From b6beb5f02e8c291f2b5f2bc7d0c15d0955336c32 Mon Sep 17 00:00:00 2001
From: Jan Tuomi <jans.tuomi@gmail.com>
Date: Fri, 2 Jul 2021 11:19:25 +0300
Subject: Implement research papers

---
 backend/index.ts | 34 +++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

(limited to 'backend/index.ts')

diff --git a/backend/index.ts b/backend/index.ts
index 9efec62..9849e50 100644
--- a/backend/index.ts
+++ b/backend/index.ts
@@ -7,12 +7,21 @@ import {
     DatabaseTransactionConnectionType,
     sql,
 } from "slonik";
+import { parse } from "node-html-parser";
+import got from "got";
+import { decode } from "he";
+
 import * as config from "./config";
-import { ValueRecord } from "./shared-types";
+import { Paper, ValueRecord } from "./shared-types";
 import { Server as SocketIOServer } from "socket.io";
 
 const app = express();
 
+app.use((req, res, next) => {
+    res.setHeader("Access-Control-Allow-Origin", "*");
+    next();
+});
+
 const httpServer = http.createServer(app);
 const io = new SocketIOServer(httpServer, {
     cors: {
@@ -53,6 +62,29 @@ io.on("connection", async (socket) => {
     });
 });
 
+const extractDataFromHtml = (html: string): Paper[] => {
+    const dom = parse(html);
+
+    const tds = dom.querySelectorAll(".gsc_a_tr .gsc_a_t a");
+    const data = tds.map(td => ({
+        url: `https://scholar.google.com${td.getAttribute("data-href") || ""}`,
+        title: decode(td.textContent),
+    }));
+
+    return data;
+};
+
+app.get("/api/papers", async (req, res) => {
+    const resp = await got("https://scholar.google.com/citations?user=HMp8VyMAAAAJ", {
+        responseType: "text",
+        encoding: "latin1",
+    });
+    const html = resp.body;
+    const papers = extractDataFromHtml(html);
+    res.contentType("json");
+    res.end(JSON.stringify(papers));
+});
+
 httpServer.listen(config.port, () => {
     console.log(`Listening on http://localhost:${config.port}`);
 });
-- 
cgit v1.3