hard tuning

gendb.sh

@@ -5,6 +5,6 @@ then
    exit 1;
 else
    echo "Index did not exist, creating..." ;
-   curl -H 'Content-Type: application/json' -H 'Accept: application/json' -X PUT -d '{ "settings" : { "number_of_shards" : 1 }, "mappings" : { "claim" : { "properties" : { "value" : { "type" : "nested" }, "suggest_name": { "type": "completion" }, "suggest_desc": { "type": "completion" } } } } }' http://localhost:9200/claims;
+   curl -H 'Content-Type: application/json' -H 'Accept: application/json' -X PUT -d '{ "settings" : { "number_of_shards" : 1 }, "mappings" : { "claim" : { "properties" : { "value" : { "type" : "nested" }, "suggest_name": { "type": "completion" }, "suggest_desc": { "type": "completion" }, "transaction_time": { "type": "date" } } } } }' http://localhost:9200/claims;
    exit 0;
 fi

server/controllers/lighthouse.js

@@ -80,6 +80,64 @@ function getResults (input) {
       },
     },
   };
+  const newerBoost = {
+    'function_score': {
+      'score_mode': 'sum', // All functions outputs get summed
+      'boost_mode': 'multiply', // The documents relevance is multiplied with the sum
+
+      'functions': [
+        {
+          // The relevancy of old posts is multiplied by at least one.
+          // Remove if you want to exclude old posts
+          'weight': 0.6,
+        },
+        {
+          // Published this week get a big boost
+          'weight': 85,
+          'gauss' : {
+            'transaction_time': { // <- Change to your date field name
+              'origin': Date.now(), // Change to current date
+              'scale' : '7d',
+              'decay' : 0.6,
+            },
+          },
+        },
+        {
+          // Published this month get a big boost
+          'weight': 60,
+          'gauss' : {
+            'transaction_time': { // <- Change to your date field name
+              'origin': Date.now(), // Change to current date
+              'scale' : '31d',
+              'decay' : 0.55,
+            },
+          },
+        },
+        {
+          // Published this month get a big boost
+          'weight': 50,
+          'gauss' : {
+            'transaction_time': { // <- Change to your date field name
+              'origin': Date.now(), // Change to current date
+              'scale' : '62d',
+              'decay' : 0.5,
+            },
+          },
+        },
+        {
+          // Published this year get a boost
+          'weight': 40,
+          'gauss' : {
+            'transaction_time': { // <- Change to your date field name
+              'origin': Date.now(), // Change to current date
+              'scale' : '356d',
+              'decay' : 0.2,
+            },
+          },
+        },
+      ],
+    },
+  };
   const funcScoreClaimWeight = { // 100 LBC adds 1 point to the score
     'function_score': {
       'field_value_factor': {
@@ -112,7 +170,9 @@ function getResults (input) {
     let conCatTerm = '';
     let phraseTerm = '';
     escapedQuery.split(' ').every((term, index) => {
-      if (index === 4) { return false }
+      if (index === 4) {
+        return false;
+      }
       phraseTerm = phraseTerm + ' ' + term;
       conCatTerm = conCatTerm + term;
       queries.push(
@@ -144,10 +204,10 @@ function getResults (input) {
           },
         },
         {
-          'prefix': { 'name': { 'value': '@' + escapedQuery, 'boost': 10 } },
+          'prefix': {'name': {'value': '@' + escapedQuery, 'boost': 10}},
         },
         {
-          'prefix': { 'name': { 'value': escapedQuery, 'boost': 10 } },
+          'prefix': {'name': {'value': escapedQuery, 'boost': 10}},
         },
       );
     });
@@ -156,7 +216,9 @@ function getResults (input) {
   const splitATD = () => {
     let queries = [];
     escapedQuery.split(' ').every((term, index) => {
-      if (index === 4) { return false }
+      if (index === 4) {
+        return false;
+      }
       queries.push({ // Contains search term in Author, Title, Description
         'query_string': {
           'query' : `*${term}*`,
@@ -277,6 +339,7 @@ function getResults (input) {
         'bool': {
           'should': [
             conBoost,
+            newerBoost,
             funcScoreClaimWeight,
             funcScoreChannelWeight,
             channelIdentifier,
@@ -360,7 +423,7 @@ function getAutoCompleteQuery (query) {
 
 function getFilters (input) {
   var filters = [];
-  var bidStateFilter = {'bool': {'must_not': {'match': { 'bid_state': 'Accepted' }}}};
+  var bidStateFilter = {'bool': {'must_not': {'match': {'bid_state': 'Accepted'}}}};
   if (input.nsfw === 'true' || input.nsfw === 'false') {
     const nsfwFilter = {'match': {'value.stream.metadata.nsfw': input.nsfw}};
     filters.push(nsfwFilter);
@@ -446,7 +509,12 @@ function getStatus () {
     rp(`http://localhost:9200/claims/_stats`)
       .then(function (data) {
         data = JSON.parse(data);
-        resolve({status: getStats(), spaceUsed: pretty(data._all.total.store.size_in_bytes, true), claimsInIndex: data._all.total.indexing.index_total, totSearches: data._all.total.search.query_total});
+        resolve({
+          status       : getStats(),
+          spaceUsed    : pretty(data._all.total.store.size_in_bytes, true),
+          claimsInIndex: data._all.total.indexing.index_total,
+          totSearches  : data._all.total.search.query_total,
+        });
       })
       .catch(function (err) {
         reject(err);
@@ -459,11 +527,13 @@ function getWashedQuery (query) {
   query = query.toLowerCase().replace(/ +/g, ' ').replace('lbry://', '');
   let splitBy = ['&', '$', ' '];
   let regex = new RegExp(splitBy.join('|'), 'gi');
-  let badWords  = [ 'from', 'with', 'not', 'can', 'all', 'are', 'for', 'but', 'and', 'the' ];
+  let badWords = ['from', 'with', 'not', 'can', 'all', 'are', 'for', 'but', 'and', 'the'];
   let words = query.split(regex);
   let sentence = [];
   words.forEach(w => {
-    if (!badWords.includes(w))      { sentence.push(w) }
+    if (!badWords.includes(w)) {
+      sentence.push(w);
+    }
   });
   query = sentence.join(' ');
 
@@ -502,12 +572,14 @@ async function update () {
 
 class LighthouseControllers {
   /* eslint-disable no-param-reassign */
+
   // Start syncing blocks...
   startSync () {
     winston.log('info', '[Importer] Started importer, indexing claims.');
     claimSync();
     // sync(); // Old Sync
   }
+
   /**
    * Search API Endpoint.
    * @param {ctx} Koa Context
@@ -515,6 +587,7 @@ class LighthouseControllers {
   async search (ctx) {
     await getResults(ctx.query).then(function (result) {
       let results = result.hits.hits;
+      console.log(results);
       let cResults = [];
       for (let pResult of results) {
         cResults.push(pResult._source);
@@ -560,6 +633,7 @@ class LighthouseControllers {
       ctx.body = clean;
     });
   }
+
   /**
    * Info about the api here
    * @param {ctx} Koa Context
@@ -597,15 +671,18 @@ class LighthouseControllers {
           ctx.body = 'OK';
         } else {
           ctx.status = 400;
-          ctx.body = 'skip auto update: pull request'; logToSlack(ctx.body);
+          ctx.body = 'skip auto update: pull request';
+          logToSlack(ctx.body);
         }
       } else {
         ctx.status = 400;
-        ctx.body = 'skip auto update: only deploys on master branch'; logToSlack(ctx.body);
+        ctx.body = 'skip auto update: only deploys on master branch';
+        logToSlack(ctx.body);
       }
     } else {
       ctx.status = 500;
-      ctx.body = 'skip auto update: could not verify webhook'; logToSlack(ctx.body);
+      ctx.body = 'skip auto update: could not verify webhook';
+      logToSlack(ctx.body);
     }
   }
 

server/utils/chainquery/index.js

@@ -19,7 +19,7 @@ import chainqueryConfig from '../../../chainquery-config.json';
 let connection = null;
 
 const esLogLevel = 'info';
-const MaxClaimsToProcessPerIteration = 100000;
+const MaxClaimsToProcessPerIteration = 100000000;
 const BatchSize = 5000;
 const loggerStream = winstonStream(winston, esLogLevel);
 const eclient = new elasticsearch.Client({
@@ -32,7 +32,7 @@ const eclient = new elasticsearch.Client({
   },
 });
 
-const queue = new ElasticQueue({elastic: eclient});
+const queue = new ElasticQueue({batchSize: 5000, concurrency: 6, elastic: eclient});
 queue.on('drain', function () {
   console.log('elasticsearch queue is drained');
 });
@@ -211,6 +211,7 @@ function getClaimsSince (time, lastID, MaxClaimsInCall) {
      p.claim_id as channel_id,
      c.bid_state,
      c.effective_amount,
+     c.transaction_time,
      COALESCE(p.effective_amount,1) as certificate_amount,
      c.claim_id as claimId,
      c.value_as_json as value
@@ -245,6 +246,7 @@ function getClaimsSince (time, lastID, MaxClaimsInCall) {
           bid_state         : r.bid_state,
           effective_amount  : r.effective_amount,
           certificate_amount: r.certificate_amount,
+          transaction_time  : new Date(r.transaction_time * 1000),
           claimId           : r.claimId,
           value             : value,
         });