summaryrefslogtreecommitdiff
path: root/src/index-terms.sh
diff options
context:
space:
mode:
authorLouie S <louie@example.com>2023-08-26 12:08:28 -0400
committerLouie S <louie@example.com>2023-08-26 12:08:28 -0400
commitc125d1179850a4742756b2580343f7fbd73fca88 (patch)
tree1621659b7c470fa477b5aae1d9077fd5ce0e87d4 /src/index-terms.sh
First commitHEADmaster
Diffstat (limited to 'src/index-terms.sh')
-rwxr-xr-xsrc/index-terms.sh33
1 files changed, 33 insertions, 0 deletions
diff --git a/src/index-terms.sh b/src/index-terms.sh
new file mode 100755
index 0000000..f74af03
--- /dev/null
+++ b/src/index-terms.sh
@@ -0,0 +1,33 @@
+#!/usr/bin/env sh
+
+# shellcheck source=./lib/create_table
+. "$(dirname "$0")"/lib/create_table
+# shellcheck source=./lib/insert
+. "$(dirname "$0")"/lib/insert
+
+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 "$@"