From 704e956dafe1283bae774456f5663af9ca82d7c1 Mon Sep 17 00:00:00 2001 From: bel Date: Mon, 27 Jun 2022 18:34:31 -0600 Subject: [PATCH] cleanup --- scrape.sh | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/scrape.sh b/scrape.sh index cc02bee..41c0972 100644 --- a/scrape.sh +++ b/scrape.sh @@ -1,6 +1,5 @@ #! /bin/bash - main() { ensure @@ -114,6 +113,7 @@ should_scrape_book_chapter() { log foo:should_scrape_book_chapter $url if scrape_dndbeyond "$url" | grep -i "add to cart" &> /dev/null; then log "not purchased: $url" + _purge_cached_scrape_dndbeyond "$url" return 1 fi } @@ -154,9 +154,7 @@ list_chapters_in_book() { scrape_dndbeyond() { log foo:scrape_dndbeyond $@ - local d="${TMPDIR}/scrape_cache" - mkdir -p "$d" - local f="$d/$(echo "$*" | base64)" + local f="$(_scrape_dndbeyond_cache_f "$@")" if [ -f "$f" ]; then if cat "$f" | grep .; then log "foo:scrape_dndbeyond: yes cached: $f: $@" @@ -167,6 +165,19 @@ scrape_dndbeyond() { _scrape_dndbeyond "$@" | tee "$f" } +_purge_cached_scrape_dndbeyond () { + local f="$(_scrape_dndbeyond_cache_f "$@")" + if [ -f "$f" ]; then + rm -f "$f" + fi +} + +_scrape_dndbeyond_cache_f() { + local d="${TMPDIR}/scrape_cache" + mkdir -p "$d" + echo "$d/$(echo "$*" | base64 | tr -d '\n')" +} + _scrape_dndbeyond() { log foo:_scrape_dndbeyond $@ rate_limit_1s scrape_dndbeyond @@ -182,7 +193,7 @@ _scrape_dndbeyond() { -H 'sec-fetch-user: ?1' \ -H 'sec-fetch-dest: document' \ -H 'accept-language: en-US,en;q=0.9' \ - -H 'Cookie: Preferences=undefined; Preferences=undefined; _pxhd=qvJb-g60h1HSymUFEiAmb-Noh35mNbUmcFgIE0qEiisCj/BqXrgrrsRUlODrjAaSGknpuHBn0S2b4rGTBFvaIA==:5/4beZX4ojNLKbAeZH9rQFcGIQ/ijS7lK29wvhLkgB4O0L2GFpRuU5FOpRztEiKg9EoUZr2cPB2Z8w3DKVkaKJPlwcIt23DM5PgQLZEPoA8=; ResponsiveSwitch.DesktopMode=1; LoginState=0ae9ef26-a1ae-4cb2-8af7-f18c21b99811; G_ENABLED_IDPS=google; CobaltSession=eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..FawE7bpRlluZsWq7Iny1xQ.nBX_qdZq5varuXNGkl70QeJ03JIRT3sxxQi6iCzazt-SjlK1hEuINX0sK85gJKVT.8wlUwv…sername=squeaky2x3; UserInfo={"UserId":109926924,"UserJoinDate":"2021-03-25","UserSessionId":"5ce909a9-34aa-4b15-9b6d-00ee48e13e05"}; AWSELB=17A593B6CA59C3C4856B812F84CD401A582EF08335AE71D0834BC2696233572C85266F7F9F4B5BF89DD4952FDA9AD462250417F8B7C4975F1633599CC9D18D74C3B57CBA; AWSELBCORS=17A593B6CA59C3C4856B812F84CD401A582EF08335AE71D0834BC2696233572C85266F7F9F4B5BF89DD4952FDA9AD462250417F8B7C4975F1633599CC9D18D74C3B57CBA; ddbSiteBanner:bb279d28-f265-49d5-ad41-6d716915bbdf=true; WarningNotification.Lock=1' \ + -H 'Cookie: Preferences=undefined; Preferences=undefined; _pxhd=qvJb-g60h1HSymUFEiAmb-Noh35mNbUmcFgIE0qEiisCj/BqXrgrrsRUlODrjAaSGknpuHBn0S2b4rGTBFvaIA==:5/4beZX4ojNLKbAeZH9rQFcGIQ/ijS7lK29wvhLkgB4O0L2GFpRuU5FOpRztEiKg9EoUZr2cPB2Z8w3DKVkaKJPlwcIt23DM5PgQLZEPoA8=; ResponsiveSwitch.DesktopMode=1; LoginState=0ae9ef26-a1ae-4cb2-8af7-f18c21b99811; G_ENABLED_IDPS=google; CobaltSession=eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..FawE7bpRlluZsWq7Iny1xQ.nBX_qdZq5varuXNGkl70QeJ03JIRT3sxxQi6iCzazt-SjlK1hEuINX0sK85gJKVT.8wlUwv…sername=squeaky2x3; UserInfo={"UserId":109926924,"UserJoinDate":"2021-03-25","UserSessionId":"5ce909a9-34aa-4b15-9b6d-00ee48e13e05"}; ddbSiteBanner:bb279d28-f265-49d5-ad41-6d716915bbdf=true; AWSELB=17A593B6CA59C3C4856B812F84CD401A582EF083E14A2E133FB99235E1D3CC671549BC8DF3DD0A01DD4815C023F988F55E0E95275F44EAFDFB8393A8DEFA33DD0B90284D; AWSELBCORS=17A593B6CA59C3C4856B812F84CD401A582EF083E14A2E133FB99235E1D3CC671549BC8DF3DD0A01DD4815C023F988F55E0E95275F44EAFDFB8393A8DEFA33DD0B90284D; WarningNotification.Lock=1' \ -H 'dnt: 1' \ --compressed return