From faa59f192c00dadd84165fe14a90307c908e6ab2 Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Fri, 8 Mar 2019 21:05:03 +0100 Subject: possibly freshen raw papers --- scraper/s2-raw-papers.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'scraper/s2-raw-papers.py') diff --git a/scraper/s2-raw-papers.py b/scraper/s2-raw-papers.py index 612c8099..8881cda0 100644 --- a/scraper/s2-raw-papers.py +++ b/scraper/s2-raw-papers.py @@ -18,10 +18,10 @@ def fetch_raw_papers(fn): lines = read_csv(fn, keys=False) parallelize(fetch_raw_paper, lines) -def fetch_raw_paper(paper_id): +def fetch_raw_paper(paper_id, freshen=False): os.makedirs(make_raw_paper_path(paper_id), exist_ok=True) paper_fn = make_raw_paper_fn(paper_id) - if os.path.exists(paper_fn): + if os.path.exists(paper_fn) and not freshen: paper = read_json(paper_fn) else: paper = s2.raw_paper(paper_id) -- cgit v1.2.3-70-g09d2 From c0d78dc072bc9f9b6e373f49b74d98d393d0e96a Mon Sep 17 00:00:00 2001 From: Jules Laplace Date: Fri, 8 Mar 2019 21:08:13 +0100 Subject: possibly freshen raw papers --- scraper/s2-raw-papers.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'scraper/s2-raw-papers.py') diff --git a/scraper/s2-raw-papers.py b/scraper/s2-raw-papers.py index 8881cda0..67ff0b65 100644 --- a/scraper/s2-raw-papers.py +++ b/scraper/s2-raw-papers.py @@ -33,7 +33,7 @@ def fetch_raw_paper(paper_id, freshen=False): if 'responseType' in paper and paper['responseType'] == 'CANONICAL': write_json(paper_fn, paper) - paper = s2.raw_paper(data['canonicalId']) + paper = s2.raw_paper(paper['canonicalId']) paper_fn = make_raw_paper_fn(paper_id) return paper -- cgit v1.2.3-70-g09d2