How store the vector (Embedding) from .csv file in Indexes

Search, No Filter
1

I have a .csv file, which consist of 2 columns: Sentences (Thai Sentences) and Embedding (vector with length 1536) as follow:
I want to store the Embedding along with Thai sentences in Indexes. I’ve looked at examples of storing vectors and sentences from this example Hello, Pinecone!.
Then, I applied it to my problem. I got the error as follow:

ApiException: (400)
Reason: Bad Request
HTTP response headers: HTTPHeaderDict({'content-type': 'application/json', 'date': 'Fri, 28 Apr 2023 08:20:46 GMT', 'x-envoy-upstream-service-time': '0', 'content-length': '1251', 'server': 'envoy'})
HTTP response body: {"code":3,"message":"ID must be ASCII, but got อุตุฯ ประกาศฉบับ 5 เตือน 'พายุฤดูร้อน' ถล่ม 27 จังหวัด ","details":[]}

I’m not sure Is the problem caused by my sentence in Thai? If so, how can I fix this problem?

2

Hello!

Here are the general steps to vectorize a string and make it persistent using Pinecone:

  1. Vectorize the strings from the CSV file:
    In this step, you vectorize any string of your choice.

  2. Set the vectorized string in the Index’s metadata:
    Link the vectorized result with the original string and make it persistent in Pinecone. This can be achieved using metadata. For more details, please refer to the following document:
    Metadata Filtering

The sample code you referred to earlier does not include the steps to set the string in the metadata. Therefore, it seems necessary to add this part.

Please review the above information and let me know if you have any questions.