fix code style (#289)

modelscope_agent/tools/__init__.py

@@ -3,10 +3,12 @@
 from .code_interpreter_jupyter import CodeInterpreterJupyter
 from .hf_tool import HFTool
 from .image_chat_tool import ImageChatTool
+from .paraformer_asr_tool import ParaformerAsrTool
 from .phantom_tool import Phantom
 from .pipeline_tool import ModelscopePipelineTool
 from .plugin_tool import LangchainTool
 from .qwen_vl import QWenVL
+from .sambert_tts_tool import SambertTtsTool
 from .style_repaint import StyleRepaint
 from .text_address_tool import TextAddressTool
 from .text_ie_tool import TextInfoExtractTool
@@ -20,8 +22,6 @@
 from .web_browser import WebBrowser
 from .web_search import WebSearch
 from .wordart_tool import WordArtTexture
-from .paraformer_asr_tool import ParaformerAsrTool
-from .sambert_tts_tool import SambertTtsTool
 
 TOOL_INFO_LIST = {
     'text-translation-zh2en': 'TranslationZh2EnTool',

modelscope_agent/tools/paraformer_asr_tool.py

@@ -1,7 +1,7 @@
 import os
 import subprocess
 from http import HTTPStatus
-from typing import List, Any
+from typing import Any, List
 
 from modelscope_agent.tools.tool import Tool, ToolSchema
 from pydantic import ValidationError
@@ -10,9 +10,11 @@
 
 
 def _preprocess(input_file, output_file):
-    ret = subprocess.call(['ffmpeg', '-y', '-i', input_file,
-                           '-f', 's16le', '-acodec', 'pcm_s16le', '-ac', '1', '-ar', '16000',
-                           '-loglevel', 'quiet', output_file])
+    ret = subprocess.call([
+        'ffmpeg', '-y', '-i', input_file, '-f', 's16le', '-acodec',
+        'pcm_s16le', '-ac', '1', '-ar', '16000', '-loglevel', 'quiet',
+        output_file
+    ])
     if ret != 0:
         raise ValueError(f'Failed to preprocess audio file {input_file}')
 
@@ -29,7 +31,8 @@ class ParaformerAsrTool(Tool):
     def __init__(self, cfg={}):
         self.cfg = cfg.get(self.name, {})
 
-        self.api_key = self.cfg.get('dashscope_api_key', os.environ.get('DASHSCOPE_API_KEY'))
+        self.api_key = self.cfg.get('dashscope_api_key',
+                                    os.environ.get('DASHSCOPE_API_KEY'))
         if self.api_key is None:
             raise ValueError('Please set valid DASHSCOPE_API_KEY!')
 
@@ -55,8 +58,12 @@ def __call__(self, *args, **kwargs):
         pcm_file = WORK_DIR + '/' + 'audio.pcm'
         _preprocess(raw_audio_file, pcm_file)
         if not os.path.exists(pcm_file):
-            raise ValueError(f'convert audio to pcm file failed')
-        recognition = Recognition(model='paraformer-realtime-v1', format='pcm', sample_rate=16000, callback=None)
+            raise ValueError(f'convert audio to pcm file {pcm_file} failed')
+        recognition = Recognition(
+            model='paraformer-realtime-v1',
+            format='pcm',
+            sample_rate=16000,
+            callback=None)
         response = recognition.call(pcm_file)
         result = ''
         if response.status_code == HTTPStatus.OK:
@@ -65,5 +72,7 @@ def __call__(self, *args, **kwargs):
                 for sentence in sentences:
                     result += sentence['text']
         else:
-            raise ValueError(f'call paraformer asr failed, request id: {response.get_request_id()}')
+            raise ValueError(
+                f'call paraformer asr failed, request id: {response.get_request_id()}'
+            )
         return {'result': result}

modelscope_agent/tools/phantom_tool.py

@@ -1,19 +1,22 @@
 import os
-import pandas as pd
+import time
+
 import json
+import pandas as pd
 import requests
-from modelscope_agent.tools.localfile2url_utils.localfile2url import get_upload_url
+from modelscope_agent.tools.localfile2url_utils.localfile2url import \
+    get_upload_url
 from modelscope_agent.tools.tool import Tool, ToolSchema
 from pydantic import ValidationError
 from requests.exceptions import RequestException, Timeout
-import time
 
 MAX_RETRY_TIMES = 3
 WORK_DIR = os.getenv('CODE_INTERPRETER_WORK_DIR', '/tmp/ci_workspace')
 
-class Phantom(Tool):# 继承基础类Tool，新建一个继承类
-    description = '追影-放大镜'# 对这个tool的功能描述
-    name = 'phantom_image_enhancement'#tool name
+
+class Phantom(Tool):  # 继承基础类Tool，新建一个继承类
+    description = '追影-放大镜'  # 对这个tool的功能描述
+    name = 'phantom_image_enhancement'  # tool name
     """
     parameters是需要传入api tool的参数，通过api详情获取需要哪些必要入参
     其中每一个参数都是一个字典，包含name，description，required三个字段
@@ -23,20 +26,20 @@ class Phantom(Tool):# 继承基础类Tool，新建一个继承类
         'name': 'input.image_path',
         'description': '输入的待增强图片的本地相对路径',
         'required': True
-    },
-    {
+    }, {
         'name': 'parameters.upscale',
         'description': '选择需要超分的倍率，可选择1、2、3、4',
         'required': False
     }]
 
     def __init__(self, cfg={}):
-        self.cfg = cfg.get(self.name, {})# cfg注册见下一条说明，这里是通过name找到对应的cfg
+        self.cfg = cfg.get(self.name, {})  # cfg注册见下一条说明，这里是通过name找到对应的cfg
         # api url
         self.url = 'https://dashscope.aliyuncs.com/api/v1/services/enhance/image-enhancement/generation'
         # api token，可以选择注册在下面的cfg里，也可以选择将'API_TOKEN'导入环境变量
-        self.token = self.cfg.get('token', os.environ.get('DASHSCOPE_API_KEY', ''))
-        assert self.token != '', 'dashscope api token must be acquired' 
+        self.token = self.cfg.get('token',
+                                  os.environ.get('DASHSCOPE_API_KEY', ''))
+        assert self.token != '', 'dashscope api token must be acquired'
         # 验证，转换参数格式，保持即可
         try:
             all_param = {
@@ -51,12 +54,12 @@ def __init__(self, cfg={}):
         self._str = self.tool_schema.model_dump_json()
         self._function = self.parse_pydantic_model_to_openai_function(
             all_param)
-    
-    # 调用api操作函数，kwargs里是llm根据上面的parameters说明得到的对应参数  
+
+    # 调用api操作函数，kwargs里是llm根据上面的parameters说明得到的对应参数
     def __call__(self, *args, **kwargs):
         # 对入参格式调整和补充，比如解开嵌套的'.'连接的参数，还有导入你默认的一些参数，
         # 比如model，参考下面的_remote_parse_input函数。
-        
+
         remote_parsed_input = json.dumps(
             self._remote_parse_input(*args, **kwargs))
         origin_result = None
@@ -76,14 +79,11 @@ def __call__(self, *args, **kwargs):
             try:
                 # requests请求
                 response = requests.post(
-                    url=self.url,
-                    headers=headers,
-                    data=remote_parsed_input)
+                    url=self.url, headers=headers, data=remote_parsed_input)
 
                 if response.status_code != requests.codes.ok:
                     response.raise_for_status()
-                origin_result = json.loads(
-                    response.content.decode('utf-8'))
+                origin_result = json.loads(response.content.decode('utf-8'))
                 # self._parse_output是基础类Tool对output结果的一个格式调整，你可                  # 以在这里按需调整返回格式
                 self.final_result = self._parse_output(
                     origin_result, remote=True)
@@ -95,11 +95,11 @@ def __call__(self, *args, **kwargs):
                 raise ValueError(
                     f'Remote call failed with error code: {e.response.status_code},\
                     error message: {e.response.content.decode("utf-8")}')
-        
+
         raise ValueError(
             'Remote call max retry times exceeded! Please try to use local call.'
         )
-        
+
     def _remote_parse_input(self, *args, **kwargs):
         restored_dict = {}
         for key, value in kwargs.items():
@@ -114,9 +114,10 @@ def _remote_parse_input(self, *args, **kwargs):
                 # f the key does not contain ".", directly store the key-value pair into restored_dict
                 restored_dict[key] = value
         kwargs = restored_dict
-        
+
         image_path = kwargs['input'].pop('image_path', None)
-        if image_path and image_path.endswith(('.jpeg', '.png', '.jpg', '.bmp')):
+        if image_path and \
+           image_path.endswith(('.jpeg', '.png', '.jpg', '.bmp')):
             # 生成 image_url，然后设置到 kwargs['input'] 中
             # 复用dashscope公共oss
             image_path = f'file://{os.path.join(WORK_DIR, image_path)}'
@@ -127,47 +128,42 @@ def _remote_parse_input(self, *args, **kwargs):
             kwargs['input']['image_url'] = image_url
         else:
             raise ValueError('请先上传一张正确格式的图片')
-        
+
         kwargs['model'] = 'wanx-image-enhancement-v1'
         print('传给tool的参数:', kwargs)
         return kwargs
-    
+
     def get_result(self):
         result_data = json.loads(json.dumps(self.final_result['result']))
         if 'task_id' in result_data['output']:
             task_id = result_data['output']['task_id']
         get_url = f'https://dashscope.aliyuncs.com/api/v1/tasks/{task_id}'
         get_header = {'Authorization': f'Bearer {self.token}'}
-        
+
         origin_result = None
         retry_times = MAX_RETRY_TIMES
         while retry_times:
             retry_times -= 1
             try:
                 response = requests.request(
-                        'GET',
-                        url=get_url,
-                        headers=get_header
-                        )
+                    'GET', url=get_url, headers=get_header)
                 if response.status_code != requests.codes.ok:
                     response.raise_for_status()
-                origin_result = json.loads(
-                    response.content.decode('utf-8'))
+                origin_result = json.loads(response.content.decode('utf-8'))
 
-                get_result = self._parse_output(
-                    origin_result, remote=True)
+                get_result = self._parse_output(origin_result, remote=True)
                 return get_result
             except Timeout:
                 continue
             except RequestException as e:
                 raise ValueError(
                     f'Remote call failed with error code: {e.response.status_code},\
                     error message: {e.response.content.decode("utf-8")}')
-        
+
         raise ValueError(
             'Remote call max retry times exceeded! Please try to use local call.'
         )
-    
+
     def get_phantom_result(self):
         try:
             result = self.get_result()
@@ -186,16 +182,17 @@ def get_phantom_result(self):
                     # output_url = self._parse_output(result['result']['output']['result_url'])
                     output_url = {}
                     output_url['result'] = {}
-                    output_url['result']['url'] = result['result']['output']['result_url']
+                    output_url['result']['url'] = result['result']['output'][
+                        'result_url']
                     # print(output_url)
                     print(output_url)
                     return output_url
-                
+
                 elif task_status in ['FAILED', 'ERROR']:
-                    raise("任务失败")
+                    raise ('任务失败')
 
                 # 继续轮询，等待一段时间后再次调用
                 time.sleep(1)  # 等待 1 秒钟
                 result = self.get_result()
         except Exception as e:
-            print('get request Error:', str(e))
+            print('get request Error:', str(e))

modelscope_agent/tools/sambert_tts_tool.py

@@ -19,7 +19,8 @@ class SambertTtsTool(Tool):
     def __init__(self, cfg={}):
         self.cfg = cfg.get(self.name, {})
 
-        self.api_key = self.cfg.get('dashscope_api_key', os.environ.get('DASHSCOPE_API_KEY'))
+        self.api_key = self.cfg.get('dashscope_api_key',
+                                    os.environ.get('DASHSCOPE_API_KEY'))
         if self.api_key is None:
             raise ValueError('Please set valid DASHSCOPE_API_KEY!')
 
@@ -41,13 +42,16 @@ def __call__(self, *args, **kwargs):
         from dashscope.audio.tts import SpeechSynthesizer
         tts_text = kwargs['text']
         if tts_text is None or len(tts_text) == 0 or tts_text == '':
-            raise ValueError(f'tts input text is valid')
+            raise ValueError('tts input text is valid')
         os.makedirs(WORK_DIR, exist_ok=True)
         wav_file = WORK_DIR + '/sambert_tts_audio.wav'
-        response = SpeechSynthesizer.call(model='sambert-zhijia-v1', format='wav', text=tts_text)
+        response = SpeechSynthesizer.call(
+            model='sambert-zhijia-v1', format='wav', text=tts_text)
         if response.get_audio_data() is not None:
             with open(wav_file, 'wb') as f:
                 f.write(response.get_audio_data())
         else:
-            raise ValueError(f'call sambert tts failed, request id: {response.get_response().request_id}')
+            raise ValueError(
+                f'call sambert tts failed, request id: {response.get_response().request_id}'
+            )
         return {'result': AudioWrapper(wav_file)}

tests/tools/test_dashscope_asr_tts.py

@@ -1,6 +1,5 @@
 from modelscope_agent.agent import AgentExecutor
-from modelscope_agent.tools import ParaformerAsrTool
-from modelscope_agent.tools import SambertTtsTool
+from modelscope_agent.tools import ParaformerAsrTool, SambertTtsTool
 from tests.utils import MockLLM, MockOutParser, MockPromptGenerator
 
 
@@ -21,7 +20,8 @@ def test_sambert_tts():
 def test_paraformer_asr_agent():
     responses = [
         "<|startofthink|>{\"api_name\": \"paraformer_asr_utils\", \"parameters\": "
-        "{\"audio_path\": \"16k-xwlb3_local_user.wav\"}}<|endofthink|>", 'summarize'
+        "{\"audio_path\": \"16k-xwlb3_local_user.wav\"}}<|endofthink|>",
+        'summarize'
     ]
     llm = MockLLM(responses)
 

tests/tools/test_phantom.py

@@ -5,17 +5,23 @@
 
 def test_phantom():
     input = '2_local_user.png'
-    kwargs = {'input.image_path': input, 'parameters.upscale': 2, 'remote': False}
+    kwargs = {
+        'input.image_path': input,
+        'parameters.upscale': 2,
+        'remote': False
+    }
     phantom = Phantom()
     res = phantom(**kwargs)
 
     print(res)
     assert res['result']['url'].startswith('http')
 
+
 def test_phantom_agent():
     responses = [
         "<|startofthink|>{\"api_name\": \"phantom_image_enhancement\", \"parameters\": "
-        "{\"input.image_path\": \"2_local_user.png\"}}<|endofthink|>", 'summarize'
+        "{\"input.image_path\": \"2_local_user.png\"}}<|endofthink|>",
+        'summarize'
     ]
     llm = MockLLM(responses)
 
@@ -34,4 +40,4 @@ def test_phantom_agent():
     res = agent.run('2倍超分')
     print(res)
 
-    assert res[0]['result']['url'].startswith('http')
+    assert res[0]['result']['url'].startswith('http')