allenai / allenai/ir_datasets

apply SourceDocIter elsewhere

未關閉
#102 0 則留言 0 個 reaction 已指派 0 人 在 GitHub 檢視
chore
主要語言
Python
星號
391
分支
58
PR 合併指標
30 天內沒有已合併 PR

描述

In #101 (C4 + TREC Health Misinformation 2021), I abstracted much of the annoying bits of writing an iterator over document sources into base classes. This should make adding new large datasets considerably easier, with less boilerplate. I should go back and see which prior document collections could be simplified by making use of this.

I believe the datasets that could benefit from this would be:
- [ ] `gov2`
- [ ] `msmarco-passage-v2`
- [ ] `tweets2013-ia`
- [ ] `clueweb09` & `clueweb12`
- [ ] Maybe even the standard docstore implementation?

貢獻指南

這個儲存庫沒有索引到貢獻指南

評估

這個 Issue 還沒有評估資料。

把新 issue 寄到你的電子郵件信箱

精選適合新手參與的 GitHub issue 摘要。