WebBaseLoader interprets incorrectly the web_path parameter

[mrtj]

System Info

LangChain version: 0.0.304

Who can help?

@eyurt

Information

• The official example notebooks/scripts
• My own modified scripts

Related Components

• LLMs/Chat Models
• Embedding Models
• Prompts / Prompt Templates / Prompt Selectors
• Output Parsers
• Document Loaders
• Vector Stores / Retrievers
• Memory
• Agents / Agent Executors
• Tools / Toolkits
• Chains
• Callbacks/Tracing
• Async

Reproduction

from langchain.document_loaders.web_base import WebBaseLoader
loader = WebBaseLoader("https://www.google.com")
docs = loader.load()

Result: RequestException is raised with the message "Invalid URL 'h': No scheme supplied. Perhaps you meant https://h?"

Expected behavior

The page contents correctly loads to documents.

[mrtj]

Background

The latest update to WebBaseLoader introduces a regression: if you pass a string to the (now deprecated) web_path parameter, it will be incorrectly interpreted and parsed as a sequence, and as a result when you call the load() method, an invalid uri (just the first letter of the original uri) will be passed to beatifulsoup.

Debugging hints

If you pass a string to web_path, the following condition will be evaluated true and as a result, the single characters of the string will be saved as a list to web_paths:

langchain/libs/langchain/langchain/document_loaders/web_base.py

Lines 79 to 80 in 44489e7

	elif isinstance(web_path, Sequence):
	self.web_paths = list(web_path)

[mrtj]

Duplicate of #11095