diff options
Diffstat (limited to 'scraper/s2-papers.py')
| -rw-r--r-- | scraper/s2-papers.py | 18 |
1 files changed, 0 insertions, 18 deletions
diff --git a/scraper/s2-papers.py b/scraper/s2-papers.py index f38bb800..bf77a734 100644 --- a/scraper/s2-papers.py +++ b/scraper/s2-papers.py @@ -42,23 +42,5 @@ def fetch_papers(index, depth): paper = fetch_paper(paper_id) # get all of the paper's citations -def fetch_paper(paper_id): - os.makedirs('./datasets/s2/papers/{}/{}'.format(paper_id[0:2], paper_id), exist_ok=True) - paper_fn = './datasets/s2/papers/{}/{}/paper.json'.format(paper_id[0:2], paper_id) - if os.path.exists(paper_fn): - return read_json(paper_fn) - print(paper_id) - paper = s2.paper(paper_id) - if paper is None: - print("Got none paper??") - time.sleep(random.randint(20, 30)) - paper = s2.paper(paper_id) - if paper is None: - print("Paper not found") - return None - write_json(paper_fn, paper) - time.sleep(random.randint(5, 10)) - return paper - if __name__ == '__main__': fetch_papers() |