def _text_search(self, q, fields=None):
if fields is None:
fields = [
x for x in META.keys() if META[x].get("text_search", False)
]
w = re.compile(r'\W+', re.U)
words = w.split(q)
matches = []
suffix = ""
if len(fields) == 1 and fields[0] == "subject":
suffix = " IN BOOLEAN MODE"
for k in fields:
matches.append(
"MATCH(`%s`) AGAINST ('%s'%s)" % (k, " ".join(words), suffix)
)
m = "+".join(matches)
q = "SELECT r.id, (%s) AS score FROM savoirs_record AS r \
LEFT JOIN savoirs_record_listsets AS rl ON r.id = rl.record_id \
LEFT JOIN savoirs_listset AS l ON rl.listset_id = l.spec \
WHERE (%s) AND r.validated = 1 \
AND (l.validated = 1 OR l.validated IS NULL) \
GROUP BY r.id \
HAVING score > 0 ORDER BY score DESC" % (m, m)
from django.db import connection
cursor = connection.cursor()
cursor.execute(q)
rc = cursor.fetchall()
return rc
def decode(self, field, data):
"""
Décode la structure provenant de la BD.
"""
if field in META.keys() and META[field]['type'] == 'array':
return data.split(self.separator)
else:
return data
def encode(self, field, data):
"""
Encode la structure de donnée moissonnée pour la BD.
"""
if field in META.keys() and META[field]['type'] == 'array':
return self.separator.join(data)
else:
return data
def _load(self, id):
"""
Recupérer la structure de métadonnées pour un record selon un `id`.
"""
r = Record.all_objects.get(id=id)
meta = {'id': id}
for k in META.keys():
if hasattr(r, k):
v = getattr(r, k)
if v is not None:
meta[k] = self.encoder.decode(k, v)
return meta
def menage(self,):
"""
Applique sur tous les records, la fonction de corrections de string
sur les données moissonnées.
"""
for r in Record.all_objects.all():
for k in META.keys():
v = getattr(r, k)
if k not in ['uri', 'identifier']:
v = self.encoder.propre(v)
setattr(r, k, v)
r.save()
def _update_record(self, r, metadata):
for k in metadata.keys():
if hasattr(self, k):
method = getattr(self, k)
method(r, metadata[k])
else:
setattr(r, k, self.encoder.encode(k, metadata[k]))
r.last_checksum = hashlib.md5(str(metadata)).hexdigest()
r.last_update = datetime.datetime.today()
# stocke des chaînes de caractères propre en BD en provenance
# des données moissonnées
for k in META.keys():
v = getattr(r, k)
if k not in ['uri', 'identifier']:
v = self.encoder.propre(v)
setattr(r, k, v)
r.save()