GPTSimpleVectorIndex throwing

[epicshardz]

Getting some sort of Authentication error during embedding.
Running in local Jupyter notebook and tried in google colabs.
Testing with only "paul_graham_essay.txt"
Did pip install llama-index and gpt-index, both same result.
To answer the first question yes, my API key is correct and I ran the following where "my actual key" is replaced with key from openai:

import os
os.environ["OPENAI_API_KEY"] ="my actual key"

Line of code that threw error:

index = GPTSimpleVectorIndex(documents, llm_predictor=llm_predictor)

Thanks in advanced!

Error: AuthenticationError                       Traceback (most recent call last)
File [c:\Python310\lib\site-packages\tenacity\__init__.py:409](file:///C:/Python310/lib/site-packages/tenacity/__init__.py:409), in Retrying.__call__(self, fn, *args, **kwargs)
    408 try:
--> 409     result = fn(*args, **kwargs)
    410 except BaseException:  # noqa: B902

File [c:\Python310\lib\site-packages\gpt_index\embeddings\openai.py:147](file:///C:/Python310/lib/site-packages/gpt_index/embeddings/openai.py:147), in get_embeddings(list_of_text, engine)
    145 list_of_text = [text.replace("\n", " ") for text in list_of_text]
--> 147 data = openai.Embedding.create(input=list_of_text, engine=engine).data
    148 data = sorted(data, key=lambda x: x["index"])  # maintain the same order as input.

File [~\AppData\Roaming\Python\Python310\site-packages\openai\api_resources\embedding.py:33](https://file+.vscode-resource.vscode-cdn.net/d%3A/Downloaded/Testing/~/AppData/Roaming/Python/Python310/site-packages/openai/api_resources/embedding.py:33), in Embedding.create(cls, *args, **kwargs)
     32 try:
---> 33     response = super().create(*args, **kwargs)
     35     # If a user specifies base64, we'll just return the encoded string.
     36     # This is only for the default case.

File [~\AppData\Roaming\Python\Python310\site-packages\openai\api_resources\abstract\engine_api_resource.py:149](https://file+.vscode-resource.vscode-cdn.net/d%3A/Downloaded/Testing/~/AppData/Roaming/Python/Python310/site-packages/openai/api_resources/abstract/engine_api_resource.py:149), in EngineAPIResource.create(cls, api_key, api_base, api_type, request_id, api_version, organization, **params)
    127 @classmethod
    128 def create(
    129     cls,
   (...)
    136     **params,
    137 ):
...
--> 363     raise retry_exc from fut.exception()
    365 if self.wait:
    366     sleep = self.wait(retry_state=retry_state)

RetryError: RetryError[]

[jerryjliu]

could you try generating a new openai api key, running the above cell, and seeing if the issue still persists?

[epicshardz]

@jerryjliu That's the first thing that I did.

[jerryjliu]

hmm. could you try setting export OPENAI_API_KEY= in the terminal, and then running those commands in a python interpreter?

[jerryjliu]

@mrsmirf bump on this

[jerryjliu]

@mrsmirf going to close for now, feel free to reopen if you have an update on the issue

[g-benton]

I'm also experience this issue while trying to use a custom LLM (instead of the OpenAI API)

Below is a minimum working example using the Paul Graham essay as the document.

from llama_index import GPTSimpleVectorIndex, SimpleDirectoryReader, PromptHelper, LLMPredictor, ServiceContext
import torch
from langchain.llms.base import LLM
from transformers import pipeline
from typing import Optional, List, Mapping, Any

# define prompt helper
# set maximum input size
max_input_size = 512
# set number of output tokens
num_output = 512
# set maximum chunk overlap
max_chunk_overlap = 20
prompt_helper = PromptHelper(max_input_size, num_output, max_chunk_overlap)


class CustomLLM(LLM):
    model_name = "t5-small"
    pipeline = pipeline("text2text-generation", model=model_name, device="cuda:0", model_kwargs={"torch_dtype":torch.bfloat16})

    def _call(self, prompt: str, stop: Optional[List[str]] = None) -> str:
        response = self.pipeline(prompt)[0]["generated_text"]

        # only return newly generated tokens
        return response

    @property
    def _identifying_params(self) -> Mapping[str, Any]:
        return {"name_of_model": self.model_name}

    @property
    def _llm_type(self) -> str:
        return "custom"

llm_predictor = LLMPredictor(llm=CustomLLM())

service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, prompt_helper=prompt_helper)


documents = SimpleDirectoryReader('data').load_data()
index = GPTSimpleVectorIndex.from_documents(documents, service_context=service_context)

[HoustonMuzamhindo]

Hi @jerryjliu . I have the same error while following your tutorial on Medium: https://app.pinecone.io/organizations/-NRbDE2qTz4DKjcRqFRF/projects/us-west4-gcp:119edac/indexes/pdfextracts

I have installed packages with same version as the Colab file but when running this part, I get endless API error, though on the side the API key does work

index_set = {}
for year in years:
cur_index = GPTSimpleVectorIndex(doc_set[year])
index_set[year] = cur_index
cur_index.save_do_disk(f'index_{year}.json')

[HoustonMuzamhindo]

Update: I created my own person API key not the corporate one and it worked. Could this be something to do with it? I don't know how the API keys work. My personal one looks completely different to the one we use at the corporate.

[vishalp-simplecrm]

How to update simplevector index if some new files are added on google drive folder

import os
import pickle
from langchain import OpenAI
from flask import Flask, render_template, request
from google.auth.transport.requests import Request
from google_auth_oauthlib.flow import InstalledAppFlow
from llama_index import LLMPredictor, GPTSimpleVectorIndex, PromptHelper, ServiceContext, download_loader,MockLLMPredictor,MockEmbedding,Document
from langchain.chat_models import ChatOpenAI

os.environ['OPENAI_API_KEY'] = 'my key'

def authorize_gdocs():
google_oauth2_scopes = [
"https://www.googleapis.com/auth/drive.readonly",
"https://www.googleapis.com/auth/documents.readonly"
]
cred = None
if os.path.exists("token.pickle"):
with open("token.pickle", 'rb') as token:
cred = pickle.load(token)
if not cred or not cred.valid:
if cred and cred.expired and cred.refresh_token:
cred.refresh(Request())
else:
flow = InstalledAppFlow.from_client_secrets_file("client_secrets.json", google_oauth2_scopes)
cred = flow.run_local_server(port=0)
with open("token.pickle", 'wb') as token:
pickle.dump(cred, token)

authorize_gdocs()

GoogleDriveReader = download_loader('GoogleDriveReader')
folder_id = '1h02mlinlWRZdxjf1WotZ3GERNDSoGCmk'
loader = GoogleDriveReader()
documents = loader.load_data(folder_id=folder_id)

Define LLM

llm_predictor = LLMPredictor(llm=ChatOpenAI(temperature=0.7, model_name="gpt-3.5-turbo"))

Define prompt helper

max_input_size = 4096
num_output = 512
max_chunk_overlap = 20
chunk_size_limit = 600
prompt_helper = PromptHelper(max_input_size, num_output,max_chunk_overlap,chunk_size_limit=chunk_size_limit)
service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, prompt_helper=prompt_helper)

Create index from documents only one time then comment it

index = GPTSimpleVectorIndex.from_documents(documents, service_context=service_context)

Save your index to a index.json file

index.save_to_disk('index1.json')

index = GPTSimpleVectorIndex.load_from_disk('index1.json')

[astro313]

Update: I created my own person API key not the corporate one and it worked. Could this be something to do with it? I don't know how the API keys work. My personal one looks completely different to the one we use at the corporate.

Update: I created my own person API key not the corporate one and it worked. Could this be something to do with it? I don't know how the API keys work. My personal one looks completely different to the one we use at the corporate.

Same error with Corporate Key, don't have a personal one.

Error solved by adding: openai.api_key = os.getenv("OPENAI_API_KEY")

[gp-birender]

still getting error with both keys