From b949528c9a8bb9c3c7bd6723bd81d2c917b6d50e Mon Sep 17 00:00:00 2001
From: Louie S <louie@example.com>
Date: Wed, 19 Jun 2024 17:02:00 -0400
Subject: Update script

---
 src/index-terms.sh | 33 +++++++++++++++++++++++++++++++++
 1 file changed, 33 insertions(+)
 create mode 100755 src/index-terms.sh

(limited to 'src/index-terms.sh')

diff --git a/src/index-terms.sh b/src/index-terms.sh
new file mode 100755
index 0000000..c9344d4
--- /dev/null
+++ b/src/index-terms.sh
@@ -0,0 +1,33 @@
+#!/usr/bin/env sh
+
+# shellcheck source=../../../scripts/create_table.sh
+. "$(dirname "$0")"/../../../scripts/create_table.sh
+# shellcheck source=../../../scripts/insert.sh
+. "$(dirname "$0")"/../../../scripts/insert.sh
+
+TYPE="$1"
+shift
+DB_PATH="$1"
+shift
+
+insert_index_terms() {
+	# Get each term from an index page and insert
+	while [ -n "$1" ]; do
+		grep -Eo "<a href.*</a>:" "$1" | while read -r line; do
+			insert_term "$line"
+		done
+
+		shift
+	done
+}
+
+insert_term() {
+	LINK="$1"
+	NAME="$(echo "$LINK" | pup -p 'a text{}' | sed 's/\"\"//g' | tr -d \\n)"
+	PAGE_PATH="$(echo "$LINK" | pup -p 'a attr{href}')"
+
+	insert "$DB_PATH" "$NAME" "$TYPE" "$PAGE_PATH"
+}
+
+create_table "$DB_PATH"
+insert_index_terms "$@"
-- 
cgit