src/ldtplatform/management/commands/loadandadddata.py
author ymh <ymh.work@gmail.com>
Wed, 03 Jul 2024 14:26:30 +0200
changeset 369 1395a729018f
parent 102 e2968797bdae
permissions -rw-r--r--
Set about link in footer to https

# -*- coding: utf-8 -*-
'''
Created on Mar 22, 2013

@author: tc
'''

from django.contrib.auth.models import User, Group
from django.core.management import call_command
from django.core.management.base import BaseCommand, CommandError
from ldt.ldt_utils.models import Content, Project
from ldt.security.cache import cached_assign
from optparse import make_option
import json
import os.path


class Command(BaseCommand):
    '''
    Load users, medias, contents, project and guardian permissions from json file generated by dumpdata
    '''

    args = 'json_file'
    help = 'Load users, medias, contents, project and guardian permissions from json file generated by dumpdata'
    
    option_list = BaseCommand.option_list + (
        make_option('-u', '--ignore-users',
            dest= 'ignore_users',
            default= None,
            help= 'list of usernames to ignore (separated by comma)'
        ),
        make_option('-g', '--ignore-groups',
            dest= 'ignore_groups',
            default= None,
            help= 'list of group names to ignore (separated by comma)'
        ),
        make_option('-c', '--ignore-contents',
            dest= 'ignore_contents',
            default= None,
            help= 'list of content iri_id to ignore (separated by comma)'
        ),
        make_option('-n', '--new-group',
            dest= 'new_group',
            default= None,
            help= 'The script will create a new group and assign view permission for all new media/contents to all the new users'
        ),
    )
    
    
    def __safe_get(self, dict_arg, key, conv = lambda x: x, default= None):
        val = dict_arg.get(key, default)
        return conv(val) if val else default

    def __safe_decode(self, s):
        if not isinstance(s, basestring):
            return s
        try:
            return s.decode('utf8')
        except:
            try:
                return s.decode('latin1')
            except:
                return s.decode('utf8','replace')

    def handle(self, *args, **options):
        
        # Test path
        if len(args) != 1:
            raise CommandError("The command has no argument or too much arguments. Only one is needed : the json file path.")
        
        # Check if temporary files already exist
        path = os.path.abspath(args[0])
        dirpath = os.path.dirname(path)
        path_file1 = os.path.join(dirpath, 'temp_data1.json')
        path_file2 = os.path.join(dirpath, 'temp_data2.json')
        do_import = True
        if os.path.exists(path_file1) or os.path.exists(path_file2):
            confirm = raw_input(("""
    The folder %s contains the files temp_data1.json or temp_data2.json. These files will be overwritten.
    
    Do you want to continue ?

    Type 'y' to continue, or 'n' to quit: """) % dirpath)
            do_import = (confirm == "y")
        
        # Continue
        if do_import:
            # Init ignore list
            user_ignore_list = ["admin","AnonymousUser"]
            group_ignore_list = ["everyone","Hashcut IRI","Hashcut BPI"]
            content_ignore_list = []
            
            # Update ignore list
            ignore_users = options.get('ignore_users', None)
            ignore_groups = options.get('ignore_groups', None)
            ignore_contents = options.get('ignore_contents', None)
            if ignore_users:
                for u in ignore_users.split(","):
                    user_ignore_list.append(u)
            if ignore_groups:
                for g in ignore_groups.split(","):
                    group_ignore_list.append(g)
            if ignore_contents:
                for c in ignore_contents.split(","):
                    content_ignore_list.append(c)
            
            # Begin work...
            print("Opening file...")
            json_file = open(path,'rb')
            print("Loading datas...")
            data = json.load(json_file)
            print("%d objects found..." % len(data))
            content_pk_id = {}
            project_pk_id = {}
            # datas for file 1 : users, medias, contents, projects 
            data_file1 = []
            # datas for file 2 : guardian permissions
            data_file2 = []
            # users
            usernames = []
            for obj in data:
                if "model" in obj:
                    m = obj["model"]
                    if m!="guardian.userobjectpermission" and m!="guardian.groupobjectpermission":
                        # We remove user admin, user AnonymousUser, group everyone and users and contents in ignore list
                        # (a bit fuzzy for media and src but good for others)
                        if not ((m=="auth.user" and "username" in obj["fields"] and obj["fields"]["username"] in user_ignore_list) or \
                                (m=="auth.group" and "name" in obj["fields"] and obj["fields"]["name"] in group_ignore_list) or \
                                (m=="ldt_utils.media" and "src" in obj["fields"] and any((s+".") in obj["fields"]["src"] for s in content_ignore_list)) or \
                                (m=="ldt_utils.content" and "iri_id" in obj["fields"] and obj["fields"]["iri_id"] in content_ignore_list)):
                            data_file1.append(obj)
                        #else:
                        #    print("I don't keep from datas %s, pk = %s" % (m, obj["pk"]))
                        if "pk" in obj:
                            # For both contents and projects, we save 2 dicts [id]=pk and [pk]=id
                            # It will enable to parse and replace easily the old pk by the new ones in the permission datas
                            if m=="ldt_utils.project":
                                pk = str(obj["pk"])
                                ldt_id = obj["fields"]["ldt_id"]
                                project_pk_id[pk] = ldt_id
                            elif m=="ldt_utils.content":
                                pk = str(obj["pk"])
                                ldt_id = obj["fields"]["iri_id"]
                                content_pk_id[pk] = ldt_id
                            obj["pk"] = None
                    else:
                        obj["pk"] = None
                        data_file2.append(obj)
                    # Save usernames except AnonymousUser 
                    if m=="auth.user" and "username" in obj["fields"] and obj["fields"]["username"]!="AnonymousUser":
                        usernames.append(obj["fields"]["username"])
            json_file.close()
            #data_file1.append(project_pk_id)
            #data_file1.append(project_id_pk)
            #data_file1.append(content_pk_id)
            #data_file1.append(content_id_pk)
            
            # Check if import will fail with the usernames
            existing_usernames = User.objects.all().values_list("username", flat=True)
            for un in usernames:
                if un in existing_usernames and un not in user_ignore_list:
                    print("import will fail with username : %s" % str(un))
                    do_import = False
            
            # Check if import will fail with the contents's iri_id
            existing_iri_ids = Content.objects.all().values_list("iri_id", flat=True)
            new_iri_ids = list(content_pk_id.values())
            for iri_id in new_iri_ids:
                if iri_id in existing_iri_ids and iri_id not in content_ignore_list:
                    print("import will fail with iri_id : %s" % str(iri_id))
                    do_import = False
            if not do_import:
                print("Add the usernames and iri_id to the ignore parameters -u and -c")
                return ""
            
            # We save the datas in a file in order to simply call loaddata
            print("Writing %s..." % path_file1)
            file1 =  open(path_file1, 'w')
            json.dump(data_file1, file1, indent=2)
            file1.close()
            print("Updating permissions ids...")
            # We replace the old pk by the natural keys in the permission datas
            ignored_project_pks = []
            ignored_content_pks = []
            perm_data = []
            for obj in data_file2:
                content_type = obj["fields"]["content_type"][1]
                old_pk = obj["fields"]["object_pk"]
                if content_type =="project":
                    try:
                        obj["fields"]["object_pk"] = project_pk_id[old_pk]
                    except:
                        # The dumpdata can contain permissions for removed projects
                        ignored_project_pks.append(old_pk)
                        continue
                    # Keeping only valuables objs avoids errors when we we get the new pks
                    perm_data.append(obj)
                elif content_type == "content":
                    try:
                        obj["fields"]["object_pk"] = content_pk_id[old_pk]
                    except:
                        # The dumpdata can contain permissions for removed contents
                        ignored_content_pks.append(old_pk)
                        continue
                    # Keeping only valuables objs avoids errors when we we get the new pks
                    obj_id = obj["fields"]["object_pk"]
                    model = obj["model"] # "guardian.groupobjectpermission" or "guardian.userobjectpermission"
                    if obj_id in content_ignore_list:
                        if model=="guardian.groupobjectpermission":
                            if obj["fields"]["group"][0] in group_ignore_list:
                                #print("permissions : j'ignore %s pour le groupe %s ..." % (obj_id, obj["fields"]["group"][0]))
                                continue
                        elif model=="guardian.userobjectpermission":
                            if obj["fields"]["user"][0] in user_ignore_list:
                                #print("permissions : j'ignore %s pour le user %s ..." % (obj_id, obj["fields"]["user"][0]))
                                continue
                    perm_data.append(obj)
            # We inform the user
            print("%d project permissions were ignored because projects do not exist in the current datas." % len(ignored_project_pks))
            print("%d content permissions were ignored because contents do not exist in the current datas." % len(ignored_content_pks))
            print("Loading datas from temporary file %s ..." % path_file1)
            # Loaddata from file 1
            call_command("loaddata", path_file1)
            
            # Now users, medias, contents, projects have been saved.
            # We can get the new pk for contents and projects
            # Careful: in Python 3, dict.copy().values() will be prefered to list(dict.values())
            # We use select_related("media_obj") because it will usefull with the new group
            contents = Content.objects.filter(iri_id__in=list(content_pk_id.values())).select_related("media_obj")#.values('pk', 'iri_id')
            content_id_pk = {}
            for c in contents:
                content_id_pk[c.iri_id] = str(c.pk)
            projects = Project.objects.filter(ldt_id__in=list(project_pk_id.values())).values('pk', 'ldt_id')
            project_id_pk = {}
            for p in projects:
                project_id_pk[p["ldt_id"]] = str(p["pk"])
            
            # Now we reparse the perm_data and update with the new pks
            for obj in perm_data:
                content_type = obj["fields"]["content_type"][1]
                obj_id = obj["fields"]["object_pk"]
                if content_type=="project":
                    obj["fields"]["object_pk"] = project_id_pk[obj_id]
                elif content_type == "content":
                    obj["fields"]["object_pk"] = content_id_pk[obj_id]
            
            
            # We save the datas in a file in order to simply call loaddata
            print("Writing %s..." % path_file2)
            file2 =  open(path_file2, 'w')
            json.dump(perm_data, file2, indent=2)
            file2.close()
            print("Loading permissions from temporary file %s ..." % path_file2)
            call_command("loaddata", path_file2)
            
            # Remove temp files
            print("Removing temporary files...")
            try:
                os.remove(path_file1)
            except:
                print("Removing temporary files %s failed" % path_file1)
            try:
                os.remove(path_file2)
            except:
                print("Removing temporary files %s failed" % path_file2)
            
            # Now that all datas have been imported we can create the new group and assign permissions if asked
            new_group = options.get('new_group', None)
            if new_group and len(usernames)>0:
                print("Set view permissions for the new group %s ..." % new_group)
                # Get or create group
                new_grp, _ = Group.objects.get_or_create(name=new_group)
                # Add users to the group
                users = User.objects.filter(username__in=usernames)
                for u in users:
                    new_grp.user_set.add(u)
                # Get all contents and medias
                for c in contents:
                    cached_assign('view_content', new_grp, c)
                    cached_assign('view_media', new_grp, c.media_obj)
                
            print("Indexing imported projects ...")
            call_command('reindex', projects=True, no_content=True)
        
        # This is the end
        print("This is the end")