src/hdalab/management/commands/query_wikipedia_category.py
changeset 271 8f77cf71ab02
parent 114 c59383cc9940
child 571 d9642be7c937
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/hdalab/management/commands/query_wikipedia_category.py	Tue Jun 17 10:25:33 2014 +0200
@@ -0,0 +1,396 @@
+# -*- coding: utf-8 -*-
+'''
+Created on Jun 7, 2011
+
+@author: ymh
+'''
+
+from django.conf import settings
+from django.core.management.base import NoArgsCommand
+from django.core.management.color import no_style
+from hdabo.models import Tag
+from hdalab.models import WpCategory, TagWpCategory, TagInfobox, InfoboxParameter
+from optparse import make_option
+from wikitools import api,wiki
+import sys
+import re
+import itertools
+from hdabo import utils
+from django.db.models import Count
+from django.db import transaction
+
+
+TYPES_MASK_DICT = {
+        u'visible': 0b001,
+        u'hidden': 0b010,
+        u'infobox': 0b100,
+        u'all': 0b111,
+    }
+
+START_PATTERN = re.compile(u"\{\{\s?Infobox\s+([^|]+)", re.M|re.U|re.I)
+END_PATTERN = re.compile(u"\{\{|\}\}", re.M|re.U)
+SPLIT_PATTERN = re.compile("\s*?\|\s*([\w]+[^=|]*)\s*=", re.U|re.M)
+DELIMITER_PATTERN = re.compile("\{{2,3}|\}{2,3}|\[\[|\]\]|\[|\]")
+COMMENT_PATTERN = re.compile("<!--.*?-->",re.U|re.M)
+
+
+
+class Command(NoArgsCommand):
+    '''
+    query and update wikipedia for tag title.
+    '''
+    options = ''
+    help = """query and update wikipedia for tag title."""
+    
+    option_list = NoArgsCommand.option_list + (
+        make_option('--all',
+            action='store_true',
+            dest='all',
+            default=False,
+            help='force all tags to be updated, not only those not yet processed'),
+        make_option('--force',
+            action='store_true',
+            dest='force',
+            default=False,
+            help='ask no questions'),
+        make_option('--random',
+            action='store_true',
+            dest='random',
+            default=False,
+            help='randomize query on tags'),
+        make_option('--site',
+            action='store',
+            type='string',
+            dest='site_url',
+            default="http://fr.wikipedia.org/w/api.php",
+            help='the url for the wikipedia site'),
+        make_option('--limit',
+            action='store',
+            type='int',
+            dest='limit',
+            default= -1,
+            help='number of tag to process'),
+        make_option('--start',
+            action='store',
+            type='int',
+            dest='start',
+            default=0,
+            help='number of tag to ignore'),
+        make_option('--type',
+            action='append',
+            dest='types',
+            type='choice',
+            choices=['visible','hidden', 'infobox', 'all'],
+            default=[],
+            help='what type of query to perform : visible : visible categories, hidden : hidden categories, infobox: infoboxes, all: all of them. This option can be assed multiple times'),
+        make_option('--use-label',
+            action='store_true',
+            dest='use_label',
+            default=False,
+            help='use label instead of pageid to query wikipedia'),
+        make_option('--tag',
+            action='append',
+            dest='tags',
+            type='string',
+            default=[],
+            help='the tag to query'),
+
+    )
+    
+    
+#    def process_wp_response(self, label, response):        
+#
+#        query_dict = response['query']
+#        # get page if multiple pages or none -> return Tag.null_result
+#        pages = query_dict.get("pages", {})
+#        if len(pages) > 1 or len(pages) == 0:
+#            return None, Tag.TAG_URL_STATUS_DICT["null_result"], None, None
+#        
+#        page = pages.values()[0]
+#        
+#        if u"invalid" in page or u"missing" in page:
+#            return None, Tag.TAG_URL_STATUS_DICT["null_result"], None, None
+#
+#        url = page.get(u'fullurl', None)
+#        pageid = page.get(u'pageid', None)
+#        new_label = page[u'title']
+#        
+#        if self.__is_homonymie(page):
+#            status = Tag.TAG_URL_STATUS_DICT["homonyme"]
+#        elif u"redirect" in page:
+#            status = Tag.TAG_URL_STATUS_DICT["redirection"]
+#        else:
+#            status = Tag.TAG_URL_STATUS_DICT["match"]
+#        
+#        return new_label, status, url, pageid 
+
+    def query_all_categories(self, hidden, site, pageid, use_label):
+        
+        clshow = 'hidden' if hidden else '!hidden'
+        params = {'action':'query', 'titles' if use_label else 'pageids': pageid, 'prop':'categories', 'clshow': clshow}
+        
+        clcontinue = ""        
+        res = []
+        
+        while clcontinue is not None:
+            if clcontinue:
+                params['clcontinue'] = clcontinue
+                
+            wpquery = api.APIRequest(site, params) #@UndefinedVariable
+            response = wpquery.query()
+            
+            if self.verbosity > 1:
+                print "Query infoboxes : " + repr(wpquery.request.get_full_url()+"?"+wpquery.request.get_data())
+                print repr(response)
+
+            
+            query_dict = response.get('query', None)
+            
+            if query_dict is None:
+                return res
+            
+            pages = query_dict.get("pages", {})
+            if len(pages) > 1 or len(pages) == 0:
+                return res
+            
+            page = pages.values()[0]
+                        
+            for cat in page.get('categories',[]):
+                title = cat.get('title',"")
+                title = title[title.find(":")+1:]
+                if title and clcontinue != ("%s|%s" % (pageid,title)):
+                    res.append(title)
+            
+            clcontinue = response.get('query-continue', {}).get('categories',{}).get('clcontinue', None)
+
+        if self.verbosity > 1:
+            print "Query infoboxes RES: "
+            print repr(res)
+            
+        return res
+    
+    def process_categories(self, cat_list, hidden, tag):
+        
+        for cat in cat_list:
+            wp_cat,created = WpCategory.objects.get_or_create(label=cat) #@UnusedVariable
+            TagWpCategory.objects.get_or_create(tag=tag, wp_category=wp_cat, hidden=hidden)
+            
+                
+    def query_infoboxes(self, site, pageid, use_label):
+        
+        res = []
+        params = {'action':'query', 'titles' if use_label else 'pageids': pageid, 'prop':'revisions', 'rvprop': 'ids|content'}
+        wpquery = api.APIRequest(site, params) #@UndefinedVariable
+        response = wpquery.query()
+        
+        query_dict = response.get('query', None)
+            
+        if query_dict is None:
+            return res
+            
+        pages = query_dict.get("pages", {})
+        if len(pages) > 1 or len(pages) == 0:
+            return res
+
+        page = pages.values()[0]
+        
+        if 'revisions' not in page or not page['revisions']:
+            return res
+        
+        rev = page['revisions'][0]
+        
+        content = rev['*']
+                
+        start = 0
+        depth = 0
+        current_infobox_name = None
+        current_start = 0
+        
+        while start <= len(content):
+            if depth==0:
+                resm = START_PATTERN.search(content[start:])
+                if resm is None:
+                    break
+                depth = 1
+                current_start = resm.start()+start
+                start += resm.end()+1
+                current_infobox_name = resm.group(1)                    
+            else:
+                resm = END_PATTERN.search(content[start:])
+                if resm is None:
+                    break
+                if resm.group(0) == "{{":
+                    depth += 1
+                elif resm.group(0) == "}}":
+                    depth -= 1
+                if depth == 0:
+                    res.append((content[current_start:resm.end()+start], current_infobox_name))
+                start += resm.end()+1
+
+        return_val = (rev['revid'],res)
+        
+        if self.verbosity > 1:
+            print "Query infoboxes url: " + repr(wpquery.request.get_full_url()+"?"+wpquery.request.get_data())
+            print repr(return_val)
+        
+        return return_val
+    
+    def split_infoboxes(self, src):
+        
+        start = 0
+        previous_end = 0
+        split_indexes = []
+        delimiter_stack = []
+        while start<=len(src):            
+            resd = DELIMITER_PATTERN.search(src[start:])
+            ress = SPLIT_PATTERN.search(src[start:]) if len(delimiter_stack) == 0 else None
+            startd = resd.start() if resd is not None else sys.maxint
+            starts = ress.start() if ress is not None else sys.maxint
+            if starts < startd:
+                if len(split_indexes)>0:
+                    split_indexes.append((previous_end, ress.start(0)+start))
+                split_indexes.append((ress.start(1)+start, ress.end(1)+start))
+                start += ress.end(0)
+                previous_end = start
+            elif startd < sys.maxint:
+                if resd.group().startswith("{") or resd.group().startswith("[") :
+                    delimiter_stack.append(resd.group())
+                elif len(delimiter_stack)>0 and ( (delimiter_stack[-1].startswith('{') and resd.group()[0] == '}') or (delimiter_stack[-1].startswith('[') and resd.group()[0] == ']') ) and len(delimiter_stack[-1]) == len(resd.group()):
+                    delimiter_stack.pop()
+                start += resd.end()
+            else:
+                break
+            
+        if previous_end > 0:
+            split_indexes.append((previous_end,len(src)))
+        res = [src[start:end] for start,end in split_indexes]
+        return res
+
+
+
+    def process_infoboxes(self, infobox_defs, tag):
+        
+        if not infobox_defs:
+            return
+        
+        revision_id = infobox_defs[0]
+        for infobox in infobox_defs[1]:
+            src = infobox[0].strip(' \t\n\r')            
+            name = infobox[1]
+            tag_infobox, created = TagInfobox.objects.get_or_create(tag=tag, name=name, revision_id = revision_id, defaults={'source': src})
+            if not created:
+                tag_infobox.source = src
+                tag_infobox.save()
+
+            src = COMMENT_PATTERN.sub('',src)
+            src = START_PATTERN.sub('',src[:-2]).strip()
+            
+            keyvalues = self.split_infoboxes(src)
+
+            for key,value in itertools.izip(*[itertools.islice(keyvalues, i, None, 2) for i in range(2)]):
+                param, created = InfoboxParameter.objects.get_or_create(tag_infobox=tag_infobox, param_name=key.strip(), defaults={'param_value':value.strip()})
+                if not created:
+                    param.param_value = value.strip()
+                    param.save()
+        
+    def handle_noargs(self, **options):
+        
+        self.style = no_style()
+        
+        interactive = options.get('interactive', True)
+        
+        self.verbosity = int(options.get('verbosity', '1'))
+        use_label = options.get('use_label', False)
+        
+        force = options.get('force', False)
+        
+        limit = options.get("limit", -1)
+        start = options.get("start", 0)
+        
+        site_url = options.get('site_url', settings.WIKIPEDIA_API_URL)
+        
+        random = options.get('random', False)
+        
+        types_mask = 0
+        types_list = options.get('types', [])
+        
+        if len(types_list) == 0:
+            types_mask = TYPES_MASK_DICT['all']
+        else:
+            for t in types_list:
+                types_mask |=  TYPES_MASK_DICT[t]
+                
+        if self.verbosity > 1 :
+            print "types mask %s " % (bin(types_mask))  
+        
+        if self.verbosity > 2:
+            print "option passed : " + repr(options)
+
+
+        queryset = Tag.objects.exclude(wikipedia_pageid= None)
+        
+        tag_list = options.get("tags", []);
+        
+        if tag_list:
+            queryset = queryset.filter(label__in=tag_list)
+        elif not options.get('all',False):            
+            queryset = queryset.annotate(wpc=Count('wp_categories')).filter(wpc = 0)
+        #else:
+        #    queryset = Tag.objects.filter(url_status=None)                    
+        
+        if random:
+            queryset = queryset.order_by("?")
+        else:
+            queryset = queryset.order_by("label")
+        
+        if limit >= 0:
+            queryset = queryset[start:limit]
+        elif start > 0:
+            queryset = queryset[start:]            
+        
+        if self.verbosity > 2 :
+            print "Tag Query is %s" % (queryset.query)
+        
+        site = wiki.Wiki(site_url) #@UndefinedVariable
+        
+        
+        count = queryset.count()
+        if self.verbosity > 1:
+            print "Processing %d tags" % (count)
+        
+        if not force and interactive:
+            confirm = raw_input("You have requested to query and replace the wikipedia information for %d tags.\n Are you sure you want to do this? \nType 'yes' to continue, or 'no' to cancel: " % (count))
+        else:
+            confirm = 'yes'
+            
+        if confirm != "yes":
+            print "wikipedia query cancelled"
+            return
+
+        
+        
+        for i, tag in enumerate(queryset):
+            
+            if self.verbosity > 1:
+                print "processing tag %s (%d/%d)" % (tag.label, i + 1, count)
+            else:
+                utils.show_progress(i + 1, count, tag.label, 60)                            
+
+            # query categories
+            wikipedia_pageid = tag.label if use_label else tag.wikipedia_pageid
+            if tag.url_status == Tag.TAG_URL_STATUS_DICT['redirection'] and tag.alternative_wikipedia_pageid is not None :
+                wikipedia_pageid = tag.alternative_label if use_label else tag.alternative_wikipedia_pageid
+
+            with transaction.commit_on_success():
+                if types_mask & TYPES_MASK_DICT['visible']:
+                    res = self.query_all_categories(False, site, wikipedia_pageid, use_label)
+                    self.process_categories(res, False, tag)
+    
+                if types_mask & TYPES_MASK_DICT['hidden']:
+                    res = self.query_all_categories(True, site, wikipedia_pageid, use_label)
+                    self.process_categories(res, True, tag)
+                
+                if types_mask & TYPES_MASK_DICT['infobox']:
+                    res = self.query_infoboxes(site, wikipedia_pageid, use_label)
+                    self.process_infoboxes(res, tag)
+