From e72408b48c1f32c4b5b34bd3861584d98b362255 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Tue, 18 May 2021 11:14:54 -0700 Subject: sitemaps: PDF sitemaps --- extra/sitemap/pdf_urls_query.sh | 14 ++++++++++++++ 1 file changed, 14 insertions(+) create mode 100755 extra/sitemap/pdf_urls_query.sh (limited to 'extra/sitemap/pdf_urls_query.sh') diff --git a/extra/sitemap/pdf_urls_query.sh b/extra/sitemap/pdf_urls_query.sh new file mode 100755 index 0000000..fb1a4b8 --- /dev/null +++ b/extra/sitemap/pdf_urls_query.sh @@ -0,0 +1,14 @@ +#!/usr/bin/env bash + +set -e # fail on error +set -u # fail if variable not set in substitution +set -o pipefail # fail if part of a '|' command fails + + +# query for specific works; about 8.6 million circa 2021-04-29 +fatcat-cli search scholar 'doc_type:work (fulltext.access_type:ia_file OR fulltext.access_type:wayback) (NOT biblio.arxiv_id:*) (NOT biblio.pmcid:*) (NOT biblio.publisher_type:big5) (year:<1926 OR tags:*)' --index-json --limit 0 \ + | pv -l \ + | jq '[.fulltext.access_type, .fulltext.access_url] | @tsv' -r \ + | rg -v '^null' \ + | ./transform_access_url.py \ + | split --lines 20000 - sitemap-pdfs- -d -a 5 --additional-suffix .txt -- cgit v1.2.3