yuze.zyz 1 年之前
父節點
當前提交
da93cfc1d0
共有 4 個文件被更改,包括 19 次插入4 次删除
  1. 1 0
      pdf2zh/doclayout.py
  2. 1 1
      pdf2zh/gui.py
  3. 11 1
      pdf2zh/high_level.py
  4. 6 2
      pdf2zh/translator.py

+ 1 - 0
pdf2zh/doclayout.py

@@ -78,6 +78,7 @@ class OnnxModel(DocLayoutModel):
                 "wybxc/DocLayout-YOLO-DocStructBench-onnx": "AI-ModelScope/DocLayout-YOLO-DocStructBench-onnx"
             }
             from modelscope import snapshot_download
+
             model_dir = snapshot_download(repo_mapping[repo_id])
             pth = os.path.join(model_dir, filename)
         else:

+ 1 - 1
pdf2zh/gui.py

@@ -168,7 +168,7 @@ def translate_file(
     _envs = {}
     for i, env in enumerate(translator.envs.items()):
         _envs[env[0]] = envs[i]
-    
+
     print(f"Files before translation: {os.listdir(output)}")
 
     def progress_bar(t: tqdm.tqdm):

+ 11 - 1
pdf2zh/high_level.py

@@ -92,7 +92,17 @@ def translate_patch(
     rsrcmgr = PDFResourceManager()
     layout = {}
     device = TranslateConverter(
-        rsrcmgr, vfont, vchar, thread, layout, lang_in, lang_out, service, resfont, noto, kwarg.get('envs', {})
+        rsrcmgr,
+        vfont,
+        vchar,
+        thread,
+        layout,
+        lang_in,
+        lang_out,
+        service,
+        resfont,
+        noto,
+        kwarg.get('envs', {}),
     )
 
     assert device is not None

+ 6 - 2
pdf2zh/translator.py

@@ -240,7 +240,9 @@ class AzureOpenAITranslator(BaseTranslator):
         "AZURE_OPENAI_MODEL": "gpt-4o-mini",
     }
 
-    def __init__(self, lang_in, lang_out, model, base_url=None, api_key=None, envs=None):
+    def __init__(
+        self, lang_in, lang_out, model, base_url=None, api_key=None, envs=None
+    ):
         self.set_envs(envs)
         base_url = self.envs["AZURE_OPENAI_BASE_URL"]
         if not model:
@@ -271,7 +273,9 @@ class ModelScopeTranslator(OpenAITranslator):
         "MODELSCOPE_MODEL": "Qwen/Qwen2.5-32B-Instruct",
     }
 
-    def __init__(self, lang_in, lang_out, model, base_url=None, api_key=None, envs=None):
+    def __init__(
+        self, lang_in, lang_out, model, base_url=None, api_key=None, envs=None
+    ):
         self.set_envs(envs)
         base_url = "https://api-inference.modelscope.cn/v1"
         api_key = self.envs["MODELSCOPE_API_KEY"]