mrshu commited on
Commit
3adce9d
·
verified ·
1 Parent(s): df4b3e7

Card and licence: takedown policy, provenance-mapping note

Browse files
Files changed (2) hide show
  1. LICENSE-DATA.md +3 -2
  2. README.md +5 -3
LICENSE-DATA.md CHANGED
@@ -36,5 +36,6 @@ the Parliament. They are kept as metadata and removed from the canonical retriev
36
  headers and co-signatory lines; residual mentions remain (e.g. 12 train queries with a "Question supported by …"
37
  list, and MEPs or other public figures named in question/answer content). `subject`, `header` and `raw_text` are
38
  unmasked. See docs/datasheet.md "Known issues in v1.0". Removal requests concerning personal data can be filed on
39
- the code repository's issue tracker (https://github.com/NaiveNeuron/ParlIR-ecir-2027/issues). They take effect in
40
- the next release.
 
 
36
  headers and co-signatory lines; residual mentions remain (e.g. 12 train queries with a "Question supported by …"
37
  list, and MEPs or other public figures named in question/answer content). `subject`, `header` and `raw_text` are
38
  unmasked. See docs/datasheet.md "Known issues in v1.0". Removal requests concerning personal data can be filed on
39
+ the code repository's issue tracker (https://github.com/NaiveNeuron/ParlIR-ecir-2027/issues). The maintainers review
40
+ each request. Accepted removals apply to the main branch and the next release, while published, pinned revisions
41
+ stay unchanged so that reported results remain reproducible, unless the law requires otherwise.
README.md CHANGED
@@ -122,8 +122,9 @@ only by the evaluator's as-of filter for the citation tasks. For those tasks, sy
122
  earlier eligible answers to the questions they replied to, as the question-matching baselines do.
123
 
124
  The text-only export (`v1.0-textonly`) is the safest choice for evaluation. Corpus and queries carry only a random
125
- id and the masked text, the dates sit in evaluator-only files, and the id mapping is a separate download. All
126
- stored main-experiment runs score the same on it. It omits the links that question matching needs.
 
127
 
128
  ## Baselines (test split)
129
 
@@ -161,7 +162,8 @@ Source texts: European Parliament Open Data Portal, © European Union, CC BY 4.0
161
  under Decision 2011/833/EU (CC BY 4.0 via C(2019) 1655), Council answers under Decision (EU) 2017/1842. The texts
162
  were modified (headers removed, identifiers masked). ParlIR annotations: CC BY 4.0. See `LICENSE-DATA.md`.
163
  Questions are signed by MEPs acting as public office-holders. Removal requests can be filed on the code
164
- repository's issue tracker.
 
165
 
166
  ## Citation
167
 
 
122
  earlier eligible answers to the questions they replied to, as the question-matching baselines do.
123
 
124
  The text-only export (`v1.0-textonly`) is the safest choice for evaluation. Corpus and queries carry only a random
125
+ id and the masked text, the dates sit in evaluator-only files, and the id mapping is a separate download
126
+ (`v1.0/textonly-provenance/`). That mapping is public, so do not use it as a retrieval feature. All stored
127
+ main-experiment runs score the same on the export. It omits the links that question matching needs.
128
 
129
  ## Baselines (test split)
130
 
 
162
  under Decision 2011/833/EU (CC BY 4.0 via C(2019) 1655), Council answers under Decision (EU) 2017/1842. The texts
163
  were modified (headers removed, identifiers masked). ParlIR annotations: CC BY 4.0. See `LICENSE-DATA.md`.
164
  Questions are signed by MEPs acting as public office-holders. Removal requests can be filed on the code
165
+ repository's issue tracker. Accepted removals apply to `main` and the next release, and published, pinned revisions
166
+ stay unchanged unless the law requires otherwise.
167
 
168
  ## Citation
169