def suggest(dataset):
"""
Suggest API, emulates Google Refine API. See:
http://code.google.com/p/google-refine/wiki/SuggestApi
"""
try:
start = int(request.args.get('start', 0))
limit = int(request.args.get('limit', 20))
except:
raise BadRequest('Invalid result range!')
dataset = type_to_dataset(dataset)
query = request.args.get('prefix', '').strip()
results = prefix_search(query, dataset)[start:start+limit]
entities = Entity.id_map(dataset, map(lambda (c,v): v, results))
matches = []
for candidate, entity_id in results:
entity = entities[entity_id]
matches.append({
'name': entity.name,
'n:type': {
'id': '/' + dataset.name,
'name': dataset.label
},
'id': url_for('entity.view', dataset=dataset.name, entity=entity_id)
})
return jsonify({
"code" : "/api/status/ok",
"status" : "200 OK",
"prefix" : query,
"result" : matches
})
def match(dataset, alias, random=False):
dataset = Dataset.find(dataset)
authz.require(authz.dataset_edit(dataset))
alias = Alias.find(dataset, alias)
random = random or request.args.get('random')=='True'
choices = match_op(alias.name, dataset,
query=request.args.get('query'))
pager = Pager(choices, '.match',
dataset=dataset.name, alias=alias.id,
limit=10)
# HACK: Fetch only the entities on the selected page.
entities = Entity.id_map(dataset, map(lambda (c,e,s): e,
pager.query[pager.offset:pager.offset+pager.limit]))
for i, (c,e,s) in enumerate(pager.query):
if e in entities:
pager.query[i] = (c, entities.get(e), s)
html = render_template('alias/match.html',
dataset=dataset, alias=alias, choices=pager,
random=random)
choice = 'INVALID' if alias.is_invalid else alias.entity_id
if len(choices) and choice is None:
c, e, s = choices[0]
choice = 'INVALID' if s <= 50 else e.id
return htmlfill.render(html, force_defaults=False,
defaults={'choice': choice,
'name': alias.name,
'query': request.args.get('query', ''),
'random': random})
def view(dataset, entity):
dataset = Dataset.find(dataset)
entity = Entity.find(dataset, entity)
print entity.data
format = response_format()
if format == 'json':
return jsonify(entity)
query = request.args.get('query', '').strip().lower()
choices = match_op(entity.name, dataset)
choices = filter(lambda (c,e,s): e != entity.id, choices)
if len(query):
choices = filter(lambda (c,e,s): query in Entity.find(dataset,e).name.lower(),
choices)
# THIS is very inefficient - rather do this
# differently
pager = Pager(choices, '.view', dataset=dataset.name,
entity=entity.id, limit=10)
# HACK: Fetch only the entities on the selected page.
entities = Entity.id_map(dataset, map(lambda (c,v,s): v,
pager.query[pager.offset:pager.offset+pager.limit]))
for i, (c,e,s) in enumerate(pager.query):
if e in entities:
pager.query[i] = (c, entities.get(e), s)
return render_template('entity/view.html', dataset=dataset,
entity=entity, entities=pager, query=query)
def __iter__(self):
rp = db.engine.execute(self.lq)
rows = rp.fetchall()
ids = [r[0] for r in rows]
entities = Entity.id_map(ids)
for (id, score) in rows:
yield {'score': int(score), 'entity': entities.get(id)}
def reconcile_op(dataset, query):
try:
limit = max(1, min(100, int(query.get('limit'))))
except ValueError: limit = 5
except TypeError: limit = 5
filters = [(p.get('p'), p.get('v')) for p in query.get('properties', [])]
if dataset is None:
dataset = type_to_dataset(query.get('type', ''))
results = match(query.get('query', ''), dataset)[:limit]
entities = Entity.id_map(dataset, map(lambda (c,e,s): e, results))
matches = []
skip = False
for (candidate, entity_id, score) in results:
entity = entities[entity_id]
for key, fv in filters:
if entity.data.get(key) != fv:
skip = True
if skip:
continue
id = url_for('entity.view', dataset=dataset.name, entity=entity.id)
uri = url_for('entity.view', dataset=dataset.name, entity=entity.id, _external=True)
matches.append({
'name': entity.name,
'score': score,
'type': [{
'id': '/' + dataset.name,
'name': dataset.label
}],
'id': id,
'uri': uri,
'match': score==100
})
return {
'result': matches,
'num': len(results)
}
