庫成功。")
|
108 |
except Exception as e:
|
109 |
raise ValueError(f"初始化 Chroma 向量資料庫時出現錯誤: {e}")
|
110 |
|
|
|
141 |
return chat_history
|
142 |
|
143 |
except Exception as e:
|
144 |
+
logger.error(f"Error in handle_query: {e}")
|
145 |
return chat_history + [("系統", f"出現錯誤: {str(e)}")]
|
146 |
|
147 |
# 定義 Gradio 的處理函數
|
148 |
def process_files(files, state):
|
149 |
+
logger.info("process_files called")
|
150 |
if files:
|
151 |
try:
|
152 |
+
logger.info(f"Received {len(files)} files")
|
153 |
saved_file_paths = []
|
154 |
for file in files:
|
155 |
+
logger.info(f"Processing file: {file.name}")
|
156 |
save_path = os.path.join(VECTORDB_DIR, file.name)
|
157 |
with open(save_path, "wb") as f:
|
158 |
f.write(file.read())
|
159 |
+
# 確認文件是否存在並檢查文件大小
|
160 |
if os.path.exists(save_path):
|
161 |
+
file_size = os.path.getsize(save_path)
|
162 |
+
if file_size > 0:
|
163 |
+
logger.info(f"File successfully saved to: {save_path} (Size: {file_size} bytes)")
|
164 |
+
else:
|
165 |
+
logger.error(f"File saved to {save_path} is empty.")
|
166 |
+
raise ValueError(f"上傳的文件 {file.name} 為空。")
|
167 |
else:
|
168 |
+
logger.error(f"Failed to save file to: {save_path}")
|
169 |
+
raise FileNotFoundError(f"無法保存文件到 {save_path}")
|
170 |
saved_file_paths.append(save_path)
|
171 |
+
# 測試 PDF 加載器,先用 PyMuPDFLoader,再用 PyPDFLoader
|
172 |
+
try:
|
173 |
+
test_pdf_loader(save_path, loader_type='PyMuPDFLoader')
|
174 |
+
except Exception as e:
|
175 |
+
logger.error(f"PyMuPDFLoader failed: {e}")
|
176 |
+
logger.info("Attempting to load with PyPDFLoader...")
|
177 |
+
test_pdf_loader(save_path, loader_type='PyPDFLoader')
|
178 |
# 列出 VECTORDB_DIR 中的所有文件
|
179 |
saved_files = os.listdir(VECTORDB_DIR)
|
180 |
+
logger.info(f"Files in VECTORDB_DIR ({VECTORDB_DIR}): {saved_files}")
|
181 |
vectordb = load_and_process_documents(saved_file_paths, loader_type='PyMuPDFLoader')
|
182 |
state['vectordb'] = vectordb
|
183 |
return "PDF 文件已成功上傳並處理。您現在可以開始提問。", state
|
184 |
except Exception as e:
|
185 |
+
logger.error(f"Error in process_files: {e}")
|
186 |
return f"處理文件時出現錯誤: {e}", state
|
187 |
else:
|
188 |
return "請上傳至少一個 PDF 文件。", state
|