From f1c61921f751126474a77259f52db85b23b1049f Mon Sep 17 00:00:00 2001
From: jazzzooo <38244149+jazzzooo@users.noreply.github.com>
Date: Wed, 13 Sep 2023 16:21:10 +0000
Subject: [PATCH] [fix] engine - Crossref

Crossref was broken on result types journal-issue and component .. The old code
had lots of assumptions, and broke during parsing.  Now the assumptions are more
explicit and checked them with the API.
---
 searx/engines/crossref.py | 82 ++++++++++++++++++++-------------------
 1 file changed, 43 insertions(+), 39 deletions(-)

diff --git a/searx/engines/crossref.py b/searx/engines/crossref.py
index e12a0da5b..c2ed7763f 100644
--- a/searx/engines/crossref.py
+++ b/searx/engines/crossref.py
@@ -1,60 +1,64 @@
 # SPDX-License-Identifier: AGPL-3.0-or-later
 # lint: pylint
-"""Semantic Scholar (Science)
-"""
-# pylint: disable=use-dict-literal
+"""CrossRef"""
 
 from urllib.parse import urlencode
-from searx.utils import html_to_text
+from datetime import datetime
 
 about = {
-    "website": 'https://www.crossref.org/',
-    "wikidata_id": 'Q5188229',
-    "official_api_documentation": 'https://github.com/CrossRef/rest-api-doc',
+    "website": "https://www.crossref.org/",
+    "wikidata_id": "Q5188229",
+    "official_api_documentation": "https://api.crossref.org",
     "use_official_api": False,
     "require_api_key": False,
-    "results": 'JSON',
+    "results": "JSON",
 }
 
-categories = ['science', 'scientific publications']
+categories = ["science", "scientific publications"]
 paging = True
-search_url = 'https://api.crossref.org/works'
+search_url = "https://api.crossref.org/works"
 
 
 def request(query, params):
-    params['url'] = search_url + '?' + urlencode(dict(query=query, offset=20 * (params['pageno'] - 1)))
+    params["url"] = search_url + "?" + urlencode({"query": query, "offset": 20 * (params["pageno"] - 1)})
     return params
 
 
 def response(resp):
-    res = resp.json()
     results = []
-    for record in res['message']['items']:
-        record_type = record['type']
-        if record_type == 'book-chapter':
-            title = record['container-title'][0]
-            if record['title'][0].lower().strip() != title.lower().strip():
-                title = html_to_text(title) + ' (' + html_to_text(record['title'][0]) + ')'
-            journal = None
+    for record in resp.json()["message"]["items"]:
+
+        if record["type"] == "component":
+            # These seem to be files published along with papers. Not something you'd search for
+            continue
+        result = {
+            "template": "paper.html",
+            "content": record.get("abstract", ""),
+            "doi": record.get("DOI"),
+            "pages": record.get("page"),
+            "publisher": record.get("publisher"),
+            "tags": record.get("subject"),
+            "type": record.get("type"),
+            "url": record.get("URL"),
+            "volume": record.get("volume"),
+        }
+        if record["type"] == "book-chapter":
+            result["title"] = record["container-title"][0]
+            if record["title"][0].lower().strip() != result["title"].lower().strip():
+                result["title"] += f" ({record['title'][0]})"
         else:
-            title = html_to_text(record['title'][0])
-            journal = record.get('container-title', [None])[0]
-        url = record.get('resource', {}).get('primary', {}).get('URL') or record['URL']
-        authors = [author.get('given', '') + ' ' + author.get('family', '') for author in record.get('author', [])]
-        isbn = record.get('isbn') or [i['value'] for i in record.get('isbn-type', [])]
-        results.append(
-            {
-                'template': 'paper.html',
-                'url': url,
-                'title': title,
-                'journal': journal,
-                'volume': record.get('volume'),
-                'type': record['type'],
-                'content': html_to_text(record.get('abstract', '')),
-                'publisher': record.get('publisher'),
-                'authors': authors,
-                'doi': record['DOI'],
-                'isbn': isbn,
-            }
-        )
+            result["title"] = record["title"][0] if "title" in record else record.get("container-title", [None])[0]
+            result["journal"] = record.get("container-title", [None])[0] if "title" in record else None
+
+        if "resource" in record and "primary" in record["resource"] and "URL" in record["resource"]["primary"]:
+            result["url"] = record["resource"]["primary"]["URL"]
+        if "published" in record and "date-parts" in record["published"]:
+            result["publishedDate"] = datetime(*(record["published"]["date-parts"][0] + [1, 1][:3]))
+        result["authors"] = [a.get("given", "") + " " + a.get("family", "") for a in record.get("author", [])]
+        result["isbn"] = record.get("isbn") or [i["value"] for i in record.get("isbn-type", [])]
+        # All the links are not PDFs, even if the URL ends with ".pdf"
+        # result["pdf_url"] = record.get("link", [{"URL": None}])[0]["URL"]
+
+        results.append(result)
+
     return results