[filter] improve Join filter

The filter will now act a little bit more like the join command
I also added some parameters that will help formatting (newLine, addBedore, addAfter)

modules/cci/pages.py

@@ -61,7 +61,7 @@ class SearchPage(HTMLPage):
         klass = BaseJobAdvert
 
         obj_url = Format('%s#%s', Env('url'), Env('id'))
-        obj_description = Join('%s\r\n',
+        obj_description = Join('\r\n',
                                'div/fieldset/*[(@class="titreParagraphe" or @class="normal")]',
                                textCleaner=CleanHTML)
         obj_title = CleanText('div/span[@class="intituleposte"]')

modules/monster/pages.py

@@ -73,7 +73,7 @@ class AdvertPage(HTMLPage):
         obj_url = BrowserURL('advert', _id=Env('_id'))
         obj_title = CleanText('//div[@id="jobcopy"]/h1[@itemprop="title"]|//div[@itemprop="title"]/h1')
         obj_description = CleanHTML('//div[@id="jobBodyContent"]|//div[@itemprop="description"]')
-        obj_contract_type = Join('%s ', '//dd[starts-with(@class, "multipledd")]')
+        obj_contract_type = Join(' ', '//dd[starts-with(@class, "multipledd")]')
         obj_society_name = CleanText('//dd[@itemprop="hiringOrganization"]')
         obj_place = CleanText('//span[@itemprop="jobLocation"]')
         obj_pay = CleanText('//span[@itemprop="baseSalary"]')

modules/regionsjob/pages.py

@@ -64,7 +64,7 @@ class AdvertPage(HTMLPage):
     class get_job_advert(ItemElement):
         klass = BaseJobAdvert
 
-        obj_description = Join('\n%s', '//div[@id="annonce-detail"]/p[@class="text"]', textCleaner=CleanHTML)
+        obj_description = Join('\n', '//div[@id="annonce-detail"]/p[@class="text"]', textCleaner=CleanHTML)
         obj_id = Env('_id')
         obj_url = BrowserURL('advert_page', _id=Env('_id'))
         obj_publication_date = Date(Regexp(CleanText('//div[@id="annonce-detail"]/p[@class="infos"]'),

modules/senscritique/pages.py

@@ -125,8 +125,8 @@ class Description(Filter):
         return Format(u'%s %s\n\n%s%s\n\n',
                       CleanText("%s/div[@class='d-rubric-inner']/h1" % header),
                       CleanText("%s/div[@class='d-rubric-inner']/small" % header),
-                      Join(u'- %s\n', "%s/ul[@class='pvi-product-specs']/li" % header),
-                      Join(u'- %s\n', "%s/ul/li" % section))(el[0])
+                      Join(u'- ', "%s/ul[@class='pvi-product-specs']/li" % header, newline=True),
+                      Join(u'- ', "%s/ul/li" % section, newline=True, addBefore=' - '))(el[0])
 
 
 class EventPage(HTMLPage):

weboob/browser/filters/standard.py

@@ -690,19 +690,31 @@ class BrowserURL(MultiFilter):
 
 
 class Join(Filter):
-    def __init__(self, pattern, selector=None, textCleaner=CleanText):
+    def __init__(self, pattern, selector=None, textCleaner=CleanText, newline=False, addBefore='', addAfter=''):
         super(Join, self).__init__(selector)
         self.pattern = pattern
         self.textCleaner = textCleaner
+        self.newline = newline
+        self.addBefore = addBefore
+        self.addAfter = addAfter
 
     @debug()
     def filter(self, el):
-        res = u''
-        for li in el:
-            res += self.pattern % self.textCleaner.clean(li)
+        items = [self.textCleaner.clean(e) for e in el]
+        items = [item for item in items if item]
 
-        return res
+        if self.newline:
+            items = ['%s\r\n' % item for item in items]
 
+        result = self.pattern.join(items)
+
+        if self.addBefore:
+            result = '%s%s' % (self.addBefore, result)
+
+        if self.addAfter:
+            result = '%s%s' % (result, self.addAfter)
+
+        return result
 
 class Eval(MultiFilter):
     """