From d6dfada1a9c77cae13596c441f7d1cb1c2b30899 Mon Sep 17 00:00:00 2001
From: ta <alt3753.7@gmail.com>
Date: Sat, 13 Aug 2022 06:04:31 +0700
Subject: [PATCH 1/2] add neeva web engine

Neeva is "the world's first ad-free, private search engine" and uses data from Apple, Bing, Yelp and "others".
They claim to crawl "hundreds of millions" of URLs a day (https://twitter.com/Neeva/status/1536447373903335426).
---
 searx/settings.yml | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/searx/settings.yml b/searx/settings.yml
index ec03819a..52b75082 100644
--- a/searx/settings.yml
+++ b/searx/settings.yml
@@ -960,6 +960,30 @@ engines:
   #   collection: 'reviews'  # name of the db collection
   #   key: 'name'  # key in the collection to search for
 
+  - name: neeva
+    engine: xpath
+    shortcut: nv
+    search_url: https://neeva.com/search?q={query}&c=All&src=Pagination&page={pageno}{time_range}
+    results_xpath: //div[@class="web-index__component-2rKiM"] | //li[@class="web-rich-deep-links__deepLink-SIbD4"]
+    url_xpath: .//a[@class="lib-doc-title__link-1b9rC"]/@href | ./h2/a/@href
+    title_xpath: .//a[@class="lib-doc-title__link-1b9rC"] | ./h2/a
+    content_xpath: >
+      .//div[@class="lib-doc-snippet__component-3ewW6"]/text() |
+      .//div[@class="lib-doc-snippet__component-3ewW6"]/*[not(self::a)] |
+      ./p
+    content_html_to_text: true
+    suggestion_xpath: //span[@class="result-related-searches__link-2ho_u"]
+    paging: true
+    disabled: true
+    categories: [general, web]
+    timeout: 5.0
+    about:
+      website: https://neeva.com
+      official_api_documentation:
+      use_official_api: false
+      require_api_key: false
+      results: HTML
+
   - name: npm
     engine: json_engine
     paging: true

From 8fa84ee9f4b2946cc55b8934832923b344d70180 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?L=C3=A9on=20Tiek=C3=B6tter?= <leon@tiekoetter.com>
Date: Sat, 13 Aug 2022 01:59:58 +0200
Subject: [PATCH 2/2] Fix time range support and add soft_max_redirects

Add custom time_range_url and time_range_map

Set soft_max_redirects = 2 to prevent "ErrorContext('searx/search/processors/online.py', 116, 'count_error(', None, '2 redirects, maximum: 0', ('200', 'OK', 'neeva.com')) True"
---
 searx/settings.yml | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/searx/settings.yml b/searx/settings.yml
index 52b75082..6cae30c4 100644
--- a/searx/settings.yml
+++ b/searx/settings.yml
@@ -963,6 +963,13 @@ engines:
   - name: neeva
     engine: xpath
     shortcut: nv
+    time_range_support: true
+    time_range_url: '&alf%5Bfreshness%5D={time_range_val}'
+    time_range_map:
+      day: 'Day'
+      week: 'Week'
+      month: 'Month'
+      year: 'Year'
     search_url: https://neeva.com/search?q={query}&c=All&src=Pagination&page={pageno}{time_range}
     results_xpath: //div[@class="web-index__component-2rKiM"] | //li[@class="web-rich-deep-links__deepLink-SIbD4"]
     url_xpath: .//a[@class="lib-doc-title__link-1b9rC"]/@href | ./h2/a/@href
@@ -977,6 +984,7 @@ engines:
     disabled: true
     categories: [general, web]
     timeout: 5.0
+    soft_max_redirects: 2
     about:
       website: https://neeva.com
       official_api_documentation: