From 45ca8da1d708f0c8a5991117bba2b62a4790eafc Mon Sep 17 00:00:00 2001
From: Michal Piechowiak <misiek.piechowiak@gmail.com>
Date: Fri, 6 Apr 2018 20:18:28 +0200
Subject: [PATCH] [gatsby-source-wordpress] take advantage of `modified` field
 in media rest endpoint to not request file if we already have it

---
 .../src/gatsby-node.js                        |  3 +-
 .../gatsby-source-wordpress/src/normalize.js  | 49 ++++++++++++++-----
 2 files changed, 38 insertions(+), 14 deletions(-)

diff --git a/packages/gatsby-source-wordpress/src/gatsby-node.js b/packages/gatsby-source-wordpress/src/gatsby-node.js
index b9fc54f522f8f..98a9695592872 100644
--- a/packages/gatsby-source-wordpress/src/gatsby-node.js
+++ b/packages/gatsby-source-wordpress/src/gatsby-node.js
@@ -34,7 +34,7 @@ exports.sourceNodes = async (
     excludedRoutes = [],
   }
 ) => {
-  const { createNode } = boundActionCreators
+  const { createNode, touchNode } = boundActionCreators
   _verbose = verboseOutput
   _siteURL = `${protocol}://${baseUrl}`
   _useACF = useACF
@@ -99,6 +99,7 @@ exports.sourceNodes = async (
     store,
     cache,
     createNode,
+    touchNode,
     _auth,
   })
 
diff --git a/packages/gatsby-source-wordpress/src/normalize.js b/packages/gatsby-source-wordpress/src/normalize.js
index 24e2d9045686d..ecc56ab8d7428 100644
--- a/packages/gatsby-source-wordpress/src/normalize.js
+++ b/packages/gatsby-source-wordpress/src/normalize.js
@@ -393,27 +393,50 @@ exports.downloadMediaFiles = async ({
   store,
   cache,
   createNode,
+  touchNode,
   _auth,
 }) =>
   Promise.all(
     entities.map(async e => {
-      let fileNode
+      let fileNodeID
       if (e.__type === `wordpress__wp_media`) {
-        try {
-          fileNode = await createRemoteFileNode({
-            url: e.source_url,
-            store,
-            cache,
-            createNode,
-            auth: _auth,
-          })
-        } catch (e) {
-          // Ignore
+        const mediaDataCacheKey = `wordpress-media-${e.wordpress_id}`
+        const cacheMediaData = await cache.get(mediaDataCacheKey)
+
+        // If we have cached media data and it wasn't modified, reuse
+        // previously created file node to not try to redownload
+        if (cacheMediaData && e.modified === cacheMediaData.modified) {
+          fileNodeID = cacheMediaData.fileNodeID
+          touchNode(cacheMediaData.fileNodeID)
+        }
+
+        // If we don't have cached data, download the file
+        if (!fileNodeID) {
+          try {
+            const fileNode = await createRemoteFileNode({
+              url: e.source_url,
+              store,
+              cache,
+              createNode,
+              auth: _auth,
+            })
+
+            if (fileNode) {
+              fileNodeID = fileNode.id
+
+              await cache.set(mediaDataCacheKey, {
+                fileNodeID,
+                modified: e.modified,
+              })
+            }
+          } catch (e) {
+            // Ignore
+          }
         }
       }
 
-      if (fileNode) {
-        e.localFile___NODE = fileNode.id
+      if (fileNodeID) {
+        e.localFile___NODE = fileNodeID
         delete e.media_details.sizes
       }