nikolas-rauscher / ARDI-Scholarly-QALD

ARDI-Project Scholarly QALD Challenge 2024
GNU General Public License v3.0
2 stars 0 forks source link

Small bug #52

Closed sefeoglu closed 1 month ago

sefeoglu commented 1 month ago

Hi, Some of wikipedia context is an empty list, so this should be checked. You can fix it with following lines:

        wiki_context = ""
        for wiki_text in example['wiki_data']:
            # sentences = ['. '.join(list(item.values())) for item in wiki_text]
            if len(wiki_text)>0:
                sentences=str(wiki_text).split('.')
                wiki_evidence = evidence_sentence_selection(
                    example['question'], sentences, conserved_percentage=0.2, max_num=40
                )
                wiki_context = '. '.join(wiki_evidence)

Best,

https://github.com/nikolas-rauscher/ARDI-Scholarly-QALD/blob/cf7d8b84c78f4a011c95bc2bfedd30e9346bb166/src/models/prepare_prompts_context.py#L47-L54