From f8012f88641b0bb378ba79393f277c8918ebe452 Mon Sep 17 00:00:00 2001
From: Jules Laplace <julescarbon@gmail.com>
Date: Thu, 18 Apr 2019 16:50:01 +0200
Subject: better sorting of papers

---
 scraper/client/paper/paper.manager.js | 25 +++++++++++++++++--------
 scraper/client/types.js               |  6 ++++++
 2 files changed, 23 insertions(+), 8 deletions(-)

(limited to 'scraper')

diff --git a/scraper/client/paper/paper.manager.js b/scraper/client/paper/paper.manager.js
index 8b25b1cc..2c5b79de 100644
--- a/scraper/client/paper/paper.manager.js
+++ b/scraper/client/paper/paper.manager.js
@@ -3,7 +3,7 @@ import { bindActionCreators } from 'redux'
 import { connect } from 'react-redux'
 
 import * as actions from '../actions'
-import { USES_DATASET } from '../types'
+import { USES_DATASET, ADDRESS_TYPE_ENUM } from '../types'
 
 import { Loader } from '../common'
 
@@ -16,13 +16,10 @@ class PaperManager extends Component {
     if (this.props.match.params.key !== oldProps.match.params.key) {
       this.props.actions.getPaperInfo(this.props.match.params.key)
     }
-    console.log('whoms?')
     if (this.props.api.paperInfo.dataset !== oldProps.api.paperInfo.dataset && this.props.api.paperInfo.dataset && this.props.api.paperInfo.dataset.key) {
-    console.log('vert?')
       this.props.actions.getVerificationsDataset(this.props.api.paperInfo.dataset.key)
     }
     if (this.props.api.verifications !== oldProps.api.verifications && this.props.api.paperInfo.dataset) {
-    console.log('updated?')
       this.updateSortedCitations()
     }
   }
@@ -37,13 +34,25 @@ class PaperManager extends Component {
     }
     const citations = paperInfo.citations.concat(unknownCitations.citations)
     let verifiedLookup = verifications[dataset.key] || {}
-    const sortedCitations = citations.map(citation => [
+    // first sort by verification status,
+    // then by a combination of its geolocated sources and paper count.
+    // penalize papers with no PDF
+    const sortedCitations = citations
+      .map(citation => [
         citation.title,
-        verifiedLookup[citation.id] ? verifiedLookup[citation.id].uses_dataset : USES_DATASET.NO_DATA,
-        citation.pdf.length,
+        verifiedLookup[citation.id]
+          ? verifiedLookup[citation.id].uses_dataset
+          : USES_DATASET.NO_DATA,
+        (citation.addresses || [])
+          .map(address => (ADDRESS_TYPE_ENUM[address.type] || 0))
+          .reduce((a, b) => (a + b), 0)
+          + (citation.pdf.length > 0
+            ? Math.max(citation.pdf.length, 2)
+            : -2
+          ),
         citation
       ])
-      .sort((a,b) => (b[1] - a[1] || b[2] - a[2] || (a[0].localeCompare(b[0]))))
+      .sort((a, b) => (b[1] - a[1] || b[2] - a[2] || (a[0].localeCompare(b[0]))))
       .map(tup => ({
         ...tup[3],
         verified: tup[1],
diff --git a/scraper/client/types.js b/scraper/client/types.js
index 22c93a89..049b2761 100644
--- a/scraper/client/types.js
+++ b/scraper/client/types.js
@@ -22,3 +22,9 @@ export const USES_DATASET = {
   UNKNOWN: "0",
   NO_DATA: "-2",
 }
+
+export const ADDRESS_TYPE_ENUM = {
+  'edu': 1,
+  'company': 3,
+  'mil': 10,
+}
\ No newline at end of file
-- 
cgit v1.2.3-70-g09d2