From f85d17604b53c50a71f48e7a1431a2a8a7bb4358 Mon Sep 17 00:00:00 2001
From: ale <ale@manalejandro.com>
Date: Mon, 14 Oct 2024 00:05:40 +0200
Subject: [PATCH] stream-json

---
 lib/apiswagger.js | 42 +++++++++++++++++++++++++++++++-----------
 package.json      |  1 +
 public/index.html |  2 +-
 public/main.js    |  5 ++---
 4 files changed, 35 insertions(+), 15 deletions(-)

diff --git a/lib/apiswagger.js b/lib/apiswagger.js
index dd94694..afb36b7 100644
--- a/lib/apiswagger.js
+++ b/lib/apiswagger.js
@@ -3,6 +3,10 @@ const nodeinfo = require('activitypub-express/pub/nodeinfo')
 module.exports = (app, client) => {
     const constant = require('./constant'),
         zlib = require('zlib'),
+        { pick } = require('stream-json/filters/Pick'),
+        { parser } = require('stream-json'),
+        { streamArray } = require('stream-json/streamers/StreamArray'),
+        { chain } = require('stream-chain'),
         clean = str => {
             return str.replace(/[/\\^$+?()`'¡¿¨!"·%&=;,\|\[\]{}]+/gmi, '')
         }
@@ -472,15 +476,24 @@ module.exports = (app, client) => {
                         }
                     }
                 }
-            })
-            const instances = result.hits?.hits?.length > 0 ? result.hits.hits : [],
-                instancescomment = instances.map(i => ({
-                    instance: i._source.instance, comment: i._source.blocks.find(block => block.domain === clean(req.params.instance)).comment
-                }))
+            }, { asStream: true, meta: false }),
+                instancescomment = [],
+                block_count = 0,
+                pipeline = chain([
+                    parser(),
+                    pick({ filter: 'hits.hits' }),
+                    streamArray(),
+                    data => {
+                        block_count++
+                        instancescomment.push({
+                            instance: data.value._source.instance, comment: data.value._source.blocks.find(block => block.domain === clean(req.params.instance)).comment
+                        })
+                    }
+                ])
+            result.pipe(pipeline)
             res.json({
-                block_count: instances.length,
-                instances: instancescomment,
-                took: result.took
+                block_count,
+                instances: instancescomment
             })
         } else {
             res.status(404).end()
@@ -501,15 +514,22 @@ module.exports = (app, client) => {
     app.use('/api/download_index', async (req, res) => {
         try {
             res.setHeader('Content-Type', 'application/gzip')
-            res.setHeader('Content-disposition', 'attachment; filename=fediblock-index.json.gz')
+            res.setHeader('Content-disposition', 'attachment; filename=fediblock-index.jsonl.gz')
             const result = await client.search({
                 index: constant.index,
                 size: 9999,
                 query: {
                     match_all: {}
                 }
-            }, { asStream: true, meta: false })
-            result.pipe(zlib.createGzip()).pipe(res)
+            }, { asStream: true, meta: false }),
+                pipeline = chain([
+                    parser(),
+                    pick({ filter: 'hits.hits' }),
+                    streamArray(),
+                    data => JSON.stringify(data.value._source) + '\n',
+                    zlib.createGzip()
+                ])
+            result.pipe(pipeline).pipe(res)
         } catch (e) {
             console.error(e)
             res.status(404).end()
diff --git a/package.json b/package.json
index aa1501c..d89f769 100644
--- a/package.json
+++ b/package.json
@@ -25,6 +25,7 @@
     "node-schedule": "^2.1.1",
     "parcel": "^2.12.0",
     "rotating-file-stream": "^3.2.5",
+    "stream-json": "^1.8.0",
     "swagger-jsdoc": "^6.2.8",
     "swagger-ui-express": "^5.0.1"
   }
diff --git a/public/index.html b/public/index.html
index b6c63fb..9469cbf 100644
--- a/public/index.html
+++ b/public/index.html
@@ -74,7 +74,7 @@
             <div class="modal-body">
                 <p>
                     <span id="blockcount"></span> public instances are block<span id="blockinstance"></span>
-                    <br /><small>(search in <span id="blocktook"></span>ms)</small>
+                    <br /><small id="blocktook"></small>
                 </p>
                 <ul id="blocklist"></ul>
             </div>
diff --git a/public/main.js b/public/main.js
index 9f96996..e834074 100644
--- a/public/main.js
+++ b/public/main.js
@@ -39,7 +39,6 @@ document.addEventListener('DOMContentLoaded', function () {
                 if (res && res.block_count >= 0) {
                     document.getElementById('blockcount').innerText = res.block_count
                     document.getElementById('blockinstance').innerText = 'ing ' + content
-                    document.getElementById('blocktook').innerText = res.took
                     var list = document.getElementById('blocklist'),
                         download = document.getElementById('download')
                     download.removeAttribute('href')
@@ -115,7 +114,7 @@ document.addEventListener('DOMContentLoaded', function () {
         fetch('/api/download_index', { signal: ac.signal }).then(async function (result) {
             var res = await result.blob(),
                 a = document.createElement('a')
-            a.download = 'fediblock-index.json.gz'
+            a.download = 'fediblock-index.jsonl.gz'
             a.href = URL.createObjectURL(res)
             a.type = 'application/gzip'
             a.target = '_blank'
@@ -213,7 +212,7 @@ document.addEventListener('DOMContentLoaded', function () {
                                 document.getElementById('blockinstance').innerHTML = 'ed by ' + (res.api ? '<a href="/api/detail_api/' + res.instance + '" title="API info for ' + res.instance + '" target="_blank">'
                                     + res.instance + '</a>' : res.instance) + (res.nodeinfo ? '&nbsp;<a href="/api/detail_nodeinfo/' + res.instance + '" title="Nodeinfo for ' + res.instance + '" target="_blank">&#9432;</a>' : '')
                                     + '<br/>Last update: ' + (new Date(res.last)).toLocaleString()
-                                document.getElementById('blocktook').innerText = res.took
+                                document.getElementById('blocktook').innerText = '(search in ' + res.took + 'ms)'
                                 if (csv.split('\n').length > 2) {
                                     download.href = window.URL.createObjectURL(new Blob([csv], { type: 'text/csv' }))
                                     download.download = 'fediblock-' + res.instance + '.csv'