def crawl_source(slug, ignore_tags=False):
Source.sync()
source = Source.by_slug(slug)
if source is None:
raise ValueError("Invalid source: %r" % slug)
source.crawler_instance.ignore_tags = ignore_tags
source.crawler_instance.crawl()
def authz_sources(action):
if action == 'read' and request.authz_sources.get('read') is None:
request.authz_sources['read'] = Source.list_user_slugs(current_user)
if action == 'write' and request.authz_sources.get('write') is None:
request.authz_sources['write'] = Source.list_user_slugs(
current_user, include_public=False) # noqa
return request.authz_sources[action] or []
def authz_sources(action):
if action == 'read' and request.authz_sources.get('read') is None:
request.authz_sources['read'] = Source.list_user_slugs(current_user)
if action == 'write' and request.authz_sources.get('write') is None:
request.authz_sources['write'] = Source.list_user_slugs(current_user,
include_public=False) # noqa
return request.authz_sources[action] or []
def crawl_source(slug, ignore_tags=False):
logging.debug('crawl source -- celery task going')
Source.sync()
source = Source.by_slug(slug)
if source is None:
raise ValueError("Invalid source: %r" % slug)
source.crawler_instance.ignore_tags = ignore_tags
source.crawler_instance.crawl()
def crawl_source(slug, ignore_tags=False):
logging.debug('crawl source -- celery task going')
Source.sync()
source = Source.by_slug(slug)
if source is None:
raise ValueError("Invalid source: %r" % slug)
source.crawler_instance.ignore_tags = ignore_tags
source.crawler_instance.crawl()
def analyze(foreign_id=None):
"""Re-analyze documents in the given source (or throughout)."""
if foreign_id:
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
analyze_source.delay(source.id)
else:
for source in Source.all():
analyze_source.delay(source.id)
def analyze(foreign_id=None):
"""Re-analyze documents in the given source (or throughout)."""
if foreign_id:
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
analyze_source.delay(source.id)
else:
for source in Source.all():
analyze_source.delay(source.id)
def get_results(query, limit):
sources = {}
for i, row in enumerate(scan_iter(query)):
if i >= limit:
return
data = {
'file_url': url_for('documents_api.file',
document_id=row.get('_id'))
}
for name, value in row.get('_source').items():
if name == 'source_id':
if value not in sources:
source = Source.by_id(value)
if source is None:
sources[value] = '[Deleted source %s]' % value
else:
sources[value] = source.label
value = sources[value]
name = 'source'
if name not in FIELDS:
continue
if isinstance(value, (list, tuple, set)):
value = ', '.join(value)
data[name] = value
yield data
def format_results(query):
sources = {}
entities = {}
results = []
for row in raw_iter(query):
src = row.get('_source')
data = {}
for name, value in src.items():
if isinstance(value, dict) or name in SKIP_FIELDS:
continue
if name == 'entities':
load_ids = []
for entity_id in value:
if entity_id not in entities:
load_ids.append(entity_id)
if len(load_ids):
for id, ent in Entity.by_id_set(load_ids).items():
entities[id] = ent.name
value = ', '.join([entities.get(e) for e in value
if entities.get(e) is not None])
if isinstance(value, (list, tuple, set)):
value = ', '.join(value)
if name == 'source_id':
# WARNING: don't to one query per row
if value not in sources:
source = Source.by_id(value)
if source is None:
sources[value] = '[Deleted source %s]' % value
else:
sources[value] = source.label
value = sources[value]
data[name] = value
results.append(data)
return results
def update(slug):
authz.require(authz.source_write(slug))
source = obj_or_404(Source.by_slug(slug))
source.update(request_data(), current_user)
db.session.add(source)
db.session.commit()
return view(slug)
def crawl(slug):
logging.debug('starting a crawl of %s' % slug)
authz.require(authz.source_write(slug))
source = obj_or_404(Source.by_slug(slug))
crawl_source.delay(source.slug)
logging.debug('started crawl')
return jsonify({'status': 'ok'})
def alert_query(alert):
"""Execute the query and return a set of results."""
q = text_query(alert.query_text)
q = authz_sources_filter(q)
if alert.entity_id:
q = filter_query(q, [('entities.uuid', alert.entity_id)], OR_FIELDS)
if alert.notified_at:
q = add_filter(q, {"range": {"created_at": {"gt": alert.notified_at}}})
q = {'query': q, 'size': 150}
result, hits, output = execute_basic(TYPE_DOCUMENT, q)
sub_queries = []
sources = {}
for doc in hits.get('hits', []):
document = doc.get('_source')
document['id'] = int(doc.get('_id'))
source_id = document['source_id']
if source_id not in sources:
sources[source_id] = Source.by_id(source_id)
if sources[source_id] is None:
continue
document['source'] = sources[source_id]
document['records'] = {'results': [], 'total': 0}
sq = records_query(document['id'], alert.to_query(), size=1)
if sq is not None:
sub_queries.append(json.dumps({}))
sub_queries.append(json.dumps(sq))
output['results'].append(document)
run_sub_queries(output, sub_queries)
return output
def crawl(self, directory=None, source=None, meta={}):
source = source or directory
source = Source.create({
'foreign_id': 'directory:%s' % slugify(source),
'label': source
})
db.session.commit()
source_id = source.id
if os.path.isfile(directory):
self.crawl_file(source_id, directory, meta)
directory = directory or os.getcwd()
directory = directory.encode('utf-8')
for (dirname, dirs, files) in os.walk(directory):
dirparts = [
d for d in dirname.split(os.path.sep) if d in SKIP_DIRECTORIES
]
if len(dirparts):
continue
log.info("Descending: %r", dirname)
for file_name in files:
dirname = string_value(dirname)
file_name = string_value(file_name)
if file_name in SKIP_FILES:
continue
file_path = os.path.join(dirname, file_name)
self.crawl_file(source_id, file_path, meta)
def update(id):
authz.require(authz.source_write(id))
source = obj_or_404(Source.by_id(id))
source.update(request_data())
db.session.add(source)
db.session.commit()
return view(id)
def update(id):
authz.require(authz.source_write(id))
source = obj_or_404(Source.by_id(id))
source.update(request_data())
db.session.add(source)
db.session.commit()
return view(id)
def get_results(query, limit):
sources = {}
for i, row in enumerate(scan_iter(query)):
if i >= limit:
return
data = {
'file_url': url_for('documents_api.file',
document_id=row.get('_id'))
}
for name, value in row.get('_source').items():
if name == 'source_id':
if value not in sources:
source = Source.by_id(value)
if source is None:
sources[value] = '[Deleted source %s]' % value
else:
sources[value] = source.label
value = sources[value]
name = 'source'
if name not in FIELDS:
continue
if isinstance(value, (list, tuple, set)):
value = ', '.join(value)
data[name] = value
yield data
def update(slug):
authz.require(authz.source_write(slug))
source = obj_or_404(Source.by_slug(slug))
source.update(request_data(), current_user)
db.session.add(source)
db.session.commit()
return view(slug)
def test_delete_source(self):
source = Source.by_id(1000)
res = self.client.get('/api/1/query?q="mention fruit"')
assert res.json['total'] == 1, res.json
delete_source(source.id)
optimize_search()
res = self.client.get('/api/1/query?q="mention fruit"')
assert res.json['total'] == 0, res.json
def test_delete_source(self):
source = Source.by_id(1000)
res = self.client.get('/api/1/query?q="mention fruit"')
assert res.json['total'] == 1, res.json
delete_source(source.id)
optimize_search()
res = self.client.get('/api/1/query?q="mention fruit"')
assert res.json['total'] == 0, res.json
def source(self):
if not hasattr(self, '_source'):
self._source = Source.create({
'foreign_id': self.SOURCE_ID,
'label': self.SOURCE_LABEL or self.SOURCE_ID
})
db.session.commit()
return self._source
def crawl(self):
for base_url in SITES:
print 'Working on base_url: {}'.format(base_url)
self.attributes = SITES[base_url]
self.label = self.attributes['label']
Source.create({'label': self.label, 'foreign_id': 'blacklight'})
db.session.commit()
self.failed_articles = 0
page_count = self.get_page_count(base_url)
print "Pages: {}".format(page_count)
page_number = 1
while (page_number <= page_count):
if self.failed_articles >= FAILED_LIMIT:
log.warning('Failure limit reach: {}'.format(FAILED_LIMIT))
break
self.crawl_page(base_url, page_number, page_count)
page_number += 1
def flush(foreign_id):
"""Reset the crawler state for a given source specification."""
from aleph.index import delete_source
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
delete_source(source.id)
source.delete()
db.session.commit()
def source(self):
if not hasattr(self, '_source'):
self._source = Source.create({
'foreign_id': self.SOURCE_ID,
'label': self.SOURCE_LABEL or self.SOURCE_ID
})
db.session.commit()
db.session.add(self._source)
return self._source
def flush(foreign_id):
"""Reset the crawler state for a given source specification."""
from aleph.index import delete_source
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
delete_source(source.id)
source.delete()
db.session.commit()
def crawl_source(self, engine, foreign_id, data):
source = Source.create({
'foreign_id': foreign_id,
'label': data.get('label')
})
db.session.commit()
meta_base = data.get('meta', {})
for name, query in data.get('queries', {}).items():
self.crawl_query(engine, source, meta_base, name, query)
def crawl_source(self, engine, foreign_id, data):
source = Source.create({
'foreign_id': foreign_id,
'label': data.get('label')
})
db.session.commit()
meta_base = data.get('meta', {})
for name, query in data.get('queries', {}).items():
self.crawl_query(engine, source, meta_base, name, query)
def view(slug):
authz.require(authz.source_read(slug))
source = obj_or_404(Source.by_slug(slug))
etag_cache_keygen(source)
data = source.to_dict()
data['can_write'] = authz.source_write(slug)
if data['can_write']:
data['users'] = [u.id for u in source.users]
data['config'] = source.config
return jsonify(data)
def index():
sources = []
latest = set()
for source in Source.all_by_user(current_user):
data = source.to_dict()
data['can_write'] = authz.source_write(source.slug)
latest.add(data['updated_at'])
sources.append(data)
etag_cache_keygen(max(latest))
return jsonify({'results': sources, 'total': len(sources)})
def view(slug):
authz.require(authz.source_read(slug) and authz.is_admin())
source = obj_or_404(Source.by_slug(slug))
etag_cache_keygen(source)
data = source.to_dict()
data['can_write'] = authz.source_write(slug)
if data['can_write']:
data['users'] = [u.id for u in source.users]
data['config'] = source.config
return jsonify(data)
def index():
sources = []
latest = set()
for source in Source.all_by_user(current_user):
data = source.to_dict()
data['can_write'] = authz.source_write(source.slug)
latest.add(data['updated_at'])
sources.append(data)
if len(latest):
etag_cache_keygen(max(latest))
return jsonify({'results': sources, 'total': len(sources)})
def crawl(self):
for base_url in SITES:
print 'Working on base_url: {}'.format(base_url)
self.attributes = SITES[base_url]
self.label = self.attributes['label']
Source.create({
'label': self.label,
'foreign_id': 'blacklight'
})
db.session.commit()
self.failed_articles = 0
page_count = self.get_page_count(base_url)
print "Pages: {}".format(page_count)
page_number = 1
while (page_number <= page_count):
if self.failed_articles >= FAILED_LIMIT:
log.warning('Failure limit reach: {}'.format(FAILED_LIMIT))
break
self.crawl_page(base_url, page_number, page_count)
page_number += 1
def index(foreign_id=None):
"""Index documents in the given source (or throughout)."""
q = Document.all_ids()
if foreign_id:
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
q = q.filter(Document.source_id == source.id)
for doc_id, in q:
index_document.delay(doc_id)
if foreign_id is None:
reindex_entities()
def index(foreign_id=None):
"""Index documents in the given source (or throughout)."""
q = Document.all_ids()
if foreign_id:
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
q = q.filter(Document.source_id == source.id)
else:
delete_index()
init_search()
for doc_id, in q:
index_document.delay(doc_id)
def dc_projects():
slug = request.args.get('source')
authz.require(authz.source_read(slug))
source = obj_or_404(Source.by_slug(slug))
if not isinstance(source.crawler_instance, DocumentCloudCrawler):
return jsonify({'credentials': False})
username = request.args.get('username')
password = request.args.get('password')
projects = source.crawler_instance.get_projects(username, password)
if projects is False:
return jsonify({'credentials': False})
else:
return jsonify({'credentials': True, 'projects': projects})
def dc_projects():
slug = request.args.get('source')
authz.require(authz.source_read(slug))
source = obj_or_404(Source.by_slug(slug))
if not isinstance(source.crawler_instance, DocumentCloudCrawler):
return jsonify({'credentials': False})
username = request.args.get('username')
password = request.args.get('password')
projects = source.crawler_instance.get_projects(username, password)
if projects is False:
return jsonify({'credentials': False})
else:
return jsonify({'credentials': True,
'projects': projects})
def process_row(row, attributes):
src = row.get('_source')
data = {}
for name in attributes:
value = src.get(name)
for attr in src.get('attributes', []):
if attr.get('name') == name:
value = attr.get('value')
if name == 'entities':
objs = Entity.by_id_set([e.get('id') for e in value])
value = ', '.join([o.label for o in objs.values()])
if name == 'collection':
# WARNING: don't to one query per row
value = unicode(Source.by_slug(value) or value)
data[name] = value
return data
def process_row(row, attributes):
src = row.get('_source')
data = {}
for name in attributes:
value = src.get(name)
for attr in src.get('attributes', []):
if attr.get('name') == name:
value = attr.get('value')
if name == 'entities':
objs = Entity.by_id_set([e.get('id') for e in value])
value = ', '.join([o.label for o in objs.values()])
if name == 'collection':
# WARNING: don't to one query per row
value = unicode(Source.by_slug(value) or value)
data[name] = value
return data
def setUp(self):
super(SourcesApiTestCase, self).setUp()
self.source = Source()
self.source.foreign_id = "test"
self.source.label = "Test Collection"
self.source.category = "news"
db.session.add(self.source)
db.session.flush()
permission = Permission()
permission.role_id = Role.system(Role.SYSTEM_USER)
permission.read = True
permission.write = True
permission.resource_id = self.source.id
permission.resource_type = Permission.SOURCE
db.session.add(permission)
db.session.commit()
def convert_sources(facet):
output = {'values': []}
ids = [b.get('key') for b in facet.get('buckets', [])]
sources = Source.all_by_ids(ids).all()
for bucket in facet.get('buckets', []):
key = bucket.get('key')
for source in sources:
if source.id != key:
continue
output['values'].append({
'id': key,
'label': source.label,
'category': source.category,
'count': bucket.get('doc_count')
})
return output
def sources(action):
if not hasattr(request, 'auth_sources'):
request.auth_sources = {READ: set(), WRITE: set()}
if is_admin():
for source_id, in Source.all_ids():
request.auth_sources[READ].add(source_id)
request.auth_sources[WRITE].add(source_id)
else:
q = Permission.all()
q = q.filter(Permission.role_id.in_(request.auth_roles))
q = q.filter(Permission.resource_type == Permission.SOURCE)
for perm in q:
if perm.read:
request.auth_sources[READ].add(perm.resource_id)
if perm.write and request.logged_in:
request.auth_sources[WRITE].add(perm.resource_id)
return list(request.auth_sources.get(action, []))
def index(foreign_id=None, immediate=False):
"""Index documents in the given source (or throughout)."""
q = Document.all_ids()
if foreign_id:
source = Source.by_foreign_id(foreign_id)
if source is None:
raise ValueError("No such source: %r" % foreign_id)
q = q.filter(Document.source_id == source.id)
for doc_id, in q:
#import time; time.sleep(10) #let's not get banned
print('indexing %s' % doc_id)
if immediate: #bypass the queue
index_document(doc_id)
else:
index_document.delay(doc_id)
if foreign_id is None:
reindex_entities()
def sources(action):
if not hasattr(request, 'auth_sources'):
request.auth_sources = {READ: set(), WRITE: set()}
if is_admin():
for source_id, in Source.all_ids():
request.auth_sources[READ].add(source_id)
request.auth_sources[WRITE].add(source_id)
else:
q = Permission.all()
q = q.filter(Permission.role_id.in_(request.auth_roles))
q = q.filter(Permission.resource_type == Permission.SOURCE)
for perm in q:
if perm.read:
request.auth_sources[READ].add(perm.resource_id)
if perm.write and request.logged_in:
request.auth_sources[WRITE].add(perm.resource_id)
return list(request.auth_sources.get(action, []))
def alert_query(alert):
"""Execute the query and return a set of results."""
q = text_query(alert.query_text)
q = authz_sources_filter(q)
if alert.entity_id:
q = filter_query(q, [('entities.uuid', alert.entity_id)], OR_FIELDS)
if alert.notified_at:
q = add_filter(q, {
"range": {
"created_at": {
"gt": alert.notified_at
}
}
})
q = {
'query': q,
'size': 150
}
result, hits, output = execute_basic(TYPE_DOCUMENT, q)
sub_queries = []
sources = {}
for doc in hits.get('hits', []):
document = doc.get('_source')
document['id'] = int(doc.get('_id'))
source_id = document['source_id']
if source_id not in sources:
sources[source_id] = Source.by_id(source_id)
if sources[source_id] is None:
continue
document['source'] = sources[source_id]
document['records'] = {'results': [], 'total': 0}
sq = records_query(document['id'], alert.to_query(), size=1)
if sq is not None:
sub_queries.append(json.dumps({}))
sub_queries.append(json.dumps(sq))
output['results'].append(document)
run_sub_queries(output, sub_queries)
return output
def crawl(self, directory=None, source=None):
source = source or directory
source = Source.create({
'foreign_id': 'directory:%s' % slugify(source),
'label': source
})
if os.path.isfile(directory):
meta = self.metadata()
meta.file_name = directory
self.emit_file(source, meta, directory)
directory = directory or os.getcwd()
directory = directory.encode('utf-8')
for (dirname, dirs, files) in os.walk(directory):
log.info("Descending: %r", dirname)
for file_name in files:
if file_name in SKIP_FILES:
continue
file_path = os.path.join(dirname, file_name)
if not os.path.isfile(file_path):
continue
try:
meta = self.metadata()
if isinstance(file_name, six.text_type):
meta.source_path = file_path
else:
enc = chardet.detect(file_name)
enc = enc.get('encoding')
try:
meta.source_path = file_path.decode(enc)
except:
meta.source_path = file_path.decode('ascii', 'ignore')
self.emit_file(source, meta, file_path)
except Exception as ex:
log.exception(ex)
process.exception(process.INDEX, component=self.name,
source_location=directory,
source_id=source.id, exception=ex)
def crawl_item(self, item, source):
source_data = item.meta.get('source', {})
source_fk = source_data.pop('foreign_id', source)
if source_fk is None:
raise ValueError("No foreign_id for source given: %r" % item)
if source_fk not in self.sources:
label = source_data.get('label', source_fk)
self.sources[source_fk] = Source.create({
'foreign_id': source_fk,
'label': label
})
if source_data.get('public'):
Permission.grant_foreign(self.sources[source_fk],
Role.SYSTEM_GUEST, True, False)
db.session.commit()
log.info('Import: %r', item.identifier)
meta = self.normalize_metadata(item)
ingest_file(self.sources[source_fk].id,
meta,
item.data_path,
move=False)
def crawl(self, directory=None, source=None):
source = source or directory
source = Source.create({
'foreign_id': 'directory:%s' % slugify(source),
'label': source
})
if os.path.isfile(directory):
meta = self.metadata()
meta.file_name = directory
self.emit_file(source, meta, directory)
directory = directory or os.getcwd()
directory = directory.encode('utf-8')
for (dirname, dirs, files) in os.walk(directory):
log.info("Descending: %r", dirname)
for file_name in files:
if file_name in SKIP_FILES:
continue
file_path = os.path.join(dirname, file_name)
if not os.path.isfile(file_path):
continue
try:
meta = self.metadata()
if isinstance(file_name, six.text_type):
meta.source_path = file_path
else:
enc = chardet.detect(file_name)
enc = enc.get('encoding')
try:
meta.source_path = file_path.decode(enc)
except:
meta.source_path = file_path.decode('ascii', 'ignore')
self.emit_file(source, meta, file_path)
except Exception as ex:
log.exception(ex)
def sources():
"""List all sources."""
for source in Source.all():
print source.id, source.foreign_id, source.label
def index():
q = Source.all_by_ids(ids=authz.sources(authz.READ))
return jsonify(Pager(q))
def view(id):
authz.require(authz.source_read(id))
source = obj_or_404(Source.by_id(id))
return jsonify(source)
def process(id):
authz.require(authz.source_write(id))
source = obj_or_404(Source.by_id(id))
analyze_source.delay(source.id)
return jsonify({'status': 'ok'})
def process(id):
authz.require(authz.source_write(id))
source = obj_or_404(Source.by_id(id))
analyze_source.delay(source.id)
return jsonify({'status': 'ok'})
def view(id):
authz.require(authz.source_read(id))
source = obj_or_404(Source.by_id(id))
return jsonify(source)
def index():
q = Source.all_by_ids(ids=authz.sources(authz.READ))
return jsonify(Pager(q))
def upgrade():
log.info("Beginning database migration...")
migrate.upgrade()
log.info("Reconfiguring the search index...")
init_search()
Source.sync()
def create_source(self, **data):
if 'foreign_id' not in data:
data['foreign_id'] = self.name
return Source.create(data)
def process(slug):
authz.require(authz.source_write(slug))
source = obj_or_404(Source.by_slug(slug))
process_collection.delay(source.slug)
return jsonify({'status': 'ok'})
def crawl(slug):
authz.require(authz.source_write(slug))
source = obj_or_404(Source.by_slug(slug))
crawl_source.delay(source.slug)
return jsonify({'status': 'ok'})
def create():
authz.require(authz.logged_in())
src = Source.create(request_data(), current_user)
db.session.commit()
return view(src.slug)
