Removed redundant and potentially error cause validation for single doc OpenAI embedding

[Hase-U]

The line here should actually be a length comparison with the text as token
https://github.com/hwchase17/langchain/blob/18ec22fe56049aaea446406daab6d66d172dd48f/langchain/embeddings/openai.py#L210

But realistically, there is no need to use a function like len(encode(text)) here, and we can use self._get_len_safe_embeddings by default.
All langchain users will need to install tiktoken, but it's natural to think that using tiktoken is also necessary when using openai's embedding.

So the difference from the current situation is

• tiktoken needs to be installed.
• all docs will be encoded at the client side and sent to openai.

[Hase-U]

Also, this PR means that the use of _get_len_safe_embeddings is completely defaulted.
So please note that this PR over here needs to be merged first. (#3778)

[Hase-U]

#3811

Also, as pointed out here, embeddings/openai.py imports tiktoken in a different way than elsewhere in langchain, so I adjusted it accordingly.
However, since this is a problem of how to write the code, I think that nothing has changed functionally.
the attribute error seems to be caused by other reason

[shawnesquivel]

I reached the same conclusion on my fork. #3811 has more documentation on why this is the right fix.

[shawnesquivel]

Was your OpenAIEmbeddings model parameter OK with the default (text-embedding-ada-002)? Personally, I had to change model to "gpt2" as per #3811.

File "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/tiktoken/registry.py", line 60, in get_encoding
    raise ValueError(f"Unknown encoding {encoding_name}")
ValueError: Unknown encoding text-embedding-ada-002

https://github.com/hwchase17/langchain/blob/master/langchain/embeddings/openai.py#L107
Change:
model: str = "gpt2"

[leo-gan]

@Hase-U Hi , could you, please, resolve the merging issues? After that ping me and I push this PR for the review. Thanks!

[efriis]

Closing because the PR wouldn't line up with the current directory structure of the library (would need to be in /libs/langchain/langchain instead of /langchain). Feel free to reopen against the current head if it's still relevant!