diff --git a/.cursor/rules/weclone-rules.mdc b/.cursor/rules/weclone-rules.mdc
index df98ef9..26b1a8b 100644
--- a/.cursor/rules/weclone-rules.mdc
+++ b/.cursor/rules/weclone-rules.mdc
@@ -15,9 +15,3 @@ alwaysApply: true
 - Unless I ask you to, code comments don't need to be excessive.
 - Prefer using the encapsulated logger `from weclone.utils.log import logger` for printing.
 - When retrieving values from a parameter dictionary read from a configuration file, the `get` method should be preferred whenever possible.
-
-
-
-
-
-
diff --git a/dataset/res_csv/pt/dataset_info.json b/dataset/res_csv/pt/dataset_info.json
index e1ee546..cdd436c 100644
--- a/dataset/res_csv/pt/dataset_info.json
+++ b/dataset/res_csv/pt/dataset_info.json
@@ -3,4 +3,4 @@
     "columns": {
         "prompt": "c"
     }
-}}
\ No newline at end of file
+}}
diff --git a/dataset/res_csv/sft/dataset_info.json b/dataset/res_csv/sft/dataset_info.json
index 3055c42..f1133df 100644
--- a/dataset/res_csv/sft/dataset_info.json
+++ b/dataset/res_csv/sft/dataset_info.json
@@ -31,4 +31,4 @@
             "assistant_tag": "assistant"
         }
     }
-}
\ No newline at end of file
+}
diff --git a/dataset/test_data-privacy.json b/dataset/test_data-privacy.json
index 2ac3119..02461af 100644
--- a/dataset/test_data-privacy.json
+++ b/dataset/test_data-privacy.json
@@ -221,4 +221,4 @@
             "有没有什么特别的收藏？"
         ]
     ]
-}
\ No newline at end of file
+}
diff --git a/dataset/test_data.json b/dataset/test_data.json
index 6f9c386..7358806 100644
--- a/dataset/test_data.json
+++ b/dataset/test_data.json
@@ -154,4 +154,4 @@
             "注意安全啊。"
         ]
     ]
-}
\ No newline at end of file
+}
diff --git a/ds_config.json b/ds_config.json
index 92ba1a4..8a06442 100644
--- a/ds_config.json
+++ b/ds_config.json
@@ -25,4 +25,4 @@
     "train_batch_size": "auto",
     "train_micro_batch_size_per_gpu": "auto",
     "wall_clock_breakdown": false
-}
\ No newline at end of file
+}
diff --git a/examples/mllm.template.jsonc b/examples/mllm.template.jsonc
index 6d1502f..cb5ba94 100644
--- a/examples/mllm.template.jsonc
+++ b/examples/mllm.template.jsonc
@@ -82,4 +82,4 @@
         "max_length": 50,
         "top_p": 0.65
     }
-}
\ No newline at end of file
+}
diff --git a/tests/full_pipe.jsonc b/tests/full_pipe.jsonc
index e0270a4..2f43e62 100644
--- a/tests/full_pipe.jsonc
+++ b/tests/full_pipe.jsonc
@@ -86,4 +86,4 @@
         "max_length": 50,
         "top_p": 0.65
     }
-}
\ No newline at end of file
+}
diff --git a/tests/test_full_pipe.py b/tests/test_full_pipe.py
index ea06739..c6c4f0b 100644
--- a/tests/test_full_pipe.py
+++ b/tests/test_full_pipe.py
@@ -1,12 +1,14 @@
-import pytest
-from unittest import mock
-import sys
+import functools
 import os
 import shutil
-import functools
 import subprocess
+import sys
 import time
-from typing import Union, Optional, cast
+from typing import Optional, Union, cast
+from unittest import mock
+
+import pytest
+
 from weclone.utils.log import logger
 
 sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
diff --git a/weclone-audio/src/Llasa/infer.py b/weclone-audio/src/Llasa/infer.py
index bde7a4a..1053e80 100644
--- a/weclone-audio/src/Llasa/infer.py
+++ b/weclone-audio/src/Llasa/infer.py
@@ -1,12 +1,11 @@
 import os
+
 import soundfile as sf
 from text_to_speech import TextToSpeech
 
-
 sample_audio_text = "对，这就是我万人敬仰的太乙真人，虽然有点婴儿肥，但也掩不住我逼人的帅气。"  # 示例音频文本
 sample_audio_path = os.path.join(os.path.dirname(__file__), "sample.wav")  # 示例音频路径
 tts = TextToSpeech(sample_audio_path, sample_audio_text)
 target_text = "晚上好啊"  # 生成目标文本
 result = tts.infer(target_text)
 sf.write(os.path.join(os.path.dirname(__file__), "output.wav"), result[1], result[0])  # 保存生成音频
-
diff --git a/weclone-audio/src/Llasa/text_to_speech.py b/weclone-audio/src/Llasa/text_to_speech.py
index 2bb468b..9a81d78 100644
--- a/weclone-audio/src/Llasa/text_to_speech.py
+++ b/weclone-audio/src/Llasa/text_to_speech.py
@@ -1,131 +1,125 @@
-import os
-from transformers import AutoTokenizer, AutoModelForCausalLM
-import torch
-import soundfile as sf
-from xcodec2.modeling_xcodec2 import XCodec2Model
-import torchaudio
-
-
-class TextToSpeech:
-    def __init__(self, sample_audio_path, sample_audio_text):
-        self.sample_audio_text = sample_audio_text
-        # 初始化模型
-        llasa_3b = "HKUSTAudio/Llasa-3B"
-        xcodec2 = "HKUSTAudio/xcodec2"
-
-        self.tokenizer = AutoTokenizer.from_pretrained(llasa_3b)
-        self.llasa_3b_model = AutoModelForCausalLM.from_pretrained(
-            llasa_3b,
-            trust_remote_code=True,
-            device_map="auto",
-        )
-        self.llasa_3b_model.eval()
-
-        self.xcodec_model = XCodec2Model.from_pretrained(xcodec2)
-        self.xcodec_model.eval().cuda()
-
-        # 处理音频
-        waveform, sample_rate = torchaudio.load(sample_audio_path)
-        if len(waveform[0]) / sample_rate > 15:
-            print("已将音频裁剪至前15秒。")
-            waveform = waveform[:, : sample_rate * 15]
-
-        # 检查音频是否为立体声
-        if waveform.size(0) > 1:
-            waveform_mono = torch.mean(waveform, dim=0, keepdim=True)
-        else:
-            waveform_mono = waveform
-
-        self.prompt_wav = torchaudio.transforms.Resample(
-            orig_freq=sample_rate, new_freq=16000
-        )(waveform_mono)
-
-        # Encode the prompt wav
-        vq_code_prompt = self.xcodec_model.encode_code(input_waveform=self.prompt_wav)
-        vq_code_prompt = vq_code_prompt[0, 0, :]
-        self.speech_ids_prefix = self.ids_to_speech_tokens(vq_code_prompt)
-        self.speech_end_id = self.tokenizer.convert_tokens_to_ids("<|SPEECH_GENERATION_END|>")
-
-    def ids_to_speech_tokens(self, speech_ids):
-        speech_tokens_str = []
-        for speech_id in speech_ids:
-            speech_tokens_str.append(f"<|s_{speech_id}|>")
-        return speech_tokens_str
-
-    def extract_speech_ids(self, speech_tokens_str):
-        speech_ids = []
-        for token_str in speech_tokens_str:
-            if token_str.startswith("<|s_") and token_str.endswith("|>"):
-                num_str = token_str[4:-2]
-                num = int(num_str)
-                speech_ids.append(num)
-            else:
-                print(f"Unexpected token: {token_str}")
-        return speech_ids
-
-    @torch.inference_mode()
-    def infer(self, target_text):
-        if len(target_text) == 0:
-            return None
-        elif len(target_text) > 300:
-            print("文本过长，请保持在300字符以内。")
-            target_text = target_text[:300]
-
-        input_text = self.sample_audio_text + " " + target_text
-
-        formatted_text = (
-            f"<|TEXT_UNDERSTANDING_START|>{input_text}<|TEXT_UNDERSTANDING_END|>"
-        )
-
-        chat = [
-            {
-                "role": "user",
-                "content": "Convert the text to speech:" + formatted_text,
-            },
-            {
-                "role": "assistant",
-                "content": "<|SPEECH_GENERATION_START|>"
-                + "".join(self.speech_ids_prefix),
-            },
-        ]
-
-        input_ids = self.tokenizer.apply_chat_template(
-            chat, tokenize=True, return_tensors="pt", continue_final_message=True
-        )
-        input_ids = input_ids.to("cuda")
-
-        outputs = self.llasa_3b_model.generate(
-            input_ids,
-            max_length=2048,
-            eos_token_id=self.speech_end_id,
-            do_sample=True,
-            top_p=1,
-            temperature=0.8,
-        )
-        generated_ids = outputs[0][input_ids.shape[1] - len(self.speech_ids_prefix): -1]
-
-        speech_tokens = self.tokenizer.batch_decode(
-            generated_ids, skip_special_tokens=True
-        )
-
-        speech_tokens = self.extract_speech_ids(speech_tokens)
-        speech_tokens = torch.tensor(speech_tokens).cuda().unsqueeze(0).unsqueeze(0)
-
-        gen_wav = self.xcodec_model.decode_code(speech_tokens)
-        gen_wav = gen_wav[:, :, self.prompt_wav.shape[1]:]
-
-        return (16000, gen_wav[0, 0, :].cpu().numpy())
-
-
-if __name__ == "__main__":
-    # 如果遇到问题，请尝试将参考音频转换为WAV或MP3格式，将其裁剪至15秒以内，并缩短提示文本。
-    sample_audio_text = "对，这就是我万人敬仰的太乙真人，虽然有点婴儿肥，但也掩不住我逼人的帅气。"
-    sample_audio_path = os.path.join(os.path.dirname(__file__), "sample.wav")
-
-    tts = TextToSpeech(sample_audio_path, sample_audio_text)
-    target_text = "晚上好啊，吃了吗您"
-    result = tts.infer(target_text)
-    sf.write(os.path.join(os.path.dirname(__file__), "output.wav"), result[1], result[0])
-    target_text = "我是老北京正黄旗！"
-    result = tts.infer(target_text)
-    sf.write(os.path.join(os.path.dirname(__file__), "output1.wav"), result[1], result[0])
+import os
+
+import soundfile as sf
+import torch
+import torchaudio
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from xcodec2.modeling_xcodec2 import XCodec2Model
+
+
+class TextToSpeech:
+    def __init__(self, sample_audio_path, sample_audio_text):
+        self.sample_audio_text = sample_audio_text
+        # 初始化模型
+        llasa_3b = "HKUSTAudio/Llasa-3B"
+        xcodec2 = "HKUSTAudio/xcodec2"
+
+        self.tokenizer = AutoTokenizer.from_pretrained(llasa_3b)
+        self.llasa_3b_model = AutoModelForCausalLM.from_pretrained(
+            llasa_3b,
+            trust_remote_code=True,
+            device_map="auto",
+        )
+        self.llasa_3b_model.eval()
+
+        self.xcodec_model = XCodec2Model.from_pretrained(xcodec2)
+        self.xcodec_model.eval().cuda()
+
+        # 处理音频
+        waveform, sample_rate = torchaudio.load(sample_audio_path)
+        if len(waveform[0]) / sample_rate > 15:
+            print("已将音频裁剪至前15秒。")
+            waveform = waveform[:, : sample_rate * 15]
+
+        # 检查音频是否为立体声
+        if waveform.size(0) > 1:
+            waveform_mono = torch.mean(waveform, dim=0, keepdim=True)
+        else:
+            waveform_mono = waveform
+
+        self.prompt_wav = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)(waveform_mono)
+
+        # Encode the prompt wav
+        vq_code_prompt = self.xcodec_model.encode_code(input_waveform=self.prompt_wav)
+        vq_code_prompt = vq_code_prompt[0, 0, :]
+        self.speech_ids_prefix = self.ids_to_speech_tokens(vq_code_prompt)
+        self.speech_end_id = self.tokenizer.convert_tokens_to_ids("<|SPEECH_GENERATION_END|>")
+
+    def ids_to_speech_tokens(self, speech_ids):
+        speech_tokens_str = []
+        for speech_id in speech_ids:
+            speech_tokens_str.append(f"<|s_{speech_id}|>")
+        return speech_tokens_str
+
+    def extract_speech_ids(self, speech_tokens_str):
+        speech_ids = []
+        for token_str in speech_tokens_str:
+            if token_str.startswith("<|s_") and token_str.endswith("|>"):
+                num_str = token_str[4:-2]
+                num = int(num_str)
+                speech_ids.append(num)
+            else:
+                print(f"Unexpected token: {token_str}")
+        return speech_ids
+
+    @torch.inference_mode()
+    def infer(self, target_text):
+        if len(target_text) == 0:
+            return None
+        elif len(target_text) > 300:
+            print("文本过长，请保持在300字符以内。")
+            target_text = target_text[:300]
+
+        input_text = self.sample_audio_text + " " + target_text
+
+        formatted_text = f"<|TEXT_UNDERSTANDING_START|>{input_text}<|TEXT_UNDERSTANDING_END|>"
+
+        chat = [
+            {
+                "role": "user",
+                "content": "Convert the text to speech:" + formatted_text,
+            },
+            {
+                "role": "assistant",
+                "content": "<|SPEECH_GENERATION_START|>" + "".join(self.speech_ids_prefix),
+            },
+        ]
+
+        input_ids = self.tokenizer.apply_chat_template(
+            chat, tokenize=True, return_tensors="pt", continue_final_message=True
+        )
+        input_ids = input_ids.to("cuda")
+
+        outputs = self.llasa_3b_model.generate(
+            input_ids,
+            max_length=2048,
+            eos_token_id=self.speech_end_id,
+            do_sample=True,
+            top_p=1,
+            temperature=0.8,
+        )
+        generated_ids = outputs[0][input_ids.shape[1] - len(self.speech_ids_prefix) : -1]
+
+        speech_tokens = self.tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+
+        speech_tokens = self.extract_speech_ids(speech_tokens)
+        speech_tokens = torch.tensor(speech_tokens).cuda().unsqueeze(0).unsqueeze(0)
+
+        gen_wav = self.xcodec_model.decode_code(speech_tokens)
+        gen_wav = gen_wav[:, :, self.prompt_wav.shape[1] :]
+
+        return (16000, gen_wav[0, 0, :].cpu().numpy())
+
+
+if __name__ == "__main__":
+    # 如果遇到问题，请尝试将参考音频转换为WAV或MP3格式，将其裁剪至15秒以内，并缩短提示文本。
+    sample_audio_text = "对，这就是我万人敬仰的太乙真人，虽然有点婴儿肥，但也掩不住我逼人的帅气。"
+    sample_audio_path = os.path.join(os.path.dirname(__file__), "sample.wav")
+
+    tts = TextToSpeech(sample_audio_path, sample_audio_text)
+    target_text = "晚上好啊，吃了吗您"
+    result = tts.infer(target_text)
+    sf.write(os.path.join(os.path.dirname(__file__), "output.wav"), result[1], result[0])
+    target_text = "我是老北京正黄旗！"
+    result = tts.infer(target_text)
+    sf.write(os.path.join(os.path.dirname(__file__), "output1.wav"), result[1], result[0])
diff --git a/weclone-audio/src/SparkTTS.py b/weclone-audio/src/SparkTTS.py
index e12bb17..948ae42 100644
--- a/weclone-audio/src/SparkTTS.py
+++ b/weclone-audio/src/SparkTTS.py
@@ -1,14 +1,16 @@
-import re
-import torch
-from typing import Tuple
-from pathlib import Path
-from transformers import AutoTokenizer, AutoModelForCausalLM
 import os
+import re
 import sys
+from pathlib import Path
+from typing import Tuple
+
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
 sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "./Spark-TTS")))
-from sparktts.utils.file import load_config
 from sparktts.models.audio_tokenizer import BiCodecTokenizer
-from sparktts.utils.token_parser import LEVELS_MAP, GENDER_MAP, TASK_TOKEN_MAP
+from sparktts.utils.file import load_config
+from sparktts.utils.token_parser import GENDER_MAP, LEVELS_MAP, TASK_TOKEN_MAP
 
 
 class SparkTTS:
@@ -55,18 +57,12 @@ class SparkTTS:
             Tuple[str, torch.Tensor]: Input prompt; global tokens
         """
 
-        global_token_ids, semantic_token_ids = self.audio_tokenizer.tokenize(
-            prompt_speech_path
-        )
-        global_tokens = "".join(
-            [f"<|bicodec_global_{i}|>" for i in global_token_ids.squeeze()]
-        )
+        global_token_ids, semantic_token_ids = self.audio_tokenizer.tokenize(prompt_speech_path)
+        global_tokens = "".join([f"<|bicodec_global_{i}|>" for i in global_token_ids.squeeze()])
 
         # Prepare the input tokens for the model
         if prompt_text is not None:
-            semantic_tokens = "".join(
-                [f"<|bicodec_semantic_{i}|>" for i in semantic_token_ids.squeeze()]
-            )
+            semantic_tokens = "".join([f"<|bicodec_semantic_{i}|>" for i in semantic_token_ids.squeeze()])
             inputs = [
                 TASK_TOKEN_MAP["tts"],
                 "<|start_content|>",
@@ -125,9 +121,7 @@ class SparkTTS:
         speed_label_tokens = f"<|speed_label_{speed_level_id}|>"
         gender_tokens = f"<|gender_{gender_id}|>"
 
-        attribte_tokens = "".join(
-            [gender_tokens, pitch_label_tokens, speed_label_tokens]
-        )
+        attribte_tokens = "".join([gender_tokens, pitch_label_tokens, speed_label_tokens])
 
         control_tts_inputs = [
             TASK_TOKEN_MAP["controllable_tts"],
@@ -175,9 +169,7 @@ class SparkTTS:
             prompt = self.process_prompt_control(gender, pitch, speed, text)
 
         else:
-            prompt, global_token_ids = self.process_prompt(
-                text, prompt_speech_path, prompt_text
-            )
+            prompt, global_token_ids = self.process_prompt(text, prompt_speech_path, prompt_text)
         model_inputs = self.tokenizer([prompt], return_tensors="pt").to(self.device)
 
         # Generate speech using the model
@@ -192,8 +184,7 @@ class SparkTTS:
 
         # Trim the output tokens to remove the input tokens
         generated_ids = [
-            output_ids[len(input_ids):]
-            for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+            output_ids[len(input_ids) :] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
         ]
 
         # Decode the generated tokens into text
@@ -201,9 +192,7 @@ class SparkTTS:
 
         # Extract semantic token IDs from the generated text
         pred_semantic_ids = (
-            torch.tensor([int(token) for token in re.findall(r"bicodec_semantic_(\d+)", predicts)])
-            .long()
-            .unsqueeze(0)
+            torch.tensor([int(token) for token in re.findall(r"bicodec_semantic_(\d+)", predicts)]).long().unsqueeze(0)
         )
 
         if gender is not None:
diff --git a/weclone-audio/src/get_sample_audio.py b/weclone-audio/src/get_sample_audio.py
index 8e53541..4d2d89d 100644
--- a/weclone-audio/src/get_sample_audio.py
+++ b/weclone-audio/src/get_sample_audio.py
@@ -1,18 +1,20 @@
-import os
 import argparse
+import os
+
 from pywxdump.db import MediaHandler
 
+
 def main():
     parser = argparse.ArgumentParser(description="Extract audio from WeChat database")
-    parser.add_argument("--db-path", type=str, required=True,
-                        help="Path to WeChat database file")
-    parser.add_argument("--MsgSvrID", type=str, required=True,
-                        help="Message server ID of the audio")
-    parser.add_argument("--save-path", type=str,
-                        default=os.path.join(os.path.dirname(__file__), "sample.wav"),
-                        help="Path to save the audio file (default: sample.wav in script directory)")
-    parser.add_argument("--rate", type=int, default=24000,
-                        help="Sample rate for audio conversion (default: 24000)")
+    parser.add_argument("--db-path", type=str, required=True, help="Path to WeChat database file")
+    parser.add_argument("--MsgSvrID", type=str, required=True, help="Message server ID of the audio")
+    parser.add_argument(
+        "--save-path",
+        type=str,
+        default=os.path.join(os.path.dirname(__file__), "sample.wav"),
+        help="Path to save the audio file (default: sample.wav in script directory)",
+    )
+    parser.add_argument("--rate", type=int, default=24000, help="Sample rate for audio conversion (default: 24000)")
 
     args = parser.parse_args()
 
@@ -31,5 +33,6 @@ def main():
         rate=args.rate,
     )
 
+
 if __name__ == "__main__":
     main()
diff --git a/weclone-audio/src/infer.py b/weclone-audio/src/infer.py
index 55094d5..10e3f9d 100644
--- a/weclone-audio/src/infer.py
+++ b/weclone-audio/src/infer.py
@@ -1,7 +1,7 @@
 import os
+
 import soundfile as sf
 import torch
-
 from SparkTTS import SparkTTS
 
 model = SparkTTS("weclone-audio/pretrained_models/Spark-TTS-0.5B", "cuda")
diff --git a/weclone-audio/src/server未完工/.env.example b/weclone-audio/src/server未完工/.env.example
index 627d413..71ea56e 100644
--- a/weclone-audio/src/server未完工/.env.example
+++ b/weclone-audio/src/server未完工/.env.example
@@ -11,4 +11,4 @@ REQUIRE_API_KEY=True
 
 REMOVE_FILTER=False
 
-EXPAND_API=True
\ No newline at end of file
+EXPAND_API=True
diff --git a/weclone-audio/src/server未完工/handle_text.py b/weclone-audio/src/server未完工/handle_text.py
index c4fba21..167b3bf 100644
--- a/weclone-audio/src/server未完工/handle_text.py
+++ b/weclone-audio/src/server未完工/handle_text.py
@@ -1,6 +1,8 @@
 import re
+
 import emoji
 
+
 def prepare_tts_input_with_context(text: str) -> str:
     """
     Prepares text for a TTS API by cleaning Markdown and adding minimal contextual hints
diff --git a/weclone-audio/src/server未完工/requirements.txt b/weclone-audio/src/server未完工/requirements.txt
index 4019693..f08da66 100644
--- a/weclone-audio/src/server未完工/requirements.txt
+++ b/weclone-audio/src/server未完工/requirements.txt
@@ -2,4 +2,4 @@ flask
 gevent
 python-dotenv
 edge-tts
-emoji
\ No newline at end of file
+emoji
diff --git a/weclone-audio/src/server未完工/server.py b/weclone-audio/src/server未完工/server.py
index 486ffd3..b11533e 100644
--- a/weclone-audio/src/server未完工/server.py
+++ b/weclone-audio/src/server未完工/server.py
@@ -1,13 +1,13 @@
 # server.py
 
-from flask import Flask, request, send_file, jsonify
-from gevent.pywsgi import WSGIServer
-from dotenv import load_dotenv
 import os
 
+from dotenv import load_dotenv
+from flask import Flask, jsonify, request, send_file
+from gevent.pywsgi import WSGIServer
 from handle_text import prepare_tts_input_with_context
 from tts_handler import generate_speech, get_models, get_voices
-from utils import getenv_bool, require_api_key, AUDIO_FORMAT_MIME_TYPES
+from utils import AUDIO_FORMAT_MIME_TYPES, getenv_bool, require_api_key
 
 app = Flask(__name__)
 load_dotenv()
diff --git a/weclone-audio/src/server未完工/tts_handler.py b/weclone-audio/src/server未完工/tts_handler.py
index f243acc..39dafef 100644
--- a/weclone-audio/src/server未完工/tts_handler.py
+++ b/weclone-audio/src/server未完工/tts_handler.py
@@ -1,10 +1,11 @@
-import edge_tts
 import asyncio
-import tempfile
-import subprocess
 import os
+import subprocess
+import tempfile
 from pathlib import Path
 
+import edge_tts
+
 # Language default (environment variable)
 DEFAULT_LANGUAGE = os.getenv('DEFAULT_LANGUAGE', 'en-US')
 
diff --git a/weclone-audio/src/server未完工/utils.py b/weclone-audio/src/server未完工/utils.py
index 7fe9845..c0f3ca3 100644
--- a/weclone-audio/src/server未完工/utils.py
+++ b/weclone-audio/src/server未完工/utils.py
@@ -1,9 +1,10 @@
 # utils.py
 
-from flask import request, jsonify
-from functools import wraps
 import os
+from functools import wraps
+
 from dotenv import load_dotenv
+from flask import jsonify, request
 
 load_dotenv()
 
diff --git a/weclone/cli.py b/weclone/cli.py
index 78e97c0..59d6550 100644
--- a/weclone/cli.py
+++ b/weclone/cli.py
@@ -1,12 +1,13 @@
-import click
-import commentjson
-from pathlib import Path
+import functools
 import os
 import sys
-import functools
+from pathlib import Path
+
+import click
+import commentjson
 
-from weclone.utils.log import logger, capture_output
 from weclone.utils.config import load_config
+from weclone.utils.log import capture_output, logger
 
 cli_config: dict | None = None
 
diff --git a/weclone/core/inference/offline_infer.py b/weclone/core/inference/offline_infer.py
index be7affa..e417834 100644
--- a/weclone/core/inference/offline_infer.py
+++ b/weclone/core/inference/offline_infer.py
@@ -1,20 +1,13 @@
-import json
 from typing import List, Optional, Union
 
-
-from llamafactory.data import get_dataset, get_template_and_fix_tokenizer
-from llamafactory.extras.constants import IGNORE_INDEX
+from llamafactory.data import get_template_and_fix_tokenizer
 from llamafactory.extras.misc import get_device_count
-from llamafactory.extras.packages import is_vllm_available
 from llamafactory.hparams import get_infer_args
 from llamafactory.model import load_tokenizer
 from pydantic import BaseModel
-from vllm.sampling_params import GuidedDecodingParams
-
-
 from vllm import LLM, SamplingParams
 from vllm.lora.request import LoRARequest
-
+from vllm.sampling_params import GuidedDecodingParams
 
 # 这里不需要写太好，transforms库后续更新自带vllm
 
diff --git a/weclone/core/inference/online_infer.py b/weclone/core/inference/online_infer.py
index 518cf71..8868d5e 100644
--- a/weclone/core/inference/online_infer.py
+++ b/weclone/core/inference/online_infer.py
@@ -1,26 +1,23 @@
-import json
-import time
-import requests
 from openai import OpenAI
 
+
 class OnlineLLM:
-    def __init__(self, api_key: str, base_url: str,model_name: str,default_system: str):
+    def __init__(self, api_key: str, base_url: str, model_name: str, default_system: str):
         self.api_key = api_key
         self.base_url = base_url
         self.model_name = model_name
         self.default_system = default_system
-        self.client = OpenAI(
-            api_key=self.api_key,
-            base_url=self.base_url
-        )
+        self.client = OpenAI(api_key=self.api_key, base_url=self.base_url)
 
-
-    def chat(self,prompt_text, 
-            temperature: float = 0.7, 
-            max_tokens: int = 1024, 
-            top_p: float = 0.95, 
-            stream: bool = False,
-            enable_thinking: bool = False):
+    def chat(
+        self,
+        prompt_text,
+        temperature: float = 0.7,
+        max_tokens: int = 1024,
+        top_p: float = 0.95,
+        stream: bool = False,
+        enable_thinking: bool = False,
+    ):
         messages = [
             {"role": "system", "content": self.default_system},
             {"role": "user", "content": prompt_text},
@@ -29,12 +26,10 @@ class OnlineLLM:
             model=self.model_name,
             messages=messages,
             stream=stream,
-            temperature = temperature,
+            temperature=temperature,
             max_tokens=max_tokens,
             top_p=top_p,
-            # enable_thinking=enable_thinking   适配Qwen3动态开启推理  
-
+            # enable_thinking=enable_thinking   适配Qwen3动态开启推理
         )
 
         return response
-
diff --git a/weclone/data/chat_parsers/wechat_parser.py b/weclone/data/chat_parsers/wechat_parser.py
index 5bdb857..8fc4522 100644
--- a/weclone/data/chat_parsers/wechat_parser.py
+++ b/weclone/data/chat_parsers/wechat_parser.py
@@ -1,7 +1,8 @@
-import os
 import argparse
-from pathlib import Path
+import os
 import shutil
+from pathlib import Path
+
 import pandas as pd
 from tqdm import tqdm
 
diff --git a/weclone/data/clean/strategies.py b/weclone/data/clean/strategies.py
index 81c37df..d540979 100644
--- a/weclone/data/clean/strategies.py
+++ b/weclone/data/clean/strategies.py
@@ -1,12 +1,14 @@
 import json
-import pandas as pd
+import os
 from abc import ABC, abstractmethod
 from dataclasses import dataclass
-from typing import Any, Dict, List, Union
+from typing import Any, Dict, List
+
+import pandas as pd
 from langchain_core.prompts import PromptTemplate
-from weclone.data.models import QaPair, CutMessage, QaPairScore
+
+from weclone.data.models import QaPair, QaPairScore
 from weclone.prompts.clean_data import CLEAN_PROMPT
-import os
 from weclone.utils.log import logger
 
 
diff --git a/weclone/data/models.py b/weclone/data/models.py
index 20c9d63..f19370d 100644
--- a/weclone/data/models.py
+++ b/weclone/data/models.py
@@ -1,8 +1,9 @@
 from dataclasses import dataclass
 from enum import Enum
-from typing import Union, Optional
+
 from pandas import Timestamp
 from pydantic import BaseModel
+
 from weclone.utils.i18n import MultiLangList
 
 
diff --git a/weclone/data/qa_generator.py b/weclone/data/qa_generator.py
index 5696299..1caf5db 100644
--- a/weclone/data/qa_generator.py
+++ b/weclone/data/qa_generator.py
@@ -1,20 +1,20 @@
+import json
 import os
-import sys
-import subprocess
-from typing import Dict, List, Union
 import re
+import subprocess
+import sys
+from typing import List, Union
 
 import pandas as pd
-import json
-from pandas import Timestamp
 from llamafactory.extras.packages import is_vllm_available
+from pandas import Timestamp
 
 from weclone.data.clean.strategies import LLMCleaningStrategy
 from weclone.data.clean.strategies_online import OlineLLMCleaningStrategy
+from weclone.data.models import ChatMessage, CutMessage, QaPair, skip_type_list
+from weclone.data.strategies import LLMStrategy, TimeWindowStrategy
 from weclone.utils.config import load_config
 from weclone.utils.log import logger
-from weclone.data.models import ChatMessage, CutMessage, skip_type_list, QaPair
-from weclone.data.strategies import TimeWindowStrategy, LLMStrategy
 
 
 class DataProcessor:
diff --git a/weclone/data/qa_generatorV2.py b/weclone/data/qa_generatorV2.py
index 59cc74a..f827e67 100644
--- a/weclone/data/qa_generatorV2.py
+++ b/weclone/data/qa_generatorV2.py
@@ -1,27 +1,27 @@
+import json
 import os
-import sys
-import subprocess
-from typing import Dict, List, Union
 import re
+import subprocess
+import sys
+from typing import List, Union
 
 import pandas as pd
-import json
 from pandas import Timestamp
 
 from weclone.data.clean.strategies import LLMCleaningStrategy
 from weclone.data.clean.strategies_online import OlineLLMCleaningStrategy
-from weclone.utils.config import load_config
-from weclone.utils.log import logger
 from weclone.data.models import (
     ChatMessage,
     CutMessage,
-    skip_type_list,
-    cut_type_list,
-    QaPairV2,
     Message,
+    QaPairV2,
+    cut_type_list,
+    skip_type_list,
 )
-from weclone.data.strategies import TimeWindowStrategy, LLMStrategy
+from weclone.data.strategies import LLMStrategy, TimeWindowStrategy
 from weclone.data.utils import check_image_file_exists
+from weclone.utils.config import load_config
+from weclone.utils.log import logger
 
 
 class DataProcessor:
diff --git a/weclone/data/strategies.py b/weclone/data/strategies.py
index 4a457f1..2476f0a 100644
--- a/weclone/data/strategies.py
+++ b/weclone/data/strategies.py
@@ -1,7 +1,8 @@
+from abc import ABC, abstractmethod
 from dataclasses import dataclass
 from typing import List
+
 from .models import ChatMessage
-from abc import ABC, abstractmethod
 
 
 @dataclass
@@ -11,9 +12,7 @@ class ConversationStrategy(ABC):
     is_single_chat: bool
 
     @abstractmethod
-    def is_same_conversation(
-        self, history_msg: List[ChatMessage], current_msg: ChatMessage
-    ) -> bool:
+    def is_same_conversation(self, history_msg: List[ChatMessage], current_msg: ChatMessage) -> bool:
         """判断两条消息是否属于同一个对话"""
         pass
 
@@ -24,12 +23,8 @@ class TimeWindowStrategy(ConversationStrategy):
 
     time_window: int  # 时间窗口（分钟）
 
-    def is_same_conversation(
-        self, history_msg: List[ChatMessage], current_msg: ChatMessage
-    ) -> bool:
-        time_diff = abs(
-            (current_msg.CreateTime - history_msg[-1].CreateTime)
-        ).total_seconds()
+    def is_same_conversation(self, history_msg: List[ChatMessage], current_msg: ChatMessage) -> bool:
+        time_diff = abs((current_msg.CreateTime - history_msg[-1].CreateTime)).total_seconds()
         return time_diff <= self.time_window
 
 
@@ -37,9 +32,7 @@ class TimeWindowStrategy(ConversationStrategy):
 class LLMStrategy(ConversationStrategy):
     """基于大模型判断策略"""
 
-    def is_same_conversation(
-        self, history_msg: List[ChatMessage], current_msg: ChatMessage
-    ) -> bool:
+    def is_same_conversation(self, history_msg: List[ChatMessage], current_msg: ChatMessage) -> bool:
         # 修复user_id错误，使用talker字段代替user_id
         return current_msg.talker == history_msg[-1].talker if history_msg else False
 
@@ -51,10 +44,6 @@ class CompositeStrategy(ConversationStrategy):
     strategies: List[ConversationStrategy]
     require_all: bool = True  # True表示所有策略都满足，False表示任一策略满足即可
 
-    def is_same_conversation(
-        self, history_msg: List[ChatMessage], current_msg: ChatMessage
-    ) -> bool:
-        results = [
-            s.is_same_conversation(history_msg, current_msg) for s in self.strategies
-        ]
+    def is_same_conversation(self, history_msg: List[ChatMessage], current_msg: ChatMessage) -> bool:
+        results = [s.is_same_conversation(history_msg, current_msg) for s in self.strategies]
         return all(results) if self.require_all else any(results)
diff --git a/weclone/data/utils.py b/weclone/data/utils.py
index 46f91df..085312f 100644
--- a/weclone/data/utils.py
+++ b/weclone/data/utils.py
@@ -1,5 +1,6 @@
 import os
 from pathlib import Path
+
 from weclone.utils.log import logger
 
 
@@ -34,5 +35,5 @@ def check_image_file_exists(file_path: str) -> str | bool:
 
 
 if __name__ == "__main__":
-    path = "Storage\Image\2021-08\6ce3f785b4230246639c3dd0d4a8848c.dat"
+    path = "Storage\\Image\2021-08\6ce3f785b4230246639c3dd0d4a8848c.dat"
     print(check_image_file_exists(path))
diff --git a/weclone/eval/test_model.py b/weclone/eval/test_model.py
index 59eeb2f..9dc35e6 100644
--- a/weclone/eval/test_model.py
+++ b/weclone/eval/test_model.py
@@ -1,10 +1,10 @@
 import json
+from typing import List, cast  # 导入 cast
+
 import openai
 from openai import OpenAI  # 导入 OpenAI 类
-
+from openai.types.chat import ChatCompletionMessageParam  # 导入消息参数类型
 from tqdm import tqdm
-from typing import List, Dict, cast # 导入 cast
-from openai.types.chat import ChatCompletionMessageParam # 导入消息参数类型
 
 from weclone.utils.config import load_config
 
@@ -19,10 +19,7 @@ config = {
 config = type("Config", (object,), config)()
 
 # 初始化 OpenAI 客户端
-client = OpenAI(
-    api_key="""sk-test""",
-    base_url="http://127.0.0.1:8005/v1"
-)
+client = OpenAI(api_key="""sk-test""", base_url="http://127.0.0.1:8005/v1")
 
 
 def handler_text(content: str, history: list, config):
@@ -37,14 +34,14 @@ def handler_text(content: str, history: list, config):
         typed_messages = cast(List[ChatCompletionMessageParam], messages)
         response = client.chat.completions.create(
             model=config.model,
-            messages=typed_messages, # 传递转换后的列表
-            max_tokens=50
+            messages=typed_messages,  # 传递转换后的列表
+            max_tokens=50,
         )
     except openai.APIError as e:
         history.pop()
         return "AI接口出错,请重试\n" + str(e)
 
-    resp = str(response.choices[0].message.content) # type: ignore
+    resp = str(response.choices[0].message.content)  # type: ignore
     resp = resp.replace("\n ", "")
     history.append({"role": "assistant", "content": resp})
     return resp
diff --git a/weclone/eval/web_demo.py b/weclone/eval/web_demo.py
index 98ec03e..63af1f0 100644
--- a/weclone/eval/web_demo.py
+++ b/weclone/eval/web_demo.py
@@ -1,4 +1,5 @@
 from llamafactory.webui.interface import create_web_demo
+
 from weclone.utils.config import load_config
 
 
diff --git a/weclone/server/api_service.py b/weclone/server/api_service.py
index 29d8efa..b07d027 100644
--- a/weclone/server/api_service.py
+++ b/weclone/server/api_service.py
@@ -1,9 +1,10 @@
 import os
-import uvicorn
-from llamafactory.chat import ChatModel
-from llamafactory.api.app import create_app
-from weclone.utils.config import load_config
 
+import uvicorn
+from llamafactory.api.app import create_app
+from llamafactory.chat import ChatModel
+
+from weclone.utils.config import load_config
 
 
 def main():
diff --git a/weclone/train/train_pt.py b/weclone/train/train_pt.py
index b7fd547..d4ab046 100644
--- a/weclone/train/train_pt.py
+++ b/weclone/train/train_pt.py
@@ -1,4 +1,5 @@
 from llamafactory.train.tuner import run_exp
+
 from weclone.utils.config import load_config
 
 config = load_config("train_pt")
diff --git a/weclone/train/train_sft.py b/weclone/train/train_sft.py
index 78483d7..c7c5826 100644
--- a/weclone/train/train_sft.py
+++ b/weclone/train/train_sft.py
@@ -1,11 +1,14 @@
+import json
 import os
 import sys
-import json
-from llamafactory.train.tuner import run_exp
+
 from llamafactory.extras.misc import get_current_device
+from llamafactory.train.tuner import run_exp
+
+from weclone.data.clean.strategies import LLMCleaningStrategy
 from weclone.utils.config import load_config
 from weclone.utils.log import logger
-from weclone.data.clean.strategies import LLMCleaningStrategy
+
 
 def main():
     train_config = load_config(arg_type="train_sft")
diff --git a/weclone/utils/config.py b/weclone/utils/config.py
index 06ef6af..2582f1a 100644
--- a/weclone/utils/config.py
+++ b/weclone/utils/config.py
@@ -1,7 +1,8 @@
 import os
-import commentjson
 import sys
 
+import commentjson
+
 from .log import logger
 from .tools import dict_to_argv
 
diff --git a/weclone/utils/i18n.py b/weclone/utils/i18n.py
index 9efee32..78502e4 100644
--- a/weclone/utils/i18n.py
+++ b/weclone/utils/i18n.py
@@ -1,4 +1,4 @@
-from typing import List, Dict, Optional, Tuple
+from typing import Dict, List, Optional
 
 
 class MultiLangList:
diff --git a/weclone/utils/length_cdf.py b/weclone/utils/length_cdf.py
index f4f8fd5..7d2e17b 100644
--- a/weclone/utils/length_cdf.py
+++ b/weclone/utils/length_cdf.py
@@ -15,12 +15,12 @@
 from collections import defaultdict
 
 import fire
-from tqdm import tqdm
-from weclone.utils.log import logger
-
 from llamafactory.data import get_dataset, get_template_and_fix_tokenizer
 from llamafactory.hparams import get_train_args
 from llamafactory.model import load_tokenizer
+from tqdm import tqdm
+
+from weclone.utils.log import logger
 
 
 def length_cdf(
diff --git a/weclone/utils/log.py b/weclone/utils/log.py
index ad164a8..1ebc71d 100644
--- a/weclone/utils/log.py
+++ b/weclone/utils/log.py
@@ -1,7 +1,8 @@
-from loguru import logger
 import sys
 from functools import wraps
 
+from loguru import logger
+
 logger.remove()
 
 logger.add(
diff --git a/weclone/utils/tools.py b/weclone/utils/tools.py
index 28bc29d..3c2fb11 100644
--- a/weclone/utils/tools.py
+++ b/weclone/utils/tools.py
@@ -5,5 +5,3 @@ def dict_to_argv(d):
         if v is not None:
             argv.append(str(v))
     return argv
-
-