Spaces:
Running
Running
| """Shared pytest fixtures for ResearchRadar tests.""" | |
| import tempfile | |
| from pathlib import Path | |
| import pytest | |
| from src.ingestion.base_loader import PaperRecord | |
| from src.storage.sqlite_db import SQLiteDB | |
| def sample_papers() -> list[PaperRecord]: | |
| """A small set of fake papers for testing.""" | |
| return [ | |
| PaperRecord( | |
| source_id="P18-1001", | |
| source="hf_acl_ocl", | |
| title="Attention Is All You Need (Not Really)", | |
| abstract="We propose a novel transformer architecture for machine translation.", | |
| authors=["Alice Smith", "Bob Jones"], | |
| year=2018, | |
| venue="acl", | |
| volume="long", | |
| full_text=None, | |
| url="https://example.com/paper1", | |
| ), | |
| PaperRecord( | |
| source_id="D19-1234", | |
| source="hf_acl_ocl", | |
| title="LoRA: Low-Rank Adaptation of Large Language Models", | |
| abstract="We propose LoRA, a parameter-efficient fine-tuning method using low-rank matrices.", | |
| authors=["Carol Chen"], | |
| year=2019, | |
| venue="emnlp", | |
| volume="long", | |
| full_text="Introduction\nWe study parameter-efficient methods.\n\nMethod\nLoRA decomposes weight updates.", | |
| url="https://example.com/paper2", | |
| ), | |
| PaperRecord( | |
| source_id="2022.acl-long.100", | |
| source="acl_anthology", | |
| title="Contrastive Learning for NLP: A Survey", | |
| abstract="This paper surveys contrastive learning approaches applied to NLP tasks.", | |
| authors=["Dave Wilson", "Eve Taylor"], | |
| year=2022, | |
| venue="acl", | |
| volume="long", | |
| ), | |
| ] | |
| def tmp_db(tmp_path) -> SQLiteDB: | |
| """A temporary SQLite database for testing.""" | |
| db = SQLiteDB(tmp_path / "test.db") | |
| db.create_schema() | |
| return db | |