Merge pull request #3013 from owid/gdoc-migration-normalize-headings

GDoc migration: normalize headings
owid · Dec 26, 2023 · 72667ca · 72667ca
2 parents 42c97f4 + 82fde64
commit 72667ca
Show file tree

Hide file tree

Showing 2 changed files with 32 additions and 4 deletions.
diff --git a/db/migrateWpPostsToArchieMl.ts b/db/migrateWpPostsToArchieMl.ts
@@ -17,6 +17,7 @@ import {
     withoutEmptyOrWhitespaceOnlyTextBlocks,
     convertAllWpComponentsToArchieMLBlocks,
     adjustHeadingLevels,
+    findMinimumHeadingLevel,
 } from "./model/Gdoc/htmlToEnriched.js"
 import { getRelatedCharts, isPostCitable } from "./wpdb.js"
 
@@ -92,7 +93,7 @@ const migrate = async (): Promise<void> => {
         "created_at_in_wordpress",
         "updated_at",
         "featured_image"
-    ).from(db.knexTable(Post.postsTable)) //.where("id", "=", "24808"))
+    ).from(db.knexTable(Post.postsTable)) //.where("id", "=", "54759"))
 
     for (const postRaw of rawPosts) {
         try {
@@ -143,8 +144,11 @@ const migrate = async (): Promise<void> => {
 
             // Heading levels used to start at 2, in the new layout system they start at 1
             // This function iterates all blocks recursively and adjusts the heading levels inline
-            // If the article is an entry, we also put an <hr /> above and below h1's
-            adjustHeadingLevels(archieMlBodyElements, isEntry)
+            // If the article is an entry, we also put an <hr /> above and below h1's. The adjustment
+            // pulls heading levels up so that entries end up with h1s and others with h2s at the top.
+            const minHeadingLevel =
+                findMinimumHeadingLevel(archieMlBodyElements)
+            adjustHeadingLevels(archieMlBodyElements, minHeadingLevel, isEntry)
 
             if (relatedCharts.length) {
                 const indexOfFirstHeading = archieMlBodyElements.findIndex(

diff --git a/db/model/Gdoc/htmlToEnriched.ts b/db/model/Gdoc/htmlToEnriched.ts
@@ -395,8 +395,28 @@ export function convertAllWpComponentsToArchieMLBlocks(
     })
 }
 
+export function findMinimumHeadingLevel(
+    blocks: OwidEnrichedGdocBlock[]
+): number {
+    let minBlockLevel = 6
+    for (const block of blocks) {
+        if (block.type === "heading") {
+            minBlockLevel = Math.min(block.level, minBlockLevel)
+        } else if ("children" in block) {
+            minBlockLevel = Math.min(
+                findMinimumHeadingLevel(
+                    block.children as OwidEnrichedGdocBlock[]
+                ),
+                minBlockLevel
+            )
+        }
+    }
+    return minBlockLevel
+}
+
 export function adjustHeadingLevels(
     blocks: OwidEnrichedGdocBlock[],
+    minHeadingLevel: number,
     isEntry: boolean
 ): void {
     for (let i = 0; i < blocks.length; i++) {
@@ -411,10 +431,14 @@ export function adjustHeadingLevels(
                 blocks.splice(i + 2, 0, { ...hr })
                 i += 2
             }
-            block.level = Math.max(1, block.level - 1)
+            const correction = isEntry
+                ? minHeadingLevel - 1
+                : Math.max(0, minHeadingLevel - 2)
+            block.level = block.level - correction
         } else if ("children" in block) {
             adjustHeadingLevels(
                 block.children as OwidEnrichedGdocBlock[],
+                minHeadingLevel,
                 isEntry
             )
         }