Contract: Scraper Keyword Types

Type: Internal shared enum（backend API validation） Location: shared/enums/scraper_keyword.py | Date: 2026-06-04

VALID_KEYWORD_TYPES

Controls which keyword_type values are accepted by POST /scraper-keywords.

VALID_KEYWORD_TYPES: frozenset[str] = frozenset({
    "rss",
    "arxiv_keyword",             # Legacy: system ignores at scrape time
    "arxiv_category",
    "semantic_scholar_keyword",  # New in this feature
})

Usage

Backend validation（backend/routers/scraper_keywords.py）

@field_validator("keyword_type")
@classmethod
def _check_type(cls, v: str) -> str:
    if v not in VALID_KEYWORD_TYPES:
        raise ValueError(f"keyword_type must be one of {sorted(VALID_KEYWORD_TYPES)}")
    return v

Scraper factory mapping

keyword_type	Used by	Notes
rss	RssScraper	Regex filter on RSS entry titles
arxiv_keyword	ArxivScraper	Stored but ignored at scrape time（deprecated）
arxiv_category	ArxivScraper	Category code, e.g. cs.LG
semantic_scholar_keyword	SemanticScholarScraper	Free-text search keyword

Frontend keyword_type per source

Source type	Keyword types available in UI
rss	rss
arxiv	arxiv_category only（arxiv_keyword UI removed）
semantic_scholar	semantic_scholar_keyword
blog	None