移除 langchain_community 依赖，内联 CSVLoader/TextLoader，可选化 ebooklib/bs4/mobi/html2text 导入

filetxt/loader.py

@@ -1,16 +1,31 @@
 import os
+import csv
 import codecs
 from datetime import datetime
-from langchain_community.document_loaders.csv_loader import CSVLoader
-from langchain_community.document_loaders.text import TextLoader
 from docx import Document
-from ebooklib import epub
-from bs4 import BeautifulSoup
 from pptx import Presentation
 from openpyxl import load_workbook
-import mobi
 from pypdf import PdfReader
-import html2text
+
+try:
+    from ebooklib import epub
+except ImportError:
+    epub = None
+
+try:
+    from bs4 import BeautifulSoup
+except ImportError:
+    BeautifulSoup = None
+
+try:
+    import mobi
+except ImportError:
+    mobi = None
+
+try:
+    import html2text
+except ImportError:
+    html2text = None
 
 class BaseLoader:
 	def __init__(self, file_path):
@@ -26,11 +41,15 @@ class BaseLoader:
 		raise Exception('Not implement')
 			
 class MyMobiLoader(BaseLoader):
-	def load_pagetext(self):
+    def load_pagetext(self):
-		tempdir, filepath = mobi.extract(self.filepath)
+        if mobi is None:
-		with codecs.open(filepath, "r", "utf-8") as f:
+            raise ImportError("mobi not installed. Run: pip install mobi")
-			content=f.read()
+        if html2text is None:
-			yield (0, html2text.html2text(content))
+            raise ImportError("html2text not installed. Run: pip install html2text")
+        tempdir, filepath = mobi.extract(self.filepath)
+        with codecs.open(filepath, "r", "utf-8") as f:
+            content=f.read()
+            yield (0, html2text.html2text(content))
 		
 class MyPdfLoader(BaseLoader):
 	def load_pagetext(self):
@@ -59,12 +78,11 @@ class MyPptLoader(BaseLoader):
 			yield (i,'\n'.join(txts))
 
 class MyCsvLoader(BaseLoader):
-	def load_pagetext(self):
+    def load_pagetext(self):
-		loader = CSVLoader(self.filepath)
+        with open(self.filepath, 'r', newline='', encoding='utf-8') as f:
-		docs = loader.load()
+            reader = csv.reader(f)
-		for i, d in enumerate(docs):
+            for i, row in enumerate(reader):
-			dat = (i, d.page_content) 
+                yield (i, '	'.join(row))
-			yield dat
 
 class MyExcelLoader(BaseLoader):
 	def load_pagetext(self):
@@ -97,12 +115,10 @@ class MyEpubLoader(BaseLoader):
 
 
 class MyTextLoader(BaseLoader):
-	def load_pagetext(self):
+    def load_pagetext(self):
-		loader = TextLoader(self.filepath)
+        with open(self.filepath, 'r', encoding='utf-8') as f:
-		docs = loader.load()
+            content = f.read()
-		for i, d in enumerate(docs):
+        yield (0, content)
-			dat = (i, d.page_content) 
-			yield dat
 
 class File2Text:
 	all_loaders = {