def before_index(self, data_dict):
"""
customize data sent to solr
"""
bogus_date = datetime.datetime(1, 1, 1)
dataset_schema = scheming_get_dataset_schema(
data_dict.get('type', 'unknown')
)
if dataset_schema is None:
raise ValidationError(
'Found no schema for following dataset :\n{dump}'.format(
dump=json.dumps(data_dict, indent=2)
)
)
# iterate through dataset fields defined in schema
field_schema = dict()
for dataset_field in dataset_schema['dataset_fields']:
d = dataset_field
field_schema[d['field_name']] = d
index_data_dict = {}
authors = []
# drop extras fields
for dict_key in data_dict:
if not dict_key.startswith('extras_'):
index_data_dict[dict_key] = data_dict[dict_key]
# iterate through validated data_dict fields and modify as needed
validated_data_dict = json.loads(data_dict['validated_data_dict'])
for item in validated_data_dict.keys():
value = validated_data_dict[item]
if not value and item in index_data_dict:
index_data_dict.pop(item)
continue
fs = field_schema.get(item, None)
# ignore all fields not currently in the schema
if not fs:
continue
field_type = fs.get('schema_field_type', 'string')
multivalued = fs.get('schema_multivalued', False)
if field_type == 'fluent':
for key in value.keys():
label = u'{item}_{key}'.format(
item=item,
key=key
)
index_data_dict[label] = value[key]
# for code type, the en/fr labels need to be looked up
# and sent to Solr
elif field_type == 'code':
lookup_type = fs.get('lookup', '')
if lookup_type == 'codeset':
lookup = fs.get('codeset_type', '')
elif lookup_type == 'preset':
lookup = fs.get('preset', '')[4:]
else:
lookup = fs.get('lookup', '')
if lookup and value:
label_en = u'{item}_desc_en'.format(
item=item
)
label_fr = u'{item}_desc_fr'.format(
item=item
)
if multivalued:
desc_en = []
desc_fr = []
for v in value:
if not v:
continue
desc = lookup_label(lookup, v, lookup_type)
desc_en.append(desc[u'en'])
desc_fr.append(desc[u'fr'])
index_data_dict[str(item)] = value
index_data_dict[label_en] = desc_en
index_data_dict[label_fr] = desc_fr
else:
desc = lookup_label(lookup, value, lookup_type)
index_data_dict[label_en] = desc[u'en']
index_data_dict[label_fr] = desc[u'fr']
elif field_type == 'date':
if value:
try:
date = parse(value, default=bogus_date)
if date != bogus_date:
index_data_dict[item] = date.isoformat() + 'Z'
except ValueError:
continue
elif item.endswith('_authors'):
index_data_dict[str(item)] = value
authors.extend(value)
else: # all other field types
if multivalued:
index_data_dict[str(item)] = value
else:
index_data_dict[str(item)] = value
if authors:
index_data_dict['authors'] = authors
index_data_dict['authors_initials'] = list(
set(
[strip_accents(i[0]).upper() for i in authors]
)
)
return index_data_dict
def before_index(self, data_dict):
"""
customize data sent to solr
:param data_dict:
:type data_dict dict
:returns dict
"""
dataset_schema = scheming_get_dataset_schema(data_dict.get('type'))
if dataset_schema is None:
raise ValidationError((_(
'Found no schema for following datasets:\n{dump}'.format(
dump=json.dumps(data_dict, indent=2, sort_keys=True)
)
),))
field_schema = dict(
(s['field_name'], s) for s in dataset_schema['dataset_fields']
)
index_data_dict = data_dict.copy()
for k in data_dict:
if k.startswith(u'extras_'):
index_data_dict.pop(k, None)
authors = []
default_date = datetime(1, 1, 1, 8, 30, 0, 0)
validated_data_dict = json.loads(data_dict['validated_data_dict'])
name = validated_data_dict.get(u'name')
# append dguids from the datastore
if validated_data_dict.get(u'product_id_new'):
index_data_dict[u'dguid_codes'] = []
for dguid_pkg_id in geo.get_geodescriptors_for_package(
validated_data_dict[u'product_id_new']):
index_data_dict[u'dguid_codes'].append(
helpers.get_dguid_from_pkg_id(dguid_pkg_id))
# strip the vintages from dguids to get geodescriptors
index_data_dict[u'geodescriptor_codes'] = \
[g[4:] if is_dguid(g) else g
for g in index_data_dict[u'dguid_codes'] if g]
for item, value in validated_data_dict.iteritems():
fs = field_schema.get(item)
# Do not index any field that is not currently in the schema.
if not fs:
continue
field_type = fs.get('schema_field_type', 'string')
# TODO: we're not using the multivalued schema field. Drop it?
multivalued = fs.get('schema_multivalued', False)
# Legacy issues numbers are non-numeric, which is problematic
# for sorting and external tools. We can't just use a Solr
# <copyTo> directive, as it'll fail entirely on a bad value.
if name == 'issue_number':
if value.isdigit():
index_data_dict['issue_number_int'] = int(value)
# Fluent (multilingual) fields are really dictionaries, where
# each key is the ISO language code, and the value the translated
# text. We need to unpack these into individual solr fields
# for per-language search.
if field_type == 'fluent':
if isinstance(value, dict):
index_data_dict.update(
(u'{0}_{1}'.format(item, k), v)
for k, v in value.iteritems()
)
else:
raise ValidationError((_(
'{name}: Expecting a fluent dict for {item}, '
'instead got {value!r}'.format(
name=name,
item=item,
value=value
)
), ))
# Numeric foreign keys that need to be looked up to retrieve
# their multilingual labels for searching.
elif field_type == u'code':
index_data_dict[unicode(item)] = value
# These codes can refer to a codeset (a dataset of type
# 'codeset' with a particular key), a preset (a hardcoded
# value in a Scheming schema), or another dataset (lookup).
lookup_type = fs.get(u'lookup', '')
if lookup_type == u'codeset':
lookup = fs.get(u'codeset_type', '')
elif lookup_type == u'preset':
lookup = fs.get(u'preset', '')[4:]
else:
lookup = fs.get(u'lookup', '')
if not lookup:
raise ValidationError((_(
'{name}: unable to determine lookup '
'for {item}'.format(
name=name,
item=item
)
), ))
if isinstance(value, list):
for value_to_lookup in value:
if not value_to_lookup:
continue
desc = lookup_label(
lookup,
value_to_lookup,
lookup_type
)
for k, v in desc.iteritems():
if v and not k == u'found':
n = u'{item}_desc_{key}'.format(
item=item,
key=k
)
index_data_dict.update(
{n: index_data_dict.get(n, []) + [v]}
)
else:
desc = lookup_label(lookup, value, lookup_type)
index_data_dict.update((
u'{item}_desc_{key}'.format(
item=item,
key=k
), v)
for k, v in desc.iteritems() if v and not k == u'found'
)
if item == u'geodescriptor_codes':
index_data_dict[u'dguid_codes'] = \
list(index_data_dict[u'geodescriptor_codes'])
elif field_type == 'date':
try:
date = parse(value, default=default_date)
index_data_dict[unicode(item)] = unicode(
date.isoformat()[:19] + u'Z'
)
except ValueError:
continue
elif item.endswith('_authors'):
index_data_dict[unicode(item)] = value
authors.extend(value)
else:
index_data_dict[unicode(item)] = value
if authors:
index_data_dict['authors'] = authors
index_data_dict['authors_initials'] = list(
set(
[strip_accents(i[0]).upper() for i in authors]
)
)
return index_data_dict
