Merge pull request #4773 from ClickHouse/vale-fixes

Auditing Vale Rules and Including a Report Script

Author: dhtclk

.vale.ini

@@ -1,5 +1,17 @@
 StylesPath = styles
-MinAlertLevel = warning
 
-[*.{md}]
+# Exclude with more specific patterns
+# [!docs/engines/**/*.md]
+# [!docs/development/**/*.md]
+# [!docs/interfaces/**/*.md]
+# [!docs/operations/**/*.md]
+# [!docs/sql-reference/**/*.md]
+[!docs/whats-new/**/*.md]
+[!docs/releases/**/*.md]
+
+# Only check .md files (not .mdx)
+[docs/**/*.md]
 BasedOnStyles = ClickHouse
+
+ClickHouse.SentenceLength = NO
+ClickHouse.FutureTense = NO

scripts/vale/vale-report.sh

@@ -0,0 +1,41 @@
+#!/bin/bash
+# vale-report.sh - Generate Vale report
+
+vale docs/ \
+  --glob='!docs/whats-new/**/*.md' \
+  --output=JSON > vale-report.json 2>&1
+
+# Check if vale-report.json is valid JSON
+if ! jq empty vale-report.json 2>/dev/null; then
+    echo "Error: Vale did not produce valid JSON output"
+    echo ""
+    echo "Vale output:"
+    cat vale-report.json
+    exit 1
+fi
+
+echo "================================================"
+echo "Vale Report - $(date '+%Y-%m-%d')"
+echo "================================================"
+echo ""
+
+WARNINGS=$(jq '[.[] | .[] | select(.Severity == "warning")] | length' vale-report.json)
+SUGGESTIONS=$(jq '[.[] | .[] | select(.Severity == "suggestion")] | length' vale-report.json)
+
+echo "Warnings: $WARNINGS"
+echo "Suggestions: $SUGGESTIONS"
+echo ""
+
+echo "Top Warning Rules:"
+jq -r '[.[] | .[] | select(.Severity == "warning") | .Check] | group_by(.) | map({rule: .[0], count: length}) | sort_by(-.count) | .[:5] | .[] | "\(.count)\t\(.rule)"' vale-report.json | column -t -s $'\t'
+echo ""
+
+echo "Top Files to Fix:"
+jq -r 'to_entries | map({file: .key, count: [.value[] | select(.Severity == "warning")] | length}) | map(select(.count > 0)) | sort_by(-.count) | .[:15] | .[] | "\(.count)\t\(.file)"' vale-report.json | column -t -s $'\t'
+echo ""
+
+echo "================================================"
+echo "To fix a file:"
+echo "   vale docs/path/to/file.md"
+echo "   vim docs/path/to/file.md"
+echo "================================================"

styles/ClickHouse/CodeblockFences.yml

@@ -6,7 +6,7 @@ extends: existence
 message: "Instead of '%s' for the code block, use yaml, ruby, plaintext, markdown, javascript, shell, go, python, dockerfile, or typescript."
 link: https://docs.gitlab.com/development/documentation/styleguide/#code-blocks
 vocab: false
-level: warning
+level: suggestion
 scope: raw
 raw:
   - '\`\`\`(yml|rb|text|md|bash|sh\n|js\n|golang\n|py\n|docker\n|ts|irb)'

styles/ClickHouse/FutureTense.yml

@@ -8,7 +8,7 @@ message: "Instead of future tense '%s', use present tense."
 ignorecase: true
 nonword: true
 vocab: false
-level: warning
+level: suggestion
 link: https://docs.gitlab.com/development/documentation/styleguide/word_list/#future-tense
 tokens:
   - (going to|will|won't)[ \n:]\w*

styles/ClickHouse/Uppercase.yml

@@ -13,6 +13,7 @@ first: '\b([A-Z]{3,5})\b'
 second: '(?:\b[A-Z][a-z]+ )+\(([A-Z]{3,5})\)'
 # ... with the exception of these:
 exceptions:
+  - ACH
   - ACL
   - AJAX
   - ALL
@@ -163,6 +164,7 @@ exceptions:
   - OTP
   - OWASP
   - PAT
+  - PAYG
   - PCI-DSS
   - PDF
   - PEM
@@ -255,6 +257,7 @@ exceptions:
   - URI
   - URL
   - USB
+  - USD
   - UTC
   - UTF
   - UUID