From b4749f7896994d5b75e9acf2d3226959bc1f2af6 Mon Sep 17 00:00:00 2001
From: Patrice Chalin <pchalin@gmail.com>
Date: Sun, 5 Nov 2023 17:14:51 -0500
Subject: [PATCH] Adjust dict normalization script

---
 .../blog/2023/contributing-to-otel/index.md   |  2 +-
 scripts/normalize-cspell-front-matter.pl      | 38 ++++++++++++++-----
 2 files changed, 30 insertions(+), 10 deletions(-)

diff --git a/content/en/blog/2023/contributing-to-otel/index.md b/content/en/blog/2023/contributing-to-otel/index.md
index 81e04905b1cf..52b380ecf408 100644
--- a/content/en/blog/2023/contributing-to-otel/index.md
+++ b/content/en/blog/2023/contributing-to-otel/index.md
@@ -6,7 +6,7 @@ author:
   >- # If you have only one author, then add the single name on this line in quotes.
   [Adriana Villela](https://github.com/avillela) (Lightstep),
 canonical_url: https://medium.com/cloud-native-daily/how-to-contribute-to-opentelemetry-5962e8b2447e
-cSpell:ignore: EUWG, nolan, riaan, sayin, servian
+cSpell:ignore: sayin
 ---
 
 ![Sunset over the water over an orange sky, with long grass in the foreground.](turks-sunset.jpg)
diff --git a/scripts/normalize-cspell-front-matter.pl b/scripts/normalize-cspell-front-matter.pl
index 6cb4b734a523..555c7eed4f68 100755
--- a/scripts/normalize-cspell-front-matter.pl
+++ b/scripts/normalize-cspell-front-matter.pl
@@ -9,7 +9,7 @@
 my $lineLenLimit = 79;
 my $last_file = '';
 my $last_line = '';
-my %dictionary = getSiteWideDictWords('.vscode/cspell.json', '.textlintrc.json');
+my %dictionary = getSiteWideDictWords('.vscode/cspell.json', '.textlintrc.yml');
 
 while (<>) {
   if (/^\s*(spelling: |-\s*)?cSpell:ignore:?(.*)$/
@@ -52,18 +52,38 @@ sub getSiteWideDictWords {
   my $data = $json->decode($json_text);
   my %dictionary = map { $_ => 1 } @{ $data->{words} };
 
-  # Read the .textlintrc.json file
-  $fh = FileHandle->new($textlintrc_file, "r") or die "Could not open file '$textlintrc_file': $!";
-  $json_text = join "", $fh->getlines();
 
-  # Remove JSON comments
   $json_text =~ s/^\s*\/\/.*$//mg;
+  # Merge dictionaries
+  @dictionary{keys %textlintDictionary} = values %textlintDictionary;
+
+  return %dictionary;
+}
 
-  $data = $json->decode($json_text);
+sub processTextlintRcYml {
+  my $file_path = shift;
+  my $fh = FileHandle->new($file_path, "r") or die "Could not open file '$file_path': $!";
+  my @lines = $fh->getlines();
+  $fh->close();
+
+  my %dictionary;
+  my $indentation = '';
+  my $in_terms = 0;
+  foreach my $line (@lines) {
+    chomp $line;
+    if ($line =~ /^(\s*)terms:/) {
+      $indentation = $1 || '';
+      $in_terms = 1;
+      # print STDOUT "Found terms!";
+    } elsif ($line =~ /^$indentation  - (\w[^\s]*)$/ && $in_terms) {
+      my $term = $1;
+      $dictionary{$term} = 1 if $term;
+    } elsif ($line !~ /^ / && $in_terms) {
+      $in_terms = 0;
+    }
+  }
 
-  # Add terms from .textlintrc.json to dictionary
-  my @terms = @{ $data->{rules}->{terminology}->{terms} };
-  @dictionary{@terms} = (1) x @terms;
+  die "ERROR: no words read from '$file_path'!" unless %dictionary; # sanity check
 
   return %dictionary;
 }