From 7c6afa0a21883dc8037f3d021246db24eef39b41 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@robocracy.org>
Date: Mon, 29 Nov 2021 15:02:27 -0800
Subject: clean up extra/ folder a bit

---
 extra/checks/.gitignore      |  2 --
 extra/checks/check_extid.sh  | 49 --------------------------------------------
 extra/checks/check_hashes.sh | 16 ---------------
 extra/checks/check_issnl.sh  | 15 --------------
 4 files changed, 82 deletions(-)
 delete mode 100644 extra/checks/.gitignore
 delete mode 100755 extra/checks/check_extid.sh
 delete mode 100755 extra/checks/check_hashes.sh
 delete mode 100755 extra/checks/check_issnl.sh

(limited to 'extra/checks')

diff --git a/extra/checks/.gitignore b/extra/checks/.gitignore
deleted file mode 100644
index 431c3bbc..00000000
--- a/extra/checks/.gitignore
+++ /dev/null
@@ -1,2 +0,0 @@
-*.txt
-*.tsv
diff --git a/extra/checks/check_extid.sh b/extra/checks/check_extid.sh
deleted file mode 100755
index f74f50b6..00000000
--- a/extra/checks/check_extid.sh
+++ /dev/null
@@ -1,49 +0,0 @@
-#!/usr/bin/env bash
-
-set -e -u -o pipefail
-
-export LC_ALL=C
-
-EXTID_FILE=$1
-
-zcat $EXTID_FILE \
-    | awk '{print $3 "\t" $1}' \
-    | rg -v '^\t' \
-    | sort -S 4G \
-    > doi_ident.tsv
-zcat $EXTID_FILE \
-    | awk '{print $4 "\t" $1}' \
-    | rg -v '^\t' \
-    | sort -S 4G \
-    > pmid_ident.tsv
-zcat $EXTID_FILE \
-    | awk '{print $5 "\t" $1}' \
-    | rg -v '^\t' \
-    | sort -S 4G \
-    > pmcid_ident.tsv
-zcat $EXTID_FILE \
-    | awk '{print $6 "\t" $1}' \
-    | rg -v '^\t' \
-    | sort -S 4G \
-    > wikidata_ident.tsv
-
-# these identifiers aren't fixed-width, so we need to join (sigh)
-cut -f1 doi_ident.tsv \
-    | uniq -d \
-    | join -t$'\t' - doi_ident.tsv \
-    > doi_ident.dupes.tsv
-cut -f1 pmid_ident.tsv \
-    | uniq -d \
-    | join -t$'\t' - pmid_ident.tsv \
-    > pmid_ident.dupes.tsv
-cut -f1 pmcid_ident.tsv \
-    | uniq -d \
-    | join -t$'\t' - pmcid_ident.tsv \
-    > pmcid_ident.dupes.tsv
-cut -f1 wikidata_ident.tsv \
-    | uniq -d \
-    | join -t$'\t' - wikidata_ident.tsv \
-    > wikidata_ident.dupes.tsv
-
-wc -l doi_ident.dupes.tsv pmid_ident.dupes.tsv pmcid_ident.dupes.tsv wikidata_ident.dupes.tsv >> counts.txt
-
diff --git a/extra/checks/check_hashes.sh b/extra/checks/check_hashes.sh
deleted file mode 100755
index 94102329..00000000
--- a/extra/checks/check_hashes.sh
+++ /dev/null
@@ -1,16 +0,0 @@
-#!/usr/bin/env bash
-
-set -e -u -o pipefail
-
-export LC_ALL=C
-
-HASH_FILE=$1
-
-zcat $HASH_FILE \
-    | awk '{print $3 "\t" $1}' \
-    | rg -v '^\t' \
-    | sort -S 4G \
-    | uniq -d -w 40 \
-    > sha1_ident.dupes.tsv
-
-wc -l sha1_ident.dupes.tsv >> counts.txt
diff --git a/extra/checks/check_issnl.sh b/extra/checks/check_issnl.sh
deleted file mode 100755
index a28695e7..00000000
--- a/extra/checks/check_issnl.sh
+++ /dev/null
@@ -1,15 +0,0 @@
-#!/usr/bin/env bash
-
-set -e -u -o pipefail
-
-export LC_ALL=C
-
-CONTAINER_DUMP=$1
-
-zcat $CONTAINER_DUMP \
-    | jq '[.issnl, .ident] | @tsv' -r \
-    | sort -S 4G \
-    | uniq -D -w 9 \
-    > issnl_ident.dupes.tsv
-
-wc -l issnl_ident.dupes.tsv >> counts.txt
-- 
cgit v1.2.3