LAION-AI/Open-Assistant

Handle edge case search queries better

Open

#3,259 建立於 2023年5月29日

在 GitHub 查看
 (4 留言) (0 反應) (1 負責人)Python (37,363 star) (3,263 fork)batch import
backendgood first issue

描述

We should do some preprocessing of search_query here to prevent errors in edge cases, e.g. where the query ends with an operator like & or |.

Words separated by only a space and no operator also have no effect currently, we could modify these such that & is inserted automatically.

List of valid operations I am aware of:

&
|
<->
!

The syntax also allows combinations like &! and groupings via parantheses.

https://github.com/LAION-AI/Open-Assistant/blob/main/backend/oasst_backend/prompt_repository.py#L1023

貢獻者指南

技術棧
python
領域
backend
議題類型
bug
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
blocked
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
Python basicsstring manipulationunderstanding of search operators
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
35
研究方向
Examine the function at backend/oasst backend/prompt repository.py line 1023 to understand current search query handling. Identify the parsing logic for operators (&, |, < >, !) and how the query is processed. Consider edge cases like trailing operators and missing operators between words. The comments may contain additional discussion about desired behavior. Implement preprocessing to sanitize the query and insert '&' where appropriate.