[IMP] link_tracker, http_routing: ignore requests from social bots for link tracker

Purpose
=======
Avoid counting requests from social bots (twitter, facebook, linkedin...)
when tracking a link.

Specifications
=============
Social media platforms have a specific user agent in the HTTP headers that
can be used to detect them and to not increment the click count in that case.

Task-2578902

closes odoo/odoo#78806

Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>

addons/http_routing/models/ir_http.py

@@ -327,18 +327,6 @@ class IrHttp(models.AbstractModel):
         """
         return ['web']
 
-    bots = "bot|crawl|slurp|spider|curl|wget|facebookexternalhit".split("|")
-
-    @classmethod
-    def is_a_bot(cls):
-        # We don't use regexp and ustr voluntarily
-        # timeit has been done to check the optimum method
-        user_agent = request.httprequest.environ.get('HTTP_USER_AGENT', '').lower()
-        try:
-            return any(bot in user_agent for bot in cls.bots)
-        except UnicodeDecodeError:
-            return any(bot in user_agent.encode('ascii', 'ignore') for bot in cls.bots)
-
     @classmethod
     def _get_frontend_langs(cls):
         return [code for code, _ in request.env['res.lang'].get_installed()]
@@ -385,7 +373,7 @@ class IrHttp(models.AbstractModel):
         if request.routing_iteration == 1:
             context = dict(request.context)
             path = request.httprequest.path.split('/')
-            is_a_bot = cls.is_a_bot()
+            is_a_bot = request.env['ir.http'].is_a_bot()
 
             lang_codes = [code for code, *_ in Lang.get_available()]
             nearest_lang = not func and cls.get_nearest_lang(Lang._lang_get_code(path[1]))
@@ -462,7 +450,7 @@ class IrHttp(models.AbstractModel):
             path = request.httprequest.path.split('/')
             default_lg_id = cls._get_default_lang()
             if request.routing_iteration == 1:
-                is_a_bot = cls.is_a_bot()
+                is_a_bot = request.env['ir.http'].is_a_bot()
                 nearest_lang = not func and cls.get_nearest_lang(request.env['res.lang']._lang_get_code(path[1]))
                 url_lg = nearest_lang and path[1]
 

addons/link_tracker/controller/main.py

@@ -1,6 +1,8 @@
 # -*- coding: utf-8 -*-
 # Part of Odoo. See LICENSE file for full copyright and licensing details.
 
+from werkzeug.exceptions import NotFound
+
 from odoo import http
 from odoo.http import request
 
@@ -9,11 +11,15 @@ class LinkTracker(http.Controller):
 
     @http.route('/r/<string:code>', type='http', auth='public', website=True)
     def full_url_redirect(self, code, **post):
-        country_code = request.session.geoip and request.session.geoip.get('country_code') or False
-        request.env['link.tracker.click'].sudo().add_click(
-            code,
-            ip=request.httprequest.remote_addr,
-            country_code=country_code
-        )
+        if not request.env['ir.http'].is_a_bot():
+            country_code = request.session.geoip and request.session.geoip.get(
+                'country_code') or False
+            request.env['link.tracker.click'].sudo().add_click(
+                code,
+                ip=request.httprequest.remote_addr,
+                country_code=country_code,
+            )
         redirect_url = request.env['link.tracker'].get_url_from_code(code)
-        return request.redirect(redirect_url or '', code=301, local=False)
+        if not redirect_url:
+            raise NotFound()
+        return request.redirect(redirect_url, code=301, local=False)

addons/web/models/ir_http.py

@@ -16,6 +16,15 @@ from odoo.addons.web.controllers.main import HomeStaticTemplateHelpers
 class Http(models.AbstractModel):
     _inherit = 'ir.http'
 
+    bots = ["bot", "crawl", "slurp", "spider", "curl", "wget", "facebookexternalhit", "whatsapp", "trendsmapresolver", "pinterest", "instagram"]
+
+    @classmethod
+    def is_a_bot(cls):
+        user_agent = request.httprequest.user_agent.string.lower()
+        # We don't use regexp and ustr voluntarily
+        # timeit has been done to check the optimum method
+        return any(bot in user_agent for bot in cls.bots)
+
     def webclient_rendering_context(self):
         return {
             'menu_data': request.env['ir.ui.menu'].load_menus(request.session.debug),