From 86150e990885fb8101b8b84715f61d54ab1c7c1e Mon Sep 17 00:00:00 2001 From: Albert Sawczyn Date: Mon, 30 Sep 2024 16:40:37 +0200 Subject: [PATCH] feat: update documents --- data/datasets/nsa/documents.json.dvc | 4 ++-- juddges/data/nsa/scraper.py | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/data/datasets/nsa/documents.json.dvc b/data/datasets/nsa/documents.json.dvc index b35f378..b9b349d 100644 --- a/data/datasets/nsa/documents.json.dvc +++ b/data/datasets/nsa/documents.json.dvc @@ -1,5 +1,5 @@ outs: -- md5: aee1eb69003b5ab670b9f9ffd79693d9 - size: 116861042 +- md5: f035835f18297b85698a6f84d1c098a9 + size: 117137442 hash: md5 path: documents.json diff --git a/juddges/data/nsa/scraper.py b/juddges/data/nsa/scraper.py index c6748e6..d4f34f1 100644 --- a/juddges/data/nsa/scraper.py +++ b/juddges/data/nsa/scraper.py @@ -21,7 +21,7 @@ class IncorrectPage(Exception): class NSAScraper: def __init__( - self, user_agent: str, proxy_config: dict[str, str] | None = None, wait: bool = False + self, user_agent: str, proxy_config: dict[str, str] | None = None, wait: bool = True ) -> None: self.wait = wait self.browser = mechanicalsoup.StatefulBrowser(