diff options
| author | Jules Laplace <julescarbon@gmail.com> | 2018-12-16 13:40:28 +0100 |
|---|---|---|
| committer | Jules Laplace <julescarbon@gmail.com> | 2018-12-16 13:40:28 +0100 |
| commit | 797b10edaf276305c2b64b25a564c4bdea21f555 (patch) | |
| tree | 3040270c2304ac2871f06deb7240e44cb39af35c /scraper/s2-search.py | |
| parent | 7ade5fb2b8826725a63853d8906b17c0091e4707 (diff) | |
datasetsss
Diffstat (limited to 'scraper/s2-search.py')
| -rw-r--r-- | scraper/s2-search.py | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/scraper/s2-search.py b/scraper/s2-search.py index e943053a..9ec20cc9 100644 --- a/scraper/s2-search.py +++ b/scraper/s2-search.py @@ -25,7 +25,8 @@ totalResults @click.command() @click.option('--index', '-n', default=0, help='Index of CSV (query,)') -def fetch_entries(index): +@click.option('--refresh/--no-refresh', '-f', help='Force it to query the paper API again') +def fetch_entries(index, refresh): keys, lines = read_citation_list(index) citation_lookup = [] s2 = SemanticScholarAPI() @@ -39,7 +40,7 @@ def fetch_entries(index): dump_fn = './datasets/s2/dumps/{}.json'.format(key) entry_fn = './datasets/s2/entries/{}.json'.format(key) result = None - if os.path.exists(entry_fn): + if not refresh and os.path.exists(entry_fn): result = read_json(entry_fn) else: results = s2.search(clean_title) |
