class Work(BaseDoc):
title = Text(analyzer=plain_ascii, fields={'keyword': Keyword()})
sort_title = Keyword(index=False)
uuid = Keyword(store=True)
medium = Text(fields={'keyword': Keyword()})
series = Text(fields={'keyword': Keyword()})
series_position = Keyword()
alt_titles = Text(fields={'keyword': Keyword()})
issued_date = DateRange(format='date_optional_time')
created_date = DateRange(format='date_optional_time')
is_government_document = Boolean(multi=False)
instances = Nested(Instance)
identifiers = Nested(Identifier)
subjects = Nested(Subject)
agents = Nested(Agent)
languages = Nested(Language)
@classmethod
def getFields(cls):
return [
'uuid', 'title', 'sort_title', 'sub_title', 'medium', 'series',
'series_position', 'date_modified', 'date_updated'
]
def __dir__(self):
return ['identifiers', 'subjects', 'agents', 'languages']
def __repr__(self):
return '<ESWork(title={}, uuid={})>'.format(self.title, self.uuid)
class Index:
name = os.environ.get('ES_INDEX', None)
class Work(BaseDoc):
title = Text(fields={'keyword': Keyword()})
sort_title = Keyword(index=False)
uuid = Keyword(store=True)
medium = Text(fields={'keyword': Keyword()})
series = Text(fields={'keyword': Keyword()})
series_position = Keyword()
issued = DateRange(format='date_optional_time')
issued_display = Keyword(index=False)
created = DateRange(format='date_optional_time')
created_display = Keyword(index=False)
alt_titles = Text(fields={'keyword': Keyword()})
summary = Text()
identifiers = Nested(Identifier)
subjects = Nested(Subject)
agents = Nested(Agent)
measurements = Nested(Measurement)
links = Nested(Link)
instances = Nested(Instance)
language = Nested(Language)
rights = Nested(Rights)
@classmethod
def getFields(cls):
return [
'uuid', 'title', 'sort_title', 'sub_title', 'language', 'medium',
'series', 'series_position', 'date_modified', 'date_updated'
]
class Index:
name = os.environ.get('ES_INDEX', None)
class Rights(BaseInner):
source = Keyword()
license = Keyword()
rights_statement = Text(fields={'keyword': Keyword()})
rights_reason = Text(fields={'keyword': Keyword()})
copyright_date = DateRange()
copyright_date_display = Keyword(index=False)
determination_date = DateRange()
determination_date_display = Keyword(index=False)
@classmethod
def getFields(cls):
return ['source', 'license', 'rights_statement', 'rights_reason']
class Instance(BaseInner):
title = Text(fields={'keyword': Keyword()})
sub_title = Text(fields={'keyword': Keyword()})
alt_titles = Text(fields={'keyword': Keyword()})
pub_place = Text(fields={'keyword': Keyword()})
pub_date = DateRange(format='date_optional_time')
pub_date_display = Keyword(index=False)
pub_date_sort = Date()
pub_date_sort_desc = Date()
edition = Text(fields={'keyword': Keyword()})
edition_statement = Text(fields={'keyword': Keyword()})
table_of_contents = Text()
volume = Text(fields={'keyword': Keyword()})
extent = Text()
summary = Text()
items = Nested(Item)
agents = Nested(Agent)
measurements = Nested(Measurement)
identifiers = Nested(Identifier)
links = Nested(Link)
language = Nested(Language)
rights = Nested(Rights)
covers = Nested(Link)
@classmethod
def getFields(cls):
return [
'title', 'sub_title', 'pub_place', 'edition', 'edition_statement',
'table_of_contents', 'langauge', 'extent', 'volume', 'summary'
]
class Instance(BaseInner):
title = Text(analyzer=plain_ascii, fields={'keyword': Keyword()})
sub_title = Text(analyzer=plain_ascii, fields={'keyword': Keyword()})
alt_titles = Text(fields={'keyword': Keyword()})
pub_place = Text(fields={'keyword': Keyword()})
pub_date = DateRange(format='date_optional_time')
edition = Text(fields={'keyword': Keyword()})
edition_statement = Text(fields={'keyword': Keyword()})
table_of_contents = Text()
volume = Text(fields={'keyword': Keyword()})
extent = Text()
summary = Text()
formats = Keyword()
instance_id = Integer()
edition_id = Integer()
agents = Nested(Agent)
identifiers = Nested(Identifier)
rights = Nested(Rights)
languages = Nested(Language)
@classmethod
def getFields(cls):
return [
'id', 'edition_id', 'title', 'sub_title', 'pub_place', 'edition',
'edition_statement', 'table_of_contents', 'extent', 'summary'
]
def __dir__(self):
return ['agents', 'identifiers', 'rights', 'languages', 'formats']
def cleanRels(self):
for rel in dir(self):
if isinstance(getattr(self, rel), set):
setattr(self, rel, list(getattr(self, rel)))
class Agent(BaseInner):
name = Text(fields={'keyword': Keyword()})
sort_name = Keyword(index=False)
aliases = Text(fields={'keyword': Keyword()})
lcnaf = Keyword()
viaf = Keyword()
birth_date = DateRange()
birth_display = Keyword(index=False)
death_date = DateRange()
death_display = Keyword(index=False)
biography = Text()
roles = Keyword()
links = Nested(Link)
@classmethod
def getFields(cls):
return ['name', 'sort_name', 'lcnaf', 'viaf', 'biography']
class _AggregateReportDoc(Document):
class Index:
name = "dmarc_aggregate"
xml_schema = Text()
org_name = Text()
org_email = Text()
org_extra_contact_info = Text()
report_id = Text()
date_range = DateRange()
errors = Text()
published_policy = Object(_PublishedPolicy)
source_ip_address = Ip()
source_country = Text()
source_reverse_dns = Text()
source_Base_domain = Text()
message_count = Integer
disposition = Text()
dkim_aligned = Boolean()
spf_aligned = Boolean()
passed_dmarc = Boolean()
policy_overrides = Nested(_PolicyOverride)
header_from = Text()
envelope_from = Text()
envelope_to = Text()
dkim_results = Nested(_DKIMResult)
spf_results = Nested(_SPFResult)
def add_policy_override(self, type_, comment):
self.policy_overrides.append(_PolicyOverride(type=type_,
comment=comment))
def add_dkim_result(self, domain, selector, result):
self.dkim_results.append(_DKIMResult(domain=domain,
selector=selector,
result=result))
def add_spf_result(self, domain, scope, result):
self.spf_results.append(_SPFResult(domain=domain,
scope=scope,
result=result))
def save(self, ** kwargs):
self.passed_dmarc = False
self.passed_dmarc = self.spf_aligned or self.dkim_aligned
return super().save(** kwargs)
class PartSpec(InnerDoc):
key = Keyword()
# Formats: https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-date-format.html
date_range = DateRange(format="yyyy-MM-dd HH:mm:ss.SSS||yyy-MM-dd'T'HH:mm:ss.SSS||yyyy-MM-dd HH:mm:ss"
"||yyy-MM-dd'T'HH:mm:ss||epoch_millis||epoch_second")
geo_range = DoubleRange()
class Date(BaseInner):
date = DateRange(format='date_optional_time')
date_type = Keyword()
class Meta:
using = es_connection
dynamic_templates = MetaField([
# Can't get this to work
# https://stackoverflow.com/questions/49740033/elasticsearch-6-copy-to-with-dynamic-index-mappings
# {
# "strings":
# {
# "match_mapping_type": "string",
# "mapping": {
# "type": "text",
# "copy_all": "anytext"
# }
# }
# },
{
"metadata_json":
{
"path_match": "metadata_json.identifier.identifier",
"match_mapping_type": "string",
"mapping": Keyword("not_analyzed")
}
},
{
"metadata_json":
{
"path_match": "metadata_json.titles.title",
"match_mapping_type": "string",
"mapping": Text(fields={'raw': Keyword()}),
}
},
{
"metadata_json":
{
"path_match": "metadata_json.subjects.subject",
"match_mapping_type": "string",
"mapping": Text(fields={'raw': Keyword()}),
}
},
{
"metadata_json":
{
"path_match": "metadata_json.creators.creatorName",
"match_mapping_type": "string",
"mapping": Text(fields={'raw': Keyword()}),
}
},
{
"metadata_json":
{
"path_match": "metadata_json.publicationYear",
"match_mapping_type": "string",
"mapping": Integer()
}
},
{
"metadata_json":
{
"path_match": "metadata_json.publisher",
"match_mapping_type": "string",
"mapping": Text(fields={'raw': Keyword()}),
}
},
{
"metadata_json":
{
"path_match": "metadata_json.dates.date",
"match_mapping_type": "string",
"mapping": DateRange()
}
},
{
"metadata_json":
{
"path_match": "metadata_json.geoLocations.geoLocationBox",
"match_mapping_type": "string",
"mapping": GeoShape()
}
},
{
"metadata_json":
{
"path_match": "metadata_json.geoLocations.geoLocationPoint",
"match_mapping_type": "string",
"mapping": GeoPoint()
}
}
])