Add a timeout to a single page retrieval

To be tested when history generation is available

.gitignore

@@ -65,3 +65,4 @@ venv/
 # Django stuff
 db.sqlite3
 
+_vimrc_local.vim

crawl/crawl.py

@@ -1,11 +1,10 @@
 from threading import Thread
-from queue import Queue
 from urllib.robotparser import RobotFileParser
 from urllib.error import URLError
 from urllib.parse import urlparse
 
 from ssl import CertificateError
-from random import sample, randrange
+from random import sample, randrange, randint
 import re
 from datetime import datetime, timedelta
 
@@ -15,6 +14,8 @@ import async_timeout
 
 from bs4 import BeautifulSoup, Comment
 
+from profiles.models import BrowserFingerprint, SearchEngine
+
 # Ugly hack to use this module alone instead of integrating it with Django
 # from django.conf import settings
 
@@ -26,11 +27,11 @@ MAX_PER_PAGE = 10
 
 FOOTER_URL = re.compile(".*footer.*")
 
+
 class Settings:
     USER_AGENT = 'Default User'
 
 settings = Settings()
-startup_time = datetime.min
 
 
 def url_getter(html, current_page, root_url):
@@ -72,7 +73,7 @@ def url_getter(html, current_page, root_url):
     # Works only with python >= 3.6
     links_list = list(dict.fromkeys(links_list))
 
-    forbidden_words = ['login', 'agreement', 'mailto']
+    forbidden_words = ['login', 'agreement', 'mailto', 'settings']
     links_list = [link for link in links_list if not any(word in link.lower()
                                                          for word in
                                                          forbidden_words)]
@@ -80,8 +81,6 @@ def url_getter(html, current_page, root_url):
     return links_list
 
 
-
-
 class WebsiteSchedulerMeta(type):
     """ Meta-class for WebsiteScheduler, allowing a singleton class-like
     interface, but spawning one instance per canonical website URL """
@@ -104,36 +103,47 @@ class WebsiteSchedulerMeta(type):
 
 class WebsiteScheduler(metaclass=WebsiteSchedulerMeta):
     """ Schedule the accesses to a website as of robots.txt """
-    def __init__(self, name):
+
+    search_engines = []  # Must be set by CrawlingThread.__init__
+
+    def __init__(self, name, user_agent):
         self.name = name
         self.last_crawled = datetime.fromtimestamp(0)
         self.dead = False
-        try:
-            robots_url = self.urlroot() + 'robots.txt'
-            self.robot_parser = RobotFileParser(robots_url)
-            self.robot_parser.read()  # TODO async?
-        except (URLError, CertificateError):
+        self.can_fetch_b = False
+        self.user_agent = (user_agent if user_agent is not None
+                           else settings.USER_AGENT)
+        if any(self.urlroot() in item for item in self.search_engines):
+            print("found a search engine for %s" % self.urlroot())
+            self.crawl_delay = timedelta(seconds=5)
+            self.can_fetch_b = True
+        else:
             try:
-                robots_url = self.unsafe_urlroot() + 'robots.txt'
+                robots_url = self.urlroot() + 'robots.txt'
                 self.robot_parser = RobotFileParser(robots_url)
-                self.robot_parser.read()
-            except URLError: # Almost surely an offline website.
+                self.robot_parser.read()  # TODO async?
+            except (URLError, CertificateError):
+                try:
+                    robots_url = self.unsafe_urlroot() + 'robots.txt'
+                    self.robot_parser = RobotFileParser(robots_url)
+                    self.robot_parser.read()
+                except URLError:  # Almost surely an offline website.
+                    self.dead = True
+                    self.crawl_delay = 0
+            except Exception as e:
+                print(e)
+                raise e
+            if not self.robot_parser.default_entry:
                 self.dead = True
-                self.crawl_delay = 0
-        except Exception as e:
-            print(e)
-            raise e
-        if not self.robot_parser.default_entry:
-            self.dead = True
-        if not self.dead:
-            delay = self.robot_parser.crawl_delay(settings.USER_AGENT)
-            if delay is None:
-                req_rate = self.robot_parser.request_rate(settings.USER_AGENT)
-                if req_rate is None:
-                    delay = 5
-                else:
-                    delay = req_rate.requests, req_rate.seconds
-            self.crawl_delay = timedelta(seconds=delay)
+            if not self.dead:
+                delay = self.robot_parser.crawl_delay(self.user_agent)
+                if delay is None:
+                    req_rate = self.robot_parser.request_rate(self.user_agent)
+                    if req_rate is None:
+                        delay = 5
+                    else:
+                        delay = req_rate.requests, req_rate.seconds
+                self.crawl_delay = timedelta(seconds=delay)
 
     def urlroot(self):
         ''' Get the root url for this website '''
@@ -151,7 +161,9 @@ class WebsiteScheduler(metaclass=WebsiteSchedulerMeta):
 
     def can_fetch(self, url):
         ''' Check whether this program can fetch a given page '''
-        return (not self.dead) and self.robot_parser.can_fetch(settings.USER_AGENT, url)
+        return ((self.can_fetch_b)
+                or ((not self.dead) and
+                    self.robot_parser.can_fetch(self.user_agent, url)))
 
     def fetching(self):
         ''' Tell the scheduler that a page is being fetched now '''
@@ -162,38 +174,47 @@ class CrawlingThread(Thread):
     """ A separate thread for the crawling task. This is needed to use asyncio,
     since the thread will need its own event loop. """
 
-    def __init__(self, user, url, queue):
-        global settings
-        self.queue = queue
+    def __init__(self, url):
+        engine_list = [engine.url for engine in SearchEngine.objects.all()]
+        WebsiteScheduler.search_engines = engine_list
+
+        nb_fingerprint = len(BrowserFingerprint.objects.all())
+        fingerprint = BrowserFingerprint.objects.all()[
+            randint(0, nb_fingerprint - 1)]
+        self.headers = fingerprint.serialize_headers()
+
+        self.output_tree = []
         super(CrawlingThread, self).__init__()
-        if user:
-            settings.USER_AGENT = user.serialize_headers()
         self.url = url
 
     def run(self):
-        global startup_time
         tasks = []
+
         #tasks.append(async_crawler("http://plus.google.com/+Python"))
         #tasks.append(async_crawler('https://python.org/'))
-        tasks.append(async_crawler(self.url, self.queue))
+        tasks.append(run_crawl(self.url, self.output_tree, self.headers))
 
-        loop = asyncio.new_event_loop()
-        asyncio.set_event_loop(loop)
-        startup_time = datetime.now()
-        loop.run_until_complete(asyncio.wait(tasks))
-        loop.close()
+        try:
+            loop = asyncio.new_event_loop()
+            asyncio.set_event_loop(loop)
+            loop.run_until_complete(asyncio.wait(tasks))
+        finally:
+            loop.close()
 
 
 class PageGetter:
     """ Asynchronously get a webpage, abiding by robots.txt """
 
-    def __init__(self, session, url):
+    headers = None
+
+    def __init__(self, session, url, user_agent):
         self.url = url
         self.session = session
+        self.user_agent = user_agent
 
     async def get(self, ssl=True):
         """ Actually retrieve the webpage """
-        scheduler = WebsiteScheduler(self.url)
+        scheduler = WebsiteScheduler(self.url, self.user_agent)
         if not scheduler.can_fetch(self.url):
             return None
 
@@ -203,7 +224,8 @@ class PageGetter:
             delay = scheduler.fetch_delay()
         scheduler.fetching()
         async with async_timeout.timeout(10):
-            async with self.session.get(self.url, ssl=ssl) as resp:
+            async with self.session.get(self.url, verify_ssl=ssl) as resp:
+                print("Resp status %s" % resp.status)
                 try:
                     return await resp.text()
                 except UnicodeDecodeError:
@@ -213,46 +235,89 @@ class PageGetter:
 async def async_print(url):
     """ Debug function to follow what's actually happening """
     async with aiohttp.ClientSession() as session:
-        html = await PageGetter(session, url).get(ssl=False)
+        html = await PageGetter(session, url,
+                                settings.USER_AGENT).get(ssl=False)
 
-        print('GOT {}HTML for {} at {}'.format(
+        print('GOT {}HTML for {}'.format(
             'None ' if html is None else '',
             url,
-            datetime.now() - startup_time))
+        ))
 
-async def async_crawler(url, queue):
-    queued = [url]
-    crawled = []
-    while queued and (len(crawled) < HARD_LIMIT):
-        async with aiohttp.ClientSession() as session:
-            try:
-                url = queued.pop(0)
-            except IndexError:
-                print("Error queue is empty")
-                return crawled
-            parsed_url = urlparse(url)
-            print("Crawling {}".format(url))
-            html = await PageGetter(session, url).get(ssl=False)
-            if html:
-                new_urls = url_getter(
-                    html,
-                    url,
-                    parsed_url.scheme + "://" + parsed_url.netloc
-                )
-                crawled += [url]
-                if new_urls:
-                    sampled = sample(
-                        new_urls,
-                        randrange(min(MAX_PER_PAGE, len(new_urls)))
-                    )
-                    queued += [sample_url for sample_url in sampled if
-                              sample_url not in queued and sample_url not in
-                              crawled]
-    print(crawled)
-    queue.put(crawled)
 
-if __name__ == '__main__':
-    queue = Queue()
-    crawl = CrawlingThread(None, "https://python.org/", queue)
-    crawl.start()
-    crawl.join()
+class CrawlElem:
+    ''' Describes a crawled element, to be assembled into a tree '''
+
+    def __init__(self, url, parent):
+        self.url = url
+        self.parent = parent
+
+
+async def run_crawl(url, output_tree, headers=None):
+    ''' Starts a crawling session '''
+
+    if headers is None:
+        headers = {}
+    if 'User-Agent' not in headers:
+        headers['User-Agent'] = settings.USER_AGENT
+
+    user_agent = headers['User-Agent']
+    crawled = set()
+
+    async with aiohttp.ClientSession(headers=headers) as session:
+        await async_crawler(
+            url, output_tree, crawled, user_agent, session, None)
+
+
+def simplify_url(url):
+    anchor = url.find('#')
+    if anchor >= 0:
+        url = url[:anchor]
+
+    prot = url.find('://')
+    if prot >= 0:
+        url = url[prot+3:]
+
+    if url.startswith('www.'):
+        url = url[4:]
+
+    return url
+
+
+async def async_crawler(url, out_tree, crawled, user_agent, session, parent):
+    if len(crawled) >= HARD_LIMIT:
+        return
+    crawled.add(simplify_url(url))
+    parsed_url = urlparse(url)
+    print("Crawling {}".format(url))
+    try:
+        with async_timeout.timeout(3):
+            html = await PageGetter(session, url, user_agent).get(ssl=False)
+    except asyncio.TimeoutError:
+        return
+
+    new_tasks = []
+
+    if html:
+        this_elem = CrawlElem(url, parent)
+        out_tree.append(this_elem)
+        new_urls = url_getter(
+            html,
+            url,
+            parsed_url.scheme + "://" + parsed_url.netloc
+        )
+        if new_urls:
+            sampled = sample(
+                new_urls,
+                randrange(min(MAX_PER_PAGE, len(new_urls)))
+            )
+            for sample_url in sampled:
+                if simplify_url(sample_url) not in crawled:
+                    new_tasks.append(async_crawler(
+                        sample_url, out_tree, crawled, user_agent, session,
+                        this_elem))
+    else:
+        print("No html received")
+    if len(crawled) >= HARD_LIMIT:
+        return
+    if new_tasks:
+        await asyncio.wait(new_tasks)

data/.gitignore

@@ -0,0 +1 @@
+nicknames_dict

data/email_domains.txt

@@ -0,0 +1,14 @@
+gmail.com
+protonmail.com
+riseup.net
+tutanoto.com
+outlook.fr
+fastmail.com
+yandex.com
+aim.com
+icloud.com
+yahoo.com
+fmx.fr
+mail.com
+hushmail.com
+inbox.com

data/events.json

@@ -0,0 +1,27 @@
+[
+    {
+        "name" : "Atelier Anti-Pub",
+        "date" : "07/03/2018 19:00 UTC",
+        "place" : {
+            "name" : "Centre Social Autogéré Vaydom",
+            "address" : "37 rue Marceau, Ivry-sur-Seine",
+            "lat" : "48.81787",
+            "lon" : "2.38032"
+        }
+    },
+    {
+        "name" : "Rassemblement de soutien pour Bure",
+        "date" : "27/02/2018 17:00 UTC",
+        "place" : {
+            "name" : "Place Saint-Michel",
+            "address" : "Place Saint-Michel, 75005 Paris",
+            "lat" : "48.85374",
+            "lon" : "2.34455"
+        }
+    },
+    {
+        "name" : "Création d'un serveur mail",
+        "date" : "15/02/2018 12:00 UTC",
+        "place" : "La Mutinerie"
+    }
+]

data/firstnames.txt

@@ -0,0 +1,200 @@
+Jean
+Marie
+Philippe
+Nathalie
+Michel
+Isabelle
+Alain
+Sylvie
+Patrick
+Catherine
+Nicolas
+Martine
+Christophe
+Christine
+Pierre
+Françoise
+Christian
+Valérie
+Éric
+Sandrine
+Frédéric
+Stéphanie
+Laurent
+Véronique
+Stéphane
+Sophie
+David
+Céline
+Pascal
+Chantal
+Daniel
+Patricia
+Alexandre
+Anne
+Julien
+Brigitte
+Thierry
+Julie
+Olivier
+Monique
+Bernard
+Aurélie
+Thomas
+Nicole
+Sébastien
+Laurence
+Gérard
+Annie
+Didier
+Émilie
+Dominique
+Dominique
+Vincent
+Virginie
+François
+Corinne
+Bruno
+Élodie
+Guillaume
+Christelle
+Jérôme
+Camille
+Jacques
+Caroline
+Marc
+Léa
+Maxime
+Sarah
+Romain
+Florence
+Claude
+Laetitia
+Antoine
+Audrey
+Franck
+Hélène
+Jean-Pierre
+Laura
+Anthony
+Manon
+Kévin
+Michèle
+Gilles
+Cécile
+Cédric
+Christiane
+Serge
+Béatrice
+André
+Claire
+Mathieu
+Nadine
+Benjamin
+Delphine
+Patrice
+Pauline
+Fabrice
+Karine
+Joël
+Mélanie
+Jérémy
+Marion
+Clément
+Chloe
+Arnaud
+Jacqueline
+Denis
+Elisabeth
+Paul
+Evelyne
+Lucas
+Marine
+Hervé
+Claudine
+Jean-Claude
+Anais
+Sylvain
+Lucie
+Yves
+Danielle
+Ludovic
+Carole
+Guy
+Fabienne
+Florian
+Mathilde
+Damien
+Sandra
+Alexis
+Pascale
+Mickaël
+Annick
+Quentin
+Charlotte
+Emmanuel
+Emma
+Louis
+Severine
+Benoît
+Sabrina
+Jean-Luc
+Amandine
+Fabien
+Myriam
+Francis
+Jocelyne
+Hugo
+Alexandra
+Jonathan
+Angelique
+Loïc
+Josiane
+Xavier
+Joelle
+Théo
+Agnes
+Adrien
+Mireille
+Raphaël
+Vanessa
+Jean-François
+Justine
+Grégory
+Sonia
+Robert
+Bernadette
+Michaël
+Emmanuelle
+Valentin
+Oceane
+Cyril
+Amelie
+Jean-Marc
+Clara
+René
+Maryse
+Lionel
+Anne-marie
+Yannick
+Fanny
+Enzo
+Magali
+Yannis
+Marie-christine
+Jean-Michel
+Morgane
+Baptiste
+Ines
+Matthieu
+Nadia
+Rémi
+Muriel
+Georges
+Jessica
+Aurélien
+Laure
+Nathan
+Genevieve
+Jean-Paul
+Estelle

data/interests.json

@@ -0,0 +1,55 @@
+[
+    {
+        "name": "occupation",
+        "keywords": [
+            {"keyword" : "ZAD NDDL"},
+            {"keyword" : "Organiser un squat"},
+            {"keyword" : "mobilisation et rassemblement"}
+        ],
+        "places": [
+            {"place" : "Zad NDDL"},
+            {"place" : "Zad Bure"}
+        ],
+        "websites": [
+            {"website": "zad nadir"}
+        ],
+        "events":  [
+            {"event": "Rassemblement de soutien pour Bure"}
+        ]
+    },
+    {
+        "name": "LGBT",
+        "keywords": [
+            {"keyword" : "Discrimniation sexistes, quelles actions ?"},
+            {"keyword" : "gender queer Paris"},
+            {"keyword" : "Existrans Paris"}
+        ],
+        "places": [
+            {"place" : "La Mutinerie"}
+        ],
+        "websites": [
+            {"website": "emmaclit"},
+            {"website": "paris-luttes info"}
+        ],
+        "events":  [
+            {"event": "Création d'un serveur mail"}
+        ]
+    },
+    {
+        "name": "Anti pub",
+        "keywords": [
+            {"keyword" : "Affichage JCDecaux"},
+            {"keyword" : "Anti-pub"},
+            {"keyword" : "Journée contre la publicité"}
+        ],
+        "places": [
+            {"place" : "Centre Social Autogéré Vaydom"}
+        ],
+        "websites": [
+            {"website": "paris-luttes info"}
+        ],
+        "events":  [
+            {"event": "Atelier Anti-Pub"}
+        ]
+    }
+]

data/lastnames.txt

@@ -0,0 +1,200 @@
+Martin
+Bernard
+Thomas
+Petit
+Robert
+Richard
+Durand
+Dubois
+Moreau
+Laurent
+Simon
+Michel
+Lefebvre
+Leroy
+Roux
+David
+Bertrand
+Morel
+Fournier
+Girard
+Bonnet
+Dupont
+Lambert
+Fontaine
+Rousseau
+Vincent
+Muller
+Lefevre
+Faure
+Andre
+Mercier
+Blanc
+Guerin
+Boyer
+Garnier
+Chevalier
+Francois
+Legrand
+Gauthier
+Garcia
+Perrin
+Robin
+Clement
+Morin
+Nicolas
+Henry
+Roussel
+Mathieu
+Gautier
+Masson
+Marchand
+Duval
+Denis
+Dumont
+Marie
+Lemaire
+Noel
+Meyer
+Dufour
+Meunier
+Brun
+Blanchard
+Giraud
+Joly
+Riviere
+Lucas
+Brunet
+Gaillard
+Barbier
+Arnaud
+Martinez
+Gerard
+Roche
+Renard
+Schmitt
+Roy
+Leroux
+Colin
+Vidal
+Caron
+Picard
+Roger
+Fabre
+Aubert
+Lemoine
+Renaud
+Dumas
+Lacroix
+Olivier
+Philippe
+Bourgeois
+Pierre
+Benoit
+Rey
+Leclerc
+Payet
+Rolland
+Leclercq
+Guillaume
+Lecomte
+Lopez
+Jean
+Dupuy
+Guillot
+Hubert
+Berger
+Carpentier
+Sanchez
+Dupuis
+Moulin
+Louis
+Deschamps
+Huet
+Vasseur
+Perez
+Boucher
+Fleury
+Royer
+Klein
+Jacquet
+Adam
+Paris
+Poirier
+Marty
+Aubry
+Guyot
+Carre
+Charles
+Renault
+Charpentier
+Menard
+Maillard
+Baron
+Bertin
+Bailly
+Herve
+Schneider
+Fernandez
+Le Gall
+Collet
+Leger
+Bouvier
+Julien
+Prevost
+Millet
+Perrot
+Daniel
+Le Roux
+Cousin
+Germain
+Breton
+Besson
+Langlois
+Remy
+Le Goff
+Pelletier
+Leveque
+Perrier
+Leblanc
+Barre
+Lebrun
+Marchal
+Weber
+Mallet
+Hamon
+Boulanger
+Jacob
+Monnier
+Michaud
+Rodriguez
+Guichard
+Gillet
+Etienne
+Grondin
+Poulain
+Tessier
+Chevallier
+Collin
+Chauvin
+Da Silva
+Bouchet
+Gay
+Lemaitre
+Benard
+Marechal
+Humbert
+Reynaud
+Antoine
+Hoarau
+Perret
+Barthelemy
+Cordier
+Pichon
+Lejeune
+Gilbert
+Lamy
+Delaunay
+Pasquier
+Carlier
+Laporte

data/website.json

@@ -1,93 +1,93 @@
 [
-    "website": {
+    {
         "name":"emmaclit",
         "url":"https://emmaclit.com/",
         "keywords": [
-            "keyword":"Charge mentale",
-            "keyword":"Un autre regard",
-            "keyword":"Un petit poutou",
-            "keyword":"solidarité",
-            "keyword":"dédicace"
-        },
-        "notable_pages": [
-            "webpage": "https://emmaclit.com/2017/05/09/repartition-des-taches-hommes-femmes/",
-            "webpage": "https://emmaclit.com/2016/12/01/une-famille-parmi-dautres/",
-            "webpage": "https://emmaclit.com/2017/09/11/travaille-pourquoi/"
-        ]
-    },
-    "website": {
-        "name":"paris-luttes info",
-        "url":"https//paris-luttes.info/",
-        "keywords": [
-            "keyword":"manifestations",
-            "keyword":"solidarité immigré·e·s",
-            "keyword":"grève salariés",
-            "keyword":"prison",
-            "keyword":"violence policère"
+            {"keyword":"Charge mentale"},
+            {"keyword":"Un autre regard"},
+            {"keyword":"Un petit poutou"},
+            {"keyword":"solidarité"},
+            {"keyword":"dédicace"}
         ],
         "notable_pages": [
-            "webpage": "https://paris-luttes.info/-analyse-et-reflexion-?lang=fr",
-            "webpage": "https://paris-luttes.info/comment-publier-sur-paris-luttes-134?lang=fr",
-            "webpage": "https://paris-luttes.info/pourquoi-et-comment-utiliser-tor-9013?lang=fr"
+            {"webpage": "https://emmaclit.com/2017/05/09/repartition-des-taches-hommes-femmes/"},
+            {"webpage": "https://emmaclit.com/2016/12/01/une-famille-parmi-dautres/"},
+            {"webpage": "https://emmaclit.com/2017/09/11/travaille-pourquoi/"}
         ]
     },
-    "website": {
+    {
+        "name":"paris-luttes info",
+        "url":"https://paris-luttes.info/",
+        "keywords": [
+            {"keyword":"manifestations"},
+            {"keyword":"solidarité immigré·e·s"},
+            {"keyword":"grève salariés"},
+            {"keyword":"prison"},
+            {"keyword":"violence policère"}
+        ],
+        "notable_pages": [
+            {"webpage": "https://paris-luttes.info/-analyse-et-reflexion-?lang=fr"},
+            {"webpage": "https://paris-luttes.info/comment-publier-sur-paris-luttes-134?lang=fr"},
+            {"webpage": "https://paris-luttes.info/pourquoi-et-comment-utiliser-tor-9013?lang=fr"}
+        ]
+    },
+    {
         "name":"zad nadir",
         "url":"http://zad.nadir.org/",
         "keywords": [
-            "keyword":"Écologie",
-            "keyword":"opération césar",
-            "keyword":"expulsion vinci",
-            "keyword":"adresse",
-            "keyword":"la wardine",
-            "keyword":"route des chicanes",
-            "keyword":"opposition à l'aéroport Grand Ouest"
+            {"keyword":"Écologie"},
+            {"keyword":"opération césar"},
+            {"keyword":"expulsion vinci"},
+            {"keyword":"adresse"},
+            {"keyword":"la wardine"},
+            {"keyword":"route des chicanes"},
+            {"keyword":"opposition à l'aéroport Grand Ouest"}
         ],
         "notable_pages": [
-            "webpage": "http://zad.nadir.org/spip.php?article86&lang=fr",
-            "webpage": "http://zad.nadir.org/spip.php?article515&lang=fr",
-            "webpage": "http://zad.nadir.org/spip.php?rubrique71",
-            "webpage": "https://zad.nadir.org/spip.php?rubrique70"
+            {"webpage": "http://zad.nadir.org/spip.php?article86&lang=fr"},
+            {"webpage": "http://zad.nadir.org/spip.php?article515&lang=fr"},
+            {"webpage": "http://zad.nadir.org/spip.php?rubrique71"},
+            {"webpage": "https://zad.nadir.org/spip.php?rubrique70"}
         ]
     },
-    "website": {
+    {
         "name":"Fnac",
         "url":"https://www.fnac.com/",
         "keywords": [
-            "keyword":"smartphone",
-            "keyword":"SAV",
-            "keyword":"Macbook",
-            "keyword":"TV",
-            "keyword":"PC Gaming",
-            "keyword":"DVD",
-            "keyword":"Home Cinema Philips",
-            "keyword":"Billeterie"
+            {"keyword":"smartphone"},
+            {"keyword":"SAV"},
+            {"keyword":"Macbook"},
+            {"keyword":"TV"},
+            {"keyword":"PC Gaming"},
+            {"keyword":"DVD"},
+            {"keyword":"Home Cinema Philips"},
+            {"keyword":"Billeterie"}
         ],
         "notable_pages": [
-            "webpage": "https://www.fnac.com/Informatique/shi48966/w-4#bl=MMinfo",
-            "webpage": "https://www.fnac.com/Service/default.aspx#bl=footer",
-            "webpage": "https://www.fnac.com/Ventes-Flash/shi42245/w-4#bl=marktlink1",
-            "webpage": "https://www.fnac.com/Home-cinema-barre-de-son-et-enceinte-TV/shi474916/w-4#bl=MMtvh"
+            {"webpage": "https://www.fnac.com/Informatique/shi48966/w-4#bl=MMinfo"},
+            {"webpage": "https://www.fnac.com/Service/default.aspx#bl=footer"},
+            {"webpage": "https://www.fnac.com/Ventes-Flash/shi42245/w-4#bl=marktlink1"},
+            {"webpage": "https://www.fnac.com/Home-cinema-barre-de-son-et-enceinte-TV/shi474916/w-4#bl=MMtvh"}
         ]
     },
-    "website": {
+    {
         "name":"Sea Shepherd",
         "url":"https://www.seashepherd.fr/",
         "keywords": [
-            "keyword":"pirates",
-            "keyword":"Phoques",
-            "keyword":"Paul Watson",
-            "keyword":"harponnage",
-            "keyword":"seal",
-            "keyword":"Chasse aux dauphins",
-            "keyword":"participation",
-            "keyword":"boutique"
+            {"keyword":"pirates"},
+            {"keyword":"Phoques"},
+            {"keyword":"Paul Watson"},
+            {"keyword":"harponnage"},
+            {"keyword":"seal"},
+            {"keyword":"Chasse aux dauphins"},
+            {"keyword":"participation"},
+            {"keyword":"boutique"}
         ],
         "notable_pages": [
-            "webpage": "http://www.seashepherd.fr/index.php/qui-sommes-nous",
-            "webpage": "http://nyamba.seashepherd.info/",
-            "webpage": "http://seashepherd-shop.com/en/",
-            "webpage": "http://seashepherd.fr/index.php/qui-sommes-nous/sea-shepherd-france"
+            {"webpage": "http://www.seashepherd.fr/index.php/qui-sommes-nous"},
+            {"webpage": "http://nyamba.seashepherd.info/"},
+            {"webpage": "http://seashepherd-shop.com/en/"},
+            {"webpage": "http://seashepherd.fr/index.php/qui-sommes-nous/sea-shepherd-france"}
         ]
     }
 ]

histories/migrations/0001_initial.py

@@ -0,0 +1,34 @@
+# Generated by Django 2.0.1 on 2018-02-25 19:08
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    initial = True
+
+    dependencies = [
+        ('profiles', '0001_initial'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='History',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('start_ts', models.DateTimeField(help_text='The starting timestamp of the history. Useful for cron-like structure.')),
+                ('played', models.BooleanField(default=False)),
+                ('user', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, to='profiles.Profile')),
+            ],
+        ),
+        migrations.CreateModel(
+            name='HistoryEntry',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('search', models.URLField(help_text='The url to be searched')),
+                ('timestamp', models.DateTimeField()),
+                ('history', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, to='histories.History')),
+            ],
+        ),
+    ]

histories/models.py

@@ -3,13 +3,26 @@ entries, which looks like human-based browsing, according to a dedicated user
 interests, keywords...
 """
 
+from collections import namedtuple
 import random
 from math import floor
-from queue import Queue
+from xml.etree import ElementTree as ET
+from datetime import datetime
 from django.db import models
+from django.core.exceptions import ValidationError
 import profiles.models as profiles
 from crawl import crawl
 from pinocchio.settings import HISTORY_MIN
+from .tor_runner import TorInstance
+
+
+class InvalidXml(Exception):
+    def __init__(self, what='unexpected XML data.'):
+        super().__init__()
+        self.what = what
+
+    def __str__(self):
+        return "Invalid XML: " + self.what
 
 
 class HistoryEntry(models.Model):
@@ -27,14 +40,48 @@ class HistoryEntry(models.Model):
         """
         return "{} : {}".format(self.timestamp, self.search)
 
+    def to_xml(self, xml_root):
+        entry = ET.Element('history')
+        entry_url = ET.Element('url')
+        entry_url.text = self.search
+        entry_ts = ET.Element('timestamp')
+        entry_ts.text = self.timestamp.timestamp()
+        entry.append(entry_url)
+        entry.append(entry_ts)
+        xml_root.append(entry)
+
+    @staticmethod
+    def from_xml(xml_root, in_history):
+        if xml_root.tag != 'history':
+            raise InvalidXml("expected <history> tag here.")
+        url, timestamp = None, None
+
+        for child in xml_root:
+            if child.tag == 'url':
+                url = child.text
+            elif child.tag == 'timestamp':
+                try:
+                    timestamp = datetime.fromtimestamp(child.text)
+                except TypeError:
+                    raise InvalidXml("invalid timestamp {}".format(child.text))
+            else:
+                raise InvalidXml("unknown tag {} as child of <history>".format(
+                    child.tag))
+        output = HistoryEntry()
+        output.search = url
+        output.timestamp = timestamp
+        output.history = in_history
+
+        return output
+
 
 class History(models.Model):
     """ A history for a user, containing some web connections (http, https).
     Each history is timed, in a human-behaviour manner. """
 
     start_ts = models.DateTimeField(
-        help_text='The starting timestamp of the history. Useful for cron-like '
-                  'structure.'
+        help_text=('The starting timestamp of the history. Useful for '
+                   'cron-like structure.')
 
     )
     played = models.BooleanField(default=False)
@@ -46,58 +93,134 @@ class History(models.Model):
     def return_history(self):
         """ Returns the history, sorted by increasing timestamps
         """
-        history_set = self.history_set.order_by('timestamp')
-        return history_set
+        output_history = self.historyentry_set.order_by('timestamp')
+        output_history = [(item.search, item.timestamp.date())
+                          for item in output_history]
+        return output_history
 
     def __str__(self):
         """ Returns the string representation of a history.
         """
-        history_set = self.history_set.order_by('timestamp')
-        header = "[History]:\n"
-        return header + "\n".join(history_set)
+        entries = self.historyentry_set.order_by('timestamp')
+        output = "[History]:\n"
+        for entry in entries:
+            output += str(entry) + '\n'
+        return output
 
-
-    def play_history(self):
+    def play_histories(self):
         """ Actually plays the history.
         """
         self.played = True
+        runner = TorInstance(self.return_history())
+        runnner.run()
         self.save()
 
+    def to_xml(self, xml_root):
+        ''' Exports the current history to xml '''
+        hist_node = ET.Element("history", attrib={
+            'start-ts': self.start_ts,
+            'played': 1 if self.played else 0,
+            'user': self.user.pk,
+        })
+        xml_root.append(hist_node)
+        for entry in self.historyentry_set:
+            entry.to_xml(hist_node)
+
+    @staticmethod
+    def from_xml(xml_root):
+        ''' Loads an history from an XML file '''
+
+        REQUIRED_ATTR = ['start-ts', 'played', 'user']
+
+        if xml_root.tag != 'history':
+            raise InvalidXml('unexpected node {} as root of an history'.format(
+                xml_root.tag))
+        for attr in REQUIRED_ATTR:
+            if attr not in xml_root.attrib:
+                raise InvalidXml(('missing attribute "{}" for tag of type '
+                                  'history').format(attr))
+        start_ts = xml_root.attrib['start-ts']
+        played = xml_root.attrib['played']
+        user_pk = xml_root.attrib['user']
+        users = History.objects.filter(pk=1)
+        if len(users) != 1:
+            raise InvalidXml('primary key for History {} is invalid'.format(
+                user_pk))
+
+        output = History()
+        output.start_ts = start_ts
+        output.played = played > 0
+        output.user = users[0]
+
+        for child in xml_root:
+            HistoryEntry.from_xml(child, output)
+
+        return output
+
+
+PartialHistoryEntry = namedtuple('PartialHistoryEntry',
+                                 ['url', 'timestamp'])
+
 
 def generate_partial_history(user, t_start):
     """ Generate the part of the history resulting from the crawl starting at
     the given url.
     """
+    timestamp = t_start
     result = []
     basis = generate_first_url(user)
-    result.append((basis, t_start))
-    t_start += 5* random.weibullvariate(1, 1.5)
-    queue = Queue()
-    crawler = crawl.CrawlingThread(user, basis, queue)
+    result.append(PartialHistoryEntry(basis, timestamp))
+    t_start += 5 * random.weibullvariate(1, 1.5)
+    crawler = crawl.CrawlingThread(basis)
     crawler.start()
     crawler.join()
-    urls = queue.get()
-    for url in urls:
-        t_start += 5* random.weibullvariate(1, 1.5)
-        result.append((url, t_start)
+    urls_tree = crawler.output_tree
+
+    open_time = {}
+    for elem in urls_tree:
+        url, parent = elem.url, elem.parent
+        timestamp = 0
+        if parent is None:
+            timestamp = t_start
+        else:
+            timestamp = open_time[parent] + 5 * random.weibullvariate(1, 1.5)
+        open_time[elem] = timestamp
+        result.append(PartialHistoryEntry(url, timestamp))
     return result
 
+
 def generate_first_url(user):
     """ Generate the first url of a partial history, based on the user
     information. """
-    interest = random.choice(
-        [user.interests.keywords.all(), user.interests.places.all(),
-         user.interests.websites.all(), user.interests.events.all()
-        ]
-    )
+
+    def nonempty(seq):
+        out = []
+        for elt in seq:
+            if elt:
+                out.append(elt)
+        return out
+
+    all_keywords = profiles.Keyword.objects.filter(
+        interest__profile__in=[user])
+    all_websites = profiles.Website.objects.filter(
+        interest__profile__in=[user])
+    all_places = profiles.Place.objects.filter(
+        interest__profile__in=[user])
+    all_events = profiles.Event.objects.filter(
+        interest__profile__in=[user])
+
+    interest = random.choice(nonempty([
+        all_keywords,
+        all_websites,
+        all_places,
+        all_events,
+    ]))
     search_term = random.choice(interest)
     url = search_term.generate_url(user)
     return url
 
 
-
-
-def generate_history(user, ts_start):
+def generate_history(user, start_time):
     """ Generate a new history for the user `user`, starting from timestamp
     `ts_start`.
     A few heuristics are used in order to give the impression that the history
@@ -105,19 +228,32 @@ def generate_history(user, ts_start):
     """
 
     # let's define a new history object.
-    history = History(start_ts=ts_start, user=user)
+    history = History(start_ts=start_time, user=user)
     length = HISTORY_MIN + floor(10 * random.weibullvariate(1, 1.5))
+    history.full_clean()
+    history.save()
 
-    history_line = 0
+    current_timestamp = start_time.timestamp()
 
-    while history_line < length:
-        ts_start += 5 * random.weibullvariate(1, 2.8)
-        history_list = generate_partial_history(user, ts_start)
-        ts_start = history_list[-1].timestamp + 5 * random.weibullvariate(1, 5)
+    hist_size = 0
+
+    while hist_size < length:
+        current_timestamp += 5 * random.weibullvariate(1, 2.8)
+        history_list = generate_partial_history(user, current_timestamp)
+        current_timestamp = \
+            history_list[-1].timestamp + 5 * random.weibullvariate(1, 5)
         for (url, timestamp) in history_list:
-            new_line = HistoryEntry(
-                search=url,
-                timestamp=timestamp,
-                history=history
-            )
-            new_line.save()
+            if len(url) < 200:
+                new_line = HistoryEntry(
+                    search=url,
+                    timestamp=datetime.fromtimestamp(timestamp),
+                    history=history
+                )
+                try:
+                    new_line.full_clean()
+                    new_line.save()
+                    hist_size += 1
+                except ValidationError:
+                    continue
+
+    return history

histories/tor_runner.py

@@ -0,0 +1,120 @@
+"""
+Modules that handles tor instaces creations in order to safely run histories
+"""
+
+import shutil
+import datetime as dt
+from time import sleep
+import asyncio
+import aiohttp
+from aiosocks.connector import ProxyConnector, ProxyClientRequest
+import async_timeout
+import stem.process as tor
+
+class TorInstance():
+    """
+    A tor instance object, with some useful information.
+    It is designed to be used as a worker in order to replay an history.
+    """
+    BASE_SOCKS_PORT = 40000
+    BASE_CONTROL_PORT = 20000
+    BASE_DATA_DIR = "/tmp/tor{}/"
+    TOR_RUNNER = 0
+
+    @classmethod
+    async def create(cls, history, headers):
+        """ Factory creation of tor processes"""
+        socks_port = cls.BASE_SOCKS_PORT + cls.TOR_RUNNER
+        control_port = cls.BASE_CONTROL_PORT + cls.TOR_RUNNER
+        data_dir = cls.BASE_DATA_DIR.format(cls.TOR_RUNNER)
+        TorInstance.TOR_RUNNER += 1
+        self = TorInstance()
+        self.socks_port = socks_port
+        self.control_port = control_port
+        self.data_dir = data_dir
+        self.history = history
+        self.headers = headers
+        self.proxy = "socks5://127.0.0.1:{}".format(self.socks_port)
+        self.create_session()
+        self.process = tor.launch_tor_with_config(
+            config={
+                'ControlPort' : str(control_port),
+                'SocksPort' : str(socks_port),
+                'DataDir' : data_dir
+            }
+        )
+        return self
+
+    def __init__(self):
+        self.socks_port = 0
+        self.control_port = 0
+        self.data_dir = ""
+        self.history = None
+        self.proxy = ""
+        self.headers = {}
+        self.session = None
+        self.process = None
+
+    async def run(self):
+        """ Runs the Tor Instance on the history.
+        """
+        while (self.history[0][1] - dt.datetime.now()).total_seconds >= 10:
+            print("Sleeping")
+            sleep(10)
+        while self.history:
+            item = self.history.pop(0)
+            async with async_timeout.timeout(30):
+                await(self.query(item[0]))
+                now = dt.datetime.now()
+                if now <= self.history[0][1]:
+                    sleep((self.history[0][1] - now).total_seconds())
+
+
+    def create_session(self):
+        """ Create a aiohttp session.
+        """
+        conn = ProxyConnector(remote_resolve=True)
+        self.session = aiohttp.ClientSession(
+            connector=conn,
+            headers=self.headers,
+            request_class=ProxyClientRequest
+        )
+
+
+    async def query(self, url):
+        """ Performs a query.
+        """
+        async with async_timeout.timeout(30):
+            async with self.session.get(
+                url,
+                proxy=self.proxy,
+                proxy_auth=None) as resp:
+                try:
+                    return await resp.text()
+                except UnicodeDecodeError:
+                    return None
+
+
+    def __str__(self):
+        """ Utility function """
+        return ('[TOR] SOCKSPort: {0.socks_port}, ControlPort: '
+                '{0.control_port}, DataDir: {0.data_dir}'.format(self))
+
+    async def kill(self):
+        """ Kills the process and remove the data dir"""
+        self.process.kill()
+        self.session.close()
+        shutil.rmtree(self.data_dir)
+
+
+async def main():
+    """ Test function """
+    for _ in range(3):
+        instance = await TorInstance.create(None, {"user-agent" : "Blah"})
+        await instance.query("https://python.org/")
+        print("One page received")
+        await instance.kill()
+
+if __name__ == "__main__":
+    LOOP = asyncio.get_event_loop()
+    LOOP.run_until_complete(main())

pinocchio/settings.py

@@ -97,7 +97,7 @@ USE_I18N = True
 
 USE_L10N = True
 
-USE_TZ = True
+USE_TZ = False  # We don't really care, we want POSIX timestamps
 
 
 # Static files (CSS, JavaScript, Images)

populate.sh

@@ -0,0 +1,10 @@
+#!/bin/bash
+# -*- coding: UTF8 -*-
+
+python3 manage.py import_browser_fp
+python3 manage.py import_search_engine
+python3 manage.py import_keywords
+python3 manage.py import_website
+python3 manage.py import_places
+python3 manage.py import_events
+python3 manage.py import_interests

profiles/management/commands/exportrdf.py

@@ -0,0 +1,27 @@
+from django.core.management.base import BaseCommand
+from profiles.models_rdf import RdfProfile
+from profiles import models
+
+
+class Command(BaseCommand):
+    ''' Exports database models to RDF '''
+
+    def add_arguments(self, parser):
+        pass
+
+    def handle(self, *args, **kwargs):
+        exported_models = [
+            models.Keyword,
+            models.Webpage,
+            models.Website,
+            models.Place,
+            models.Event,
+            models.BrowserFingerprint,
+            models.SearchEngine,
+            models.Interest,
+            models.Profile,
+        ]
+        output_xml = RdfProfile().serialize(
+        #    models=exported_models,
+        )
+        self.stdout.write(output_xml)

profiles/management/commands/import_events.py

@@ -0,0 +1,41 @@
+""" Small module that import events into the database.
+"""
+
+import json
+from datetime import datetime
+from django.core.management.base import BaseCommand
+from django.db import models
+from profiles.models import Place, Event
+
+def import_file(filename):
+    with open(filename, mode='r') as file:
+        data = json.load(file)
+        for event in data:
+            import_event(event)
+
+def import_place(_place):
+    place = Place(
+        name=_place.get("name", ""),
+        address=_place.get("address", ""),
+        lat=float(_place.get("lat", 0)),
+        lon=float(_place.get("lon", 0))
+    )
+    place.save()
+    return place
+
+def import_event(_event):
+    if isinstance(_event["place"], str):
+        place = Place.objects.get(name=_event["place"])
+    else:
+        place = import_place(_event["place"])
+    event = Event(
+        name=_event.get("name", ""),
+        date=datetime.strptime(_event.get("date", "01/01/1970 00:00 UTC"), "%d/%m/%Y %H:%M %Z"),
+        place=place
+    )
+    #print(event)
+    event.save()
+
+class Command(BaseCommand):
+    def handle(self, *args, **kwargs):
+        import_file("data/events.json")

profiles/management/commands/import_interests.py

@@ -0,0 +1,51 @@
+""" Small module that import interests into the database.
+"""
+
+import json
+from datetime import datetime
+from django.core.management.base import BaseCommand
+from django.db import models
+from django.core.exceptions import ObjectDoesNotExist
+from profiles.models import Keyword, Interest, Place, Website, Event
+
+def import_file(filename):
+    with open(filename, mode='r') as file:
+        data = json.load(file)
+        for interest in data:
+            import_interest(interest)
+
+
+def import_interest(_interest):
+    keywords = []
+    places = []
+    websites = []
+    for keyword in _interest.get("keywords", []):
+        try:
+            stored = Keyword.objects.get(text=keyword["keyword"])
+            keywords.append(stored)
+        except ObjectDoesNotExist:
+            new_keyword = Keyword(text=keyword["keyword"])
+            new_keyword.save()
+            keywords.append(new_keyword)
+            print("New keyword %s" % new_keyword)
+    for place in _interest.get("places", []):
+        places.append(Place.objects.get(name=place["place"]))
+    for website in _interest.get("websites", []):
+        websites.append(Website.objects.get(name=website["website"]))
+
+    interest = Interest(
+        name=_interest.get("name", ""),
+    )
+    interest.save()
+    for keyword in keywords:
+        print(keyword)
+        interest.keywords.add(keyword)
+    for place in places:
+        interest.places.add(place)
+    for website in websites:
+        interest.websites.add(website)
+    interest.save()
+
+class Command(BaseCommand):
+    def handle(self, *args, **kwargs):
+        import_file("data/interests.json")

profiles/management/commands/import_website.py

@@ -0,0 +1,46 @@
+""" Small module that import events into the database.
+"""
+
+import json
+from datetime import datetime
+from django.core.management.base import BaseCommand
+from django.db import models
+from profiles.models import Webpage, Website, Keyword
+
+def import_file(filename):
+    with open(filename, mode='r') as file:
+        data = json.load(file)
+        for website in data:
+            import_website(website)
+
+def import_website(_website):
+    keywords = []
+    webpages = []
+    for keyword in _website.get("keywords", []):
+        new_keyword = Keyword(
+            text=keyword.get("keyword", "")
+        )
+        new_keyword.save()
+        keywords.append(new_keyword)
+    for webpage in _website.get("notable_pages",[]):
+        new_webpage = Webpage(
+            url=webpage.get("keyword", "")
+        )
+        new_webpage.save()
+        webpages.append(new_webpage)
+    website = Website(
+        name=_website.get("name", ""),
+        url=_website.get("url", ""),
+    )
+    website.save()
+    for keyword in keywords:
+        website.keywords.add(keyword)
+    for webpage in webpages:
+        website.notable_pages.add(webpage)
+    print(website)
+    #website.save()
+
+
+class Command(BaseCommand):
+    def handle(self, *args, **kwargs):
+        import_file("data/website.json")

profiles/models.py

@@ -6,9 +6,41 @@ the preferred search engin, and if the user is likely to directly use urls
 or to type in the search engine.
 """
 
+import os
 import random
 from django.db import models
 
+BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+
+NICKNAMES = None
+LASTNAMES = open(BASE_DIR + "/data/lastnames.txt").read().splitlines()
+FIRSTNAMES = open(BASE_DIR + "/data/firstnames.txt").read().splitlines()
+EMAIL_DOMAINS = open(BASE_DIR + "/data/email_domains.txt").read().splitlines()
+
+
+def require_nicknames(fct):
+    def read_file(path):
+        global NICKNAMES
+        print("Trying {}".format(path))
+        with open(path, 'r') as handle:
+            NICKNAMES = handle.read().splitlines()
+
+    nicknames_files = [
+        os.path.join(BASE_DIR, 'data/nicknames_dict'),
+        "/usr/share/dict/american-english",
+    ]
+    if NICKNAMES is None:
+        for nick_file in nicknames_files:
+            try:
+                read_file(nick_file)
+                break
+            except FileNotFoundError:
+                pass
+        if NICKNAMES is None:
+            raise FileNotFoundError
+
+    return fct
+
 
 class InvalidData(Exception):
     ''' Thrown when the DB contains invalid data, and cannot perform
@@ -59,13 +91,13 @@ class Website(models.Model):
         """ Generates the url in case the interest chosen is a website.
         """
         rand = random.random()
-        if user.uses_url:
+        if user.uses_urls:
             url = self.url
         elif rand <= 0.1:
-            url = random.choice(self.notable_pages).url
+            url = random.choice(self.notable_pages.all()).url
         elif rand <= 0.8:
-            search_term_text = self.name +  " " + \
-                    random.choice(self.keywords)
+            search_term_text = self.name + " " + \
+                    random.choice(self.keywords.all())
             url = user.search_engine.search_url(search_term_text)
         else:
             url = user.search_engine.search_url(self.name)
@@ -114,7 +146,6 @@ class Event(models.Model):
         return user.search_engine.search_url(" ".join(possibilities))
 
 
-
 class BrowserFingerprint(models.Model):
     ''' A browser fingerprint, containing things like a user agent '''
 
@@ -139,11 +170,11 @@ class BrowserFingerprint(models.Model):
 
     def serialize_headers(self):
         return {
-            "Description" : str(self.description),
-            "User-Agent" : str(self.useragent),
-            "Accept-Encoding" : str(self.accept_encoding),
-            "Accept" : str(self.accept_default),
-            "Accept-Language" : str(self.accept_lang),
+            "Description": str(self.description),
+            "User-Agent": str(self.useragent),
+            "Accept-Encoding": str(self.accept_encoding),
+            "Accept": str(self.accept_default),
+            "Accept-Language": str(self.accept_lang),
         }
 
 
@@ -154,8 +185,8 @@ class SearchEngine(models.Model):
     url = models.URLField()
     query_pattern = models.CharField(max_length=256)  # This field is the
     # query pattern. It should contain a `{}`, which, when substituted with a
-    # search term (using `.format()`), must yield a URL that can be resolved to
-    # perform the search
+    # search term (using `.format()`), must yield a URL tail that can be
+    # concatenated with `url` to perform a search (eg. `?q={}` for ddg).
 
     def __str__(self):
         return self.name
@@ -163,9 +194,10 @@ class SearchEngine(models.Model):
     def search_url(self, search_term):
         ''' Obtain a url to search `search_term` with this search engine '''
         pattern = str(self.query_pattern)
+        search_term = str(search_term).replace(' ', '+')
         if '{}' not in pattern:
             raise InvalidData("Search engine {}: bad pattern".format(self))
-        return str(self.query_pattern).format(search_term)
+        return self.url + (str(self.query_pattern).format(search_term))
 
 
 class Interest(models.Model):
@@ -199,3 +231,35 @@ class Profile(models.Model):
                                       on_delete=models.CASCADE)
     browser_fingerprint = models.ForeignKey(BrowserFingerprint,
                                             on_delete=models.CASCADE)
+
+
+def generate_email(nick, first_name, last_name):
+    domain = random.choice(EMAIL_DOMAINS)
+    if random.random() < 0.3:
+        email = first_name + "." + last_name + "@" + domain
+    else:
+        email = nick + "@" + domain
+    return email
+
+
+@require_nicknames
+def create_profile(nick=None):
+    nick = "".join(random.sample(NICKNAMES, random.randrange(2, 5)))
+    first_name = random.choice(FIRSTNAMES)
+    last_name = random.choice(LASTNAMES)
+    email = generate_email(nick, first_name, last_name)
+    profile = Profile(
+        nick=nick,
+        first_name=first_name,
+        last_name=last_name,
+        email=email,
+        uses_urls=(random.random() < 0.5),
+    )
+    profile.search_engine = random.choice(SearchEngine.objects.all())
+    profile.browser_fingerprint = random.choice(BrowserFingerprint.objects.all())
+
+    profile.full_clean()
+    profile.save()
+    profile.interests.add(random.choice(Interest.objects.all()))
+    profile.save()
+    return profile

profiles/models_rdf.py

@@ -0,0 +1,131 @@
+""" RDF serialization class for profile models """
+
+import rdfserializer as rdf
+from rdfserializer import RDFModelSerialiser as RDFModelSerializer
+# ^ This was hurting my eyes way too much
+from rdfserializer import SCHEMA as schema
+from rdflib.namespace import Namespace
+
+import profiles.models as profile_models
+
+
+LOCAL_NS = Namespace('local:')
+
+
+class RdfWebpage(RDFModelSerializer):
+    """ RDF serializer for Webpage """
+
+    _type = schema.WebPage
+    model = profile_models.Webpage
+    entries = [
+        rdf.RDFSimpleField(schema.url, 'url'),
+    ]
+
+
+class RdfWebsite(RDFModelSerializer):
+    """ RDF serializer for Website """
+
+    _type = schema.WebSite
+    model = profile_models.Website
+    entries = [
+        rdf.RDFSimpleField(schema.name, 'name'),
+        rdf.RDFSimpleField(schema.url, 'url'),
+        rdf.RDFManyField(schema.keywords, 'keywords',
+                         lambda keyword: keyword.text),
+        rdf.RDFManyLinker(schema.hasPart, 'notable_pages', RdfWebpage),
+    ]
+
+
+class RdfPlace(RDFModelSerializer):
+    """ RDF serializer for Place """
+
+    _type = schema.Place
+    model = profile_models.Place
+    entries = [
+        rdf.RDFSimpleField(schema.name, 'name'),
+        rdf.RDFSimpleField(schema.address, 'address'),
+        rdf.RDFSimpleField(schema.latitude, 'lat'),
+        rdf.RDFSimpleField(schema.longitude, 'lon'),
+    ]
+
+
+class RdfEvent(RDFModelSerializer):
+    """ RDF serializer for Event """
+
+    _type = schema.Event
+    model = profile_models.Event
+    entries = [
+        rdf.RDFSimpleField(schema.name, 'name'),
+        rdf.RDFSimpleField(schema.startDate, 'date'),
+        rdf.RDFLeftBinder(schema.location, 'place', RdfPlace),
+    ]
+
+
+class RdfBrowserFingerprint(RDFModelSerializer):
+    """ RDF serializer for BrowserFingerprint """
+
+    _type = schema.Intangible
+    model = profile_models.BrowserFingerprint
+    entries = [
+        rdf.RDFSimpleField(schema.description, 'description'),
+        rdf.RDFSimpleField(LOCAL_NS.useragent, 'useragent'),
+        rdf.RDFSimpleField(LOCAL_NS.appname, 'appname'),
+        rdf.RDFSimpleField(LOCAL_NS.appversion, 'appversion'),
+        rdf.RDFSimpleField(LOCAL_NS.platform, 'platform'),
+        rdf.RDFSimpleField(LOCAL_NS.vendor, 'vendor'),
+        rdf.RDFSimpleField(LOCAL_NS.vendorsub, 'vendorsub'),
+        rdf.RDFSimpleField(LOCAL_NS.buildID, 'buildID'),
+        rdf.RDFSimpleField(LOCAL_NS.oscpu, 'oscpu'),
+        rdf.RDFSimpleField(LOCAL_NS.accept_encoding, 'accept_encoding'),
+        rdf.RDFSimpleField(LOCAL_NS.accept_default, 'accept_default'),
+        rdf.RDFSimpleField(LOCAL_NS.accept_lang, 'accept_lang'),
+        rdf.RDFSimpleField(LOCAL_NS.pixeldepth, 'pixeldepth'),
+        rdf.RDFSimpleField(LOCAL_NS.colordepth, 'colordepth'),
+        rdf.RDFSimpleField(LOCAL_NS.screens, 'screens'),
+    ]
+
+
+class RdfSearchEngine(RDFModelSerializer):
+    """ RDF serializer for SearchEngine """
+
+    _type = schema.WebSite
+    model = profile_models.SearchEngine
+    entries = [
+        rdf.RDFSimpleField(schema.url, 'url'),
+        rdf.RDFSimpleField(schema.name, 'name'),
+        rdf.RDFSimpleField(LOCAL_NS.query_pattern, 'query_pattern'),
+    ]
+
+
+class RdfInterest(RDFModelSerializer):
+    """ RDF serializer for Interest """
+
+    Interesttype = 'interest'
+    model = profile_models.Interest
+    entries = [
+        rdf.RDFSimpleField(schema.name, 'name'),
+        rdf.RDFManyField(schema.keywords, 'keywords',
+                         lambda keyword: keyword.text),
+        rdf.RDFManyLinker(schema.location, 'places', RdfPlace),
+        rdf.RDFManyLinker(schema.website, 'websites', RdfWebsite),
+        rdf.RDFManyLinker(schema.event, 'events', RdfEvent),
+    ]
+
+
+class RdfProfile(RDFModelSerializer):
+    """ RDF serializer for Profile """
+
+    _type = schema.Person
+    model = profile_models.Profile
+    entries = [
+        rdf.RDFSimpleField(LOCAL_NS.nickname, 'nick'),
+        rdf.RDFSimpleField(schema.given_name, 'first_name'),
+        rdf.RDFSimpleField(schema.family_name, 'last_name'),
+        rdf.RDFSimpleField(schema.email, 'email'),
+        rdf.RDFSimpleField(LOCAL_NS.uses_urls, 'uses_urls'),
+        rdf.RDFManyLinker(LOCAL_NS.interest, 'interests', RdfInterest),
+        rdf.RDFLeftBinder(LOCAL_NS.search_engine, 'search_engine',
+                          RdfSearchEngine),
+        rdf.RDFLeftBinder(LOCAL_NS.browser_fingerprint, 'browser_fingerprint',
+                          RdfBrowserFingerprint)
+    ]

requirements.txt

@@ -1,5 +1,5 @@
 aiodns==1.1.1
-aiohttp==3.0.1
+aiohttp==2.3.2
 async-timeout==2.0.0
 attrs==17.4.0
 cchardet==2.1.1
@@ -12,3 +12,7 @@ pycares==2.3.0
 pytz==2017.3
 yarl==1.1.1
 beautifulsoup4==4.6.0
+stem==1.6.0
+pycurl==7.43.0.1
+rdflib==4.2.2
+git+https://github.com/tobast/RDFSerializer.git