path: root/ishtar_common/serializers_utils.py

from collections import OrderedDict
from copy import deepcopy
import datetime
import json
import importlib
import os
import tempfile
from zipfile import ZipFile

from django.contrib.sites.models import Site
from django.core.serializers import serialize
from django.db.models import Q

from ishtar_common.version import get_version
from . import models


SERIALIZATION_VERSION = "1.0"


def get_model_from_filename(filename):
    filename = filename.split(".")[0]  # remove extension
    splitted = filename.split("__")
    if len(splitted) != 2:
        return
    module_name, model_name = splitted
    if module_name == "django":
        if model_name in ("Group", "Permission"):
            module = importlib.import_module("django.contrib.auth.models")
        elif model_name in ("ContentType",):
            module = importlib.import_module("django.contrib.contenttypes.models")
        else:
            return
    else:
        module = importlib.import_module(module_name + ".models")
    return getattr(module, model_name)


def serialization_info(info=None):
    site = Site.objects.get_current()
    base_info = {
        "serialize-version": SERIALIZATION_VERSION,
        "ishtar-version": get_version(),
        "domain": site.domain,
        "name": site.name,
        "date": datetime.datetime.now().isoformat(),
    }
    if info:
        base_info.update(info)
    return base_info


def archive_serialization(
    result,
    archive_dir=None,
    archive=False,
    return_empty_types=False,
    archive_name=None,
    info=None,
):
    """
    Serialize all types models to JSON
    Used for import and export scripts

    :param result: serialization results
    :param archive_dir: directory inside the archive (default None)
    :param return_empty_types: if True instead of serialization return empty
    types (default False)
    :param archive: if True return a zip file containing all the file serialized
    (default False)
    :param archive_name: path to the archive if not provided a new archive is
    created
    :return: string containing the json serialization of types unless
    return_empty_types or archive is set to True
    """
    if archive and return_empty_types:
        raise ValueError("archive and return_empty_types are incompatible")
    if return_empty_types:
        return [k for k in result if not result[k]]
    if not archive:
        return result
    archive_created = False
    if not archive_name:
        archive_created = True
        tmpdir = tempfile.mkdtemp(prefix="ishtarexport-") + os.sep
        archive_name = tmpdir + "ishtar-{}.zip".format(
            datetime.date.today().strftime("%Y-%m-%d")
        )
    if not archive_name.endswith(".zip"):
        archive_name += ".zip"
    mode = "w" if archive_created else "a"
    with tempfile.TemporaryDirectory() as tmpdirname:
        if archive_dir:
            os.mkdir(tmpdirname + os.sep + archive_dir)

        with ZipFile(archive_name, mode) as current_zip:
            if archive_created:
                base_filename = "info.json"
                filename = tmpdirname + os.sep + base_filename
                with open(filename, "w") as json_file:
                    json_file.write(json.dumps(serialization_info(info=info), indent=2))
                current_zip.write(filename, arcname=base_filename)

            for dir_name, model_name in result:
                base_filename = model_name + ".json"
                filename = tmpdirname + os.sep + base_filename
                with open(filename, "w") as json_file:
                    json_file.write(result[(dir_name, model_name)])
                arcname = base_filename
                if dir_name:
                    arcname = dir_name + os.sep + base_filename
                current_zip.write(filename, arcname=arcname)
    return archive_name


GENERIC_QUERYSET_FILTER = {
    "JsonDataSection": {"JsonDataField": "json_data_field__pk__in"},
    "Regexp": {"ImporterType": "columns__importer_type__pk__in"},
    "ImporterModel": {
        "ImporterType": [
            "importer_type_associated__pk__in",
            "importer_type_created__pk__in",
        ]
    },
    "ValueFormater": {"ImporterType": "columns__importer_type__pk__in"},
    "ImporterColumn": {"ImporterType": "importer_type__pk__in"},
    "ImporterDefault": {"ImporterType": "importer_type__pk__in"},
    "ImportTarget": {"ImporterType": "column__importer_type__pk__in"},
    "FormaterType": {"ImporterType": "targets__column__importer_type__pk__in"},
    "ImporterDefaultValues": {"ImporterType": "default_target__importer_type__pk__in"},
    "ImporterDuplicateField": {"ImporterType": "column__importer_type__pk__in"},
}


def generic_get_results(
    model_list, dirname, no_geo=True, result_queryset=None, serialization_include=None
):
    result = OrderedDict()
    for model in model_list:
        base_model_name = model.__name__
        model_name = str(model.__module__).split(".")[0] + "__" + base_model_name
        base_q = model.objects
        if result_queryset:
            if result_queryset and base_model_name in result_queryset:
                base_q = result_queryset[base_model_name]
            elif base_model_name in GENERIC_QUERYSET_FILTER:
                alt_filter = GENERIC_QUERYSET_FILTER[base_model_name]
                for k in alt_filter:
                    if k in result_queryset:
                        terms = alt_filter[k]
                        if not isinstance(terms, (list, tuple)):
                            terms = [terms]
                        ids = [r["pk"] for r in result_queryset[k].values("pk").all()]
                        q = None
                        for term in terms:
                            if not q:
                                q = Q(**{term: ids})
                            else:
                                q |= Q(**{term: ids})
                        base_q = base_q.filter(q)
                        break
        q = base_q
        recursion = None
        if hasattr(model, "parent"):
            recursion = "parent"
        elif hasattr(model, "inverse_relation"):
            recursion = "inverse_relation"
        elif hasattr(model, "children") and model.children.field.model == model:
            recursion = "children__id"
        elif hasattr(model, "child"):
            recursion = "child_id"
        if recursion:
            q = q.filter(**{recursion + "__isnull": True})

        key = (dirname, model_name)
        result[key] = serialize(
            "json",
            q.distinct().all(),
            indent=2,
            use_natural_foreign_keys=True,
            use_natural_primary_keys=True,
        )

        if recursion:
            serialized = [item["id"] for item in q.values("id").all()]
            recursion_in = recursion
            if not recursion.endswith("_id"):
                recursion_in += "_id"
            recursion_in += "__in"
            q = base_q.filter(**{recursion_in: serialized}).exclude(id__in=serialized)
            while q.count():
                v = serialize(
                    "json",
                    q.all(),
                    indent=2,
                    use_natural_foreign_keys=True,
                    use_natural_primary_keys=True,
                )
                new_result = json.loads(result[key])
                new_result += json.loads(v)
                result[key] = json.dumps(new_result, indent=2)
                serialized += [item["id"] for item in q.values("id").all()]
                q = base_q.filter(**{recursion_in: serialized}).exclude(
                    id__in=serialized
                )
            # managed circular
            q = base_q.exclude(id__in=serialized)
            if q.count():
                v = serialize(
                    "json",
                    q.all(),
                    indent=2,
                    use_natural_foreign_keys=True,
                    use_natural_primary_keys=True,
                )
                result_to_add = json.loads(v)
                result_cleaned = deepcopy(result_to_add)
                for res in result_cleaned:  # first add with no recursion
                    res["fields"][recursion] = None
                new_result = json.loads(result[key])
                new_result += result_cleaned
                new_result += result_to_add
                result[key] = json.dumps(new_result, indent=2)

        excluded_fields = [
            "history_modifier",
            "history_creator",
            "imports",
            "locked",
            "lock_user",
        ]
        if hasattr(model, "SERIALIZATION_EXCLUDE"):
            excluded_fields += list(model.SERIALIZATION_EXCLUDE)
        if no_geo:
            excluded_fields += ["center", "limit"] + [
                field.name for field in models.GeoItem._meta.get_fields()
            ]
        if serialization_include and model.__name__ in serialization_include:
            for k in serialization_include[model.__name__]:
                if k in excluded_fields:
                    excluded_fields.pop(excluded_fields.index(k))
        if excluded_fields:
            new_result = json.loads(result[key])
            for idx in range(len(new_result)):
                for excluded_field in excluded_fields:
                    if excluded_field in new_result[idx]["fields"]:
                        new_result[idx]["fields"].pop(excluded_field)
            result[key] = json.dumps(new_result, indent=2)
    return result


def generic_archive_files(model_list, archive_name=None, result_queryset=None):
    if not result_queryset:
        result_queryset = {}
    result = []
    for model in model_list:
        if model.__name__ in result_queryset.keys():
            query = result_queryset[model.__name__]
        else:
            query = model.objects
        if hasattr(model, "SERIALIZATION_FILES"):
            for item in query.all():
                for attr in model.SERIALIZATION_FILES:
                    media = getattr(item, attr)
                    try:
                        result.append((media.path, media.name))
                    except ValueError:
                        pass
    archive_created = False
    if not archive_name:
        archive_created = True
        tmpdir = tempfile.mkdtemp(prefix="ishtarexport-") + os.sep
        archive_name = tmpdir + "media.zip"
    if not archive_name.endswith(".zip"):
        archive_name += ".zip"
    mode = "w" if archive_created else "a"
    with ZipFile(archive_name, mode) as current_zip:
        for media_path, name in result:
            try:
                current_zip.write(media_path, arcname=name)
            except OSError:
                pass
    return archive_name