Yuanchun Shen
Yuanchun Shen
There are cases when the default options of transformers tokenizer don't meet our demands, but the tokenizer is wrapped inside Aligner, i.e. we wish the `sentence_to_input` function to also return...
Hi Zhen! I'm trying to reproduce your subgraph retrieval method for other datasets, but I encountered several problems. 1. I am confused about how you got the facts, i.e. `SPO.txt`....
- introduce `gettext` module for localization (literals that are fed to LLMs are now wrapped with gettext `_()`) - add support to simplified Chinese (when env vars `LANGUAGE`, `LC_ALL`, `LC_MESSAGES`,...
### Description Previously, we have implemented our own type checking mechanism due to a lacking of SQL validation phase, which relies on `SqlNode`, and compatibility issues with user-defined types. With...
## Query Information **PPL Command/Query:** ```ppl source=people | eval array = array(1, -2, 3), result = transform(array, x -> x + 2) | fields result ``` **Expected Result:** `[3,0,5]` of...
## Query Information **PPL Command/Query:** ```ppl source=opensearch-sql_test_index_account | eval age_str = CAST(age AS STRING) | bin age_str bins=3 | stats count() by age_str | sort age_str ``` **Expected Result:** |...
### Description Because [nested fields](https://docs.opensearch.org/latest/mappings/supported-field-types/nested/) are indexed as hidden documents, we cannot query them directly. Instead, we have to use the [nested query](https://docs.opensearch.org/latest/query-dsl/joining/nested/) to access them. For example, for the...
**What is the bug?** When performing `bin` command on a timestamp field with `bins` parameter specified, it fails to prepare the plan. **How can one reproduce the bug?** 1. disable...