orbitkit 0.8.60__tar.gz → 0.8.62__tar.gz

{orbitkit-0.8.60/orbitkit.egg-info → orbitkit-0.8.62}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: orbitkit
-Version: 0.8.60
+Version: 0.8.62
 Summary: This project is only for Orbit Tech internal use.
 Home-page: https://github.com/clown-0726/orbitkit
 Author: Lilu Cao

orbitkit-0.8.62/orbitkit/VERSION

@@ -0,0 +1 @@
+0.8.62

{orbitkit-0.8.60 → orbitkit-0.8.62}/orbitkit/audio_transcoder/netmind_extract_v1.py

@@ -104,7 +104,7 @@ def text_processing(netmind_data, lang, translate_model='gpt-4.1-mini'):
 def send_request_to_api(s3_remote_url, **kwargs):
     endpoint = kwargs['endpoint']
     token = kwargs['token']
-    lang = kwargs.get('lang', 'en')
+    lang = kwargs.get('lang', None)
     headers = {
         'Authorization': token,
     }
@@ -112,8 +112,9 @@ def send_request_to_api(s3_remote_url, **kwargs):
     files = {
         'model': (None, 'WhisperX'),
         'url': (None, s3_remote_url),
-        'language': (None, lang),
     }
+    if lang:
+        files['language'] = (None, lang)
 
     response = requests.post(endpoint, headers=headers, files=files)
     response.raise_for_status()
@@ -125,7 +126,7 @@ def send_request_to_api(s3_remote_url, **kwargs):
 def send_request_to_stream(file_steam, **kwargs):
     endpoint = kwargs['endpoint']
     token = kwargs['token']
-    lang = kwargs.get('lang', 'en')
+    lang = kwargs.get('lang', None)
     headers = {
         'Authorization': token,
     }
@@ -133,8 +134,9 @@ def send_request_to_stream(file_steam, **kwargs):
     files = {
         'model': (None, 'WhisperX'),
         'files': (None, file_steam),
-        'language': (None, lang),
     }
+    if lang:
+        files['language'] = (None, lang)
 
     response = requests.post(endpoint, headers=headers, files=files)
     response.raise_for_status()
@@ -143,7 +145,7 @@ def send_request_to_stream(file_steam, **kwargs):
 
 
 def request_wav_from_netmind(s3_client, s3_path=None, file_steam=None, **kwargs):
-    lang = kwargs.get('lang', 'en')
+    lang = kwargs.get('lang', None)
     folder = kwargs.get('folder', '')
     translate_model = kwargs.get('translate_model', 'gpt-4.1-mini')
     if s3_path:
@@ -166,7 +168,12 @@ def request_wav_from_netmind(s3_client, s3_path=None, file_steam=None, **kwargs)
     with open(json_netmind_wav_path, 'w', encoding='utf-8') as json_file:
         json.dump(data, json_file, ensure_ascii=False, indent=4)
 
-    net_process = text_processing(data, lang, translate_model=translate_model)
+    # lang 为 None 时跳过翻译，直接使用原始结果
+    if lang:
+        net_process = text_processing(data, lang, translate_model=translate_model)
+    else:
+        logger.info("lang=None, 跳过翻译，保持原语言")
+        net_process = data
 
     # 翻译接口处理
     json_netmind_lang_wav_path = os.path.join(folder, 'netmind_lang_wav.json')

{orbitkit-0.8.60 → orbitkit-0.8.62}/orbitkit/pdf_extractor/pdf_extractor_minerU_v1.py

@@ -338,12 +338,10 @@ class MinerUExtract:
         return self.default_extractable(pdf_pages, 'discarded_blocks', pages_body, block_seq)
 
     def default_blocks(self, pdf_pages):
-        print(pdf_pages.keys())
-        raise Exception('提取block方法异常')
+        raise Exception(f'提取block方法异常，未知参数: {list(pdf_pages.keys())}')
 
     def default_extractable(self, pdf_pages, block_key, pages_body, block_seq):
         for pages_block in pdf_pages.get(block_key, []):
-            print(pages_block['type'])
             handler_type_func = getattr(self, f'level_two_{pages_block["type"]}', self.level_two_default_blocks)
             block_seq, pages_body = handler_type_func(block_seq, pages_block, pages_body)
         return pages_body
@@ -431,9 +429,10 @@ class MinerUExtract:
         return self.level_two_txt_com(block_seq, pages_block, pages_body, 'sentence')
 
     def level_two_table(self, block_seq, pages_block, pages_body):
-        """处理 table 类型 block，sentence 直接存储 HTML 格式"""
+        """处理 table 类型 block，sentence 直接存储 HTML 格式，脚注等文本单独生成 block"""
         table_blocks = pages_block.get('blocks', [])
         table_html = ""
+        text_contents = []  # 收集 table_footnote 等文本内容
         _image_detail_arr = []
         bbox = pages_block.get('bbox', [0, 0, 0, 0])
 
@@ -444,23 +443,42 @@ class MinerUExtract:
                     span_type = _span.get('type', '')
                     if span_type == 'table' and 'html' in _span:
                         table_html = _span['html']
+                    elif 'content' in _span:  # table_footnote 等文本类型
+                        text_contents.append(_span['content'])
                     if _span.get('image_path'):
                         _image_detail_arr.append({"path": f'images/{_span["image_path"]}', "desc": ""})
 
-        _block_str = table_html
-        if _block_str or _image_detail_arr:
+        # 表格 HTML 单独一个 block
+        if table_html or _image_detail_arr:
             block_seq = block_seq + 1
             text_location = self.get_location(bbox)
             self.blocks.append({
                 "id": id_srv.get_random_short_id(),
                 "page": self.page,
                 "seq_no": block_seq,
-                "sentence": _block_str,
+                "sentence": table_html,
                 "type": 'table',
                 "image_detail": _image_detail_arr,
                 "text_location": text_location
             })
-            pages_body.append(_block_str)
+            pages_body.append(table_html)
+
+        # 脚注等文本单独一个 block
+        if text_contents:
+            block_seq = block_seq + 1
+            text_location = self.get_location(bbox)
+            _text_str = '\n'.join(text_contents)
+            self.blocks.append({
+                "id": id_srv.get_random_short_id(),
+                "page": self.page,
+                "seq_no": block_seq,
+                "sentence": _text_str,
+                "type": 'sentence',
+                "image_detail": [],
+                "text_location": text_location
+            })
+            pages_body.append(_text_str)
+
         return block_seq, pages_body
 
     def level_two_image(self, block_seq, pages_block, pages_body):
@@ -541,8 +559,7 @@ class MinerUExtract:
         return block_seq, pages_body
 
     def level_two_default_blocks(self, block_seq, pages_block, pages_body):
-        print(pages_block.keys())
-        raise Exception('提取block方法异常')
+        raise Exception(f'提取block方法异常，未知类型: {pages_block.get("type")}，参数: {list(pages_block.keys())}')
 
     def get_com_lines(self, pages_lines):
         _block_arr = []

{orbitkit-0.8.60 → orbitkit-0.8.62/orbitkit.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: orbitkit
-Version: 0.8.60
+Version: 0.8.62
 Summary: This project is only for Orbit Tech internal use.
 Home-page: https://github.com/clown-0726/orbitkit
 Author: Lilu Cao

orbitkit-0.8.60/orbitkit/VERSION

@@ -1 +0,0 @@
-0.8.60