Fix: missing parameters in by_plaintext method for PDF naive mode

rag/app/book.py

@@ -113,6 +113,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             lang = lang,
             callback = callback,
             pdf_cls = Pdf,
+            layout_recognizer = layout_recognizer,
             **kwargs
         )
 

rag/app/laws.py

@@ -172,6 +172,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             lang = lang,
             callback = callback,
             pdf_cls = Pdf,
+            layout_recognizer = layout_recognizer,
             **kwargs
         )
 

rag/app/manual.py

@@ -213,6 +213,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             lang = lang,
             callback = callback,
             pdf_cls = Pdf,
+            layout_recognizer = layout_recognizer,
             **kwargs
         )
 

rag/app/naive.py

@@ -111,13 +111,7 @@ def by_tcadp(filename, binary=None, from_page=0, to_page=100000, lang="Chinese",
 
 
 def by_plaintext(filename, binary=None, from_page=0, to_page=100000, callback=None, **kwargs):
-    parser_config = kwargs.get("parser_config", {})
-    layout_recognizer = parser_config.get("layout_recognize", "DeepDOC")
-
-    if isinstance(layout_recognizer, bool):
-        layout_recognizer = "DeepDOC" if layout_recognizer else "Plain Text"
-
-    if layout_recognizer == "Plain Text":
+    if kwargs.get("layout_recognizer", "") == "Plain Text":
         pdf_parser = PlainParser()
     else:
         vision_model = LLMBundle(kwargs["tenant_id"], LLMType.IMAGE2TEXT, llm_name=kwargs.get("layout_recognizer", ""), lang=kwargs.get("lang", "Chinese"))

rag/app/one.py

@@ -99,6 +99,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             lang = lang,
             callback = callback,
             pdf_cls = Pdf,
+            layout_recognizer = layout_recognizer,
             **kwargs
         )
 

rag/app/presentation.py

@@ -142,6 +142,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             lang = lang,
             callback = callback,
             pdf_cls = Pdf,
+            layout_recognizer = layout_recognizer,
             **kwargs
         )