def _create_test_models(n=100):
_add_test_models_to_session('Tag', n, ['name'])
_add_test_models_to_session('Speaker', n, ['first_name', 'last_name', 'dialect'])
_add_test_models_to_session('Source', n, ['author_first_name', 'author_last_name',
'title'])
_add_test_models_to_session('ElicitationMethod', n, ['name'])
_add_test_models_to_session('SyntacticCategory', n, ['name'])
_add_test_models_to_session('File', n, ['name'])
restricted_tag = h.generate_restricted_tag()
Session.add(restricted_tag)
Session.commit()
def _create_test_models(n=100):
_add_test_models_to_session('Tag', n, ['name'])
_add_test_models_to_session('Speaker', n,
['first_name', 'last_name', 'dialect'])
_add_test_models_to_session(
'Source', n, ['author_first_name', 'author_last_name', 'title'])
_add_test_models_to_session('ElicitationMethod', n, ['name'])
_add_test_models_to_session('SyntacticCategory', n, ['name'])
_add_test_models_to_session('File', n, ['name'])
restricted_tag = h.generate_restricted_tag()
Session.add(restricted_tag)
Session.commit()
def setup_app(command, conf, vars):
"""Commands to setup onlinelinguisticdatabase."""
config = load_environment(conf.global_conf, conf.local_conf)
log.info('Environment loaded.')
Base.metadata.create_all(bind=Session.bind)
filename = os.path.split(conf.filename)[-1] # e.g., production.ini, development.ini, test.ini, ...
# Create the ``store`` directory and those for file, analysis and corpora
# objects and their subdirectories. See ``lib.utils.py`` for details.
h.create_OLD_directories(config=config)
# ISO-639-3 Language data for the languages table
log.info("Retrieving ISO-639-3 languages data.")
languages = h.get_language_objects(filename, config)
# Get default users.
log.info("Creating a default administrator, contributor and viewer.")
administrator = h.generate_default_administrator(config_filename=filename)
contributor = h.generate_default_contributor(config_filename=filename)
viewer = h.generate_default_viewer(config_filename=filename)
# If we are running tests, make sure the test db contains only language data.
if filename == 'test.ini':
# Permanently drop any existing tables
Base.metadata.drop_all(bind=Session.bind, checkfirst=True)
log.info("Existing tables dropped.")
# Create the tables if they don't already exist
Base.metadata.create_all(bind=Session.bind, checkfirst=True)
log.info('Tables created.')
Session.add_all(languages + [administrator, contributor, viewer])
Session.commit()
# Not a test: add a bunch of nice defaults.
else:
# Create the _requests_tests.py script
requests_tests_path = os.path.join(config['pylons.paths']['root'], 'tests',
'scripts', '_requests_tests.py')
# This line is problematic in production apps because the
# _requests_tests.py file is not included in the build. So, I'm
# commenting it out by default.
# copyfile(requests_tests_path, '_requests_tests.py')
# Create the tables if they don't already exist
Base.metadata.create_all(bind=Session.bind, checkfirst=True)
log.info('Tables created.')
# Get default home & help pages.
log.info("Creating default home and help pages.")
homepage = h.generate_default_home_page()
helppage = h.generate_default_help_page()
# Get default application settings.
log.info("Generating default application settings.")
application_settings = h.generate_default_application_settings()
# Get default tags and categories
log.info("Creating some useful tags and categories.")
restricted_tag = h.generate_restricted_tag()
foreign_word_tag = h.generate_foreign_word_tag()
S = h.generate_s_syntactic_category()
N = h.generate_n_syntactic_category()
V = h.generate_v_syntactic_category()
# Initialize the database
log.info("Adding defaults.")
data = [administrator, contributor, viewer, homepage, helppage,
application_settings, restricted_tag, foreign_word_tag]
if config['add_language_data'] != '0':
data += languages
if config['empty_database'] == '0':
Session.add_all(data)
Session.commit()
log.info("OLD successfully set up.")
def test_aaa_initialize(self):
"""Initialize the database using pseudo-data generated from random lorem ipsum sentences.
These are located in ``onlinelinguisticdatabase/tests/data/corpora``.
The data contain morphologically analyzed sentences, their component
morphemes, and syntactic categories. The sentences have phrase
structure trees in bracket notation.
The test will try to load the lorem ipsum dataset from a MySQL/SQLite
dump file in ``onlinelinguisticdatabase/tests/data/corpora``. If the
dump file corresponding to ``loremipsum_path`` does not exist, it will
import the lorem ipsum data directly from the text files and create
the dump file so that future tests can run more speedily. The
``loremipsum100_path``, ``loremipsum1000_path``, ``loremipsum10000_path``
and ``loremipsum30000_path`` files are available and contain 100, 1000
and 10,000 sentences, respectively.
Setting the ``via_request`` variable to ``True`` will cause all of the
forms to be created via request, i.e., via
``self.app.post(url('forms))...``. This is much slower but may be
desirable since values for the morphological analysis attributes
will be generated.
.. note::
In order to run ``mysqldump`` with the MySQL user listed in
``test.ini``, that user must have permission to lock and update
tables::
mysql -u root -p<root_password>
grant lock tables, update on old_test.* to 'old'@'localhost';
.. warning::
Loading the .txt or .sql files with the ``via_request`` option set to
``True`` will take a very long time. This might be an argument for
separating the interface and logic components of the controllers so
that a "core" HTTP-less OLD application could be exposed. This
would facilitate the creation of models with system-generated data
and validation but without the HTTP overhead...
"""
########################################################################
# Configure lorem ipsum data set import
########################################################################
# Set ``loremipsum_path`` this to ``self.loremipsum100_path``,
# ``self.loremipsum1000_path`` or ``self.loremipsum10000_path``.
# WARNING: the larger ones will take a long time.
# Use the 10,000-sentence lorem ipsum dataset to ensure that
# very large corpora are handled correctly.
loremipsum_path = self.loremipsum100_path
# Set ``via_request`` to ``True`` to create all forms via HTTP requests.
via_request = True
self._add_SEARCH_to_web_test_valid_methods()
# Add an application settings so that morpheme references will work out right.
application_settings = h.generate_default_application_settings()
Session.add(application_settings)
Session.commit()
def create_model(line, categories, via_request=False):
"""Create a model (form or syncat) using the string in ``line``."""
model = 'Form'
elements = unicode(line).split('\t')
non_empty_elements = filter(None, elements)
try:
ol, mb, mg, ml, sc, sx = non_empty_elements
except Exception:
try:
ol, mb, mg, ml, sc = non_empty_elements
sx = u''
except Exception:
try:
model = 'SyntacticCategory'
n, t = non_empty_elements
except Exception:
return categories
if via_request:
if model == 'SyntacticCategory':
params = self.syntactic_category_create_params.copy()
params.update({
'name': n,
'type': t
})
params = json.dumps(params)
response = self.app.post(url('syntacticcategories'), params, self.json_headers,
self.extra_environ_admin)
cat_id = json.loads(response.body)['id']
categories[n] = cat_id
else:
params = self.form_create_params.copy()
params.update({
'transcription': ol,
'morpheme_break': mb,
'morpheme_gloss': mg,
'translations': [{'transcription': ml, 'grammaticality': u''}],
'syntax': sx,
'syntactic_category': categories.get(sc, u'')
})
params = json.dumps(params)
self.app.post(url('forms'), params, self.json_headers,
self.extra_environ_admin)
else:
if model == 'SyntacticCategory':
syntactic_category = model.SyntacticCategory()
syntactic_category.name = n
syntactic_category.type = t
Session.add(syntactic_category)
categories[n] = syntactic_category.id
else:
form = model.Form()
form.transcription = ol
form.morpheme_break = mb
form.morpheme_gloss = mg
translation = model.Translation()
translation.transcription = ml
form.translations.append(translation)
form.syntax = sx
form.syntacticcategory_id = categories.get(sc, None)
Session.add(form)
return categories
def add_loremipsum_to_db(loremipsum_path, via_request=False):
"""Add the contents of the file at ``loremipsum_path`` to the database."""
categories = {}
with open(loremipsum_path, 'r') as f:
i = 0
for l in f:
if i % 100 == 0:
if not via_request: Session.commit()
log.debug('%d lines processed' % i)
i = i + 1
categories = create_model(l.replace('\n', ''), categories,
via_request)
Session.commit()
loremipsum_path_no_ext = os.path.splitext(loremipsum_path)[0]
sqlalchemy_URL = self.config['sqlalchemy.url']
sqlalchemy_URL_list = sqlalchemy_URL.split(':')
olddump_script_path = os.path.join(self.test_scripts_path, 'olddump.sh')
oldload_script_path = os.path.join(self.test_scripts_path, 'oldload.sh')
RDBMS = sqlalchemy_URL_list[0]
if RDBMS == 'mysql':
mysql_dump_path = '%s_mysql.sql' % loremipsum_path_no_ext
username = sqlalchemy_URL_list[1][2:]
password = sqlalchemy_URL_list[2].split('@')[0]
dbname = sqlalchemy_URL_list[3].split('/')[1]
if os.path.exists(mysql_dump_path):
log.debug('The lorem ipsum MySQL dump file exists. Loading it...')
# Clear the current DB completely
h.clear_all_models(retain=[])
# Load the dump file to the DB
shell_script = '#!/bin/sh\nmysql -u %s -p%s %s < %s' % (
username, password, dbname, mysql_dump_path)
with open(oldload_script_path, 'w') as f:
f.write(shell_script)
os.chmod(oldload_script_path, 0744)
# Load the DB
with open(os.devnull, 'w') as f:
call([oldload_script_path], stdout=f, stderr=f)
# Destroy the load script
os.remove(oldload_script_path)
log.debug('Loaded.')
else:
log.debug('Have to import the lorem ipsum dataset from the text file and create the MySQL dump file.')
# Populate the database from the loremipusm text file and dump it
add_loremipsum_to_db(loremipsum_path, via_request=via_request)
# Write the DB dump shell script
shell_script = '#!/bin/sh\nmysqldump -u %s -p%s --no-create-info %s > %s' % (
username, password, dbname, mysql_dump_path)
with open(olddump_script_path, 'w') as f:
f.write(shell_script)
os.chmod(olddump_script_path, 0744)
# Dump the DB
with open(os.devnull, 'w') as f:
call([olddump_script_path], stdout=f, stderr=f)
# Destroy the dump script
os.remove(olddump_script_path)
log.debug('Imported and dumped.')
elif RDBMS == 'sqlite' and h.command_line_program_installed('sqlite3'):
sqlite_dump_path = '%s_sqlite.sql' % loremipsum_path_no_ext
sqlite_db = sqlalchemy_URL.split('/')[-1]
dbpath = os.path.join(self.here, sqlite_db)
if os.path.exists(sqlite_dump_path):
log.debug('The lorem ipsum SQLite dump file exists. Loading it...')
# Clear the current DB completely
h.clear_all_models(retain=[])
# Load the dump file to the DB
shell_script = '#!/bin/sh\nsqlite3 %s < %s' % (
dbpath, sqlite_dump_path)
with open(oldload_script_path, 'w') as f:
f.write(shell_script)
os.chmod(oldload_script_path, 0744)
# Load the DB
with open(os.devnull, 'w') as f:
call([oldload_script_path], stdout=f, stderr=f)
# Destroy the load script
os.remove(oldload_script_path)
log.debug('Loaded.')
else:
log.debug('Have to import the lorem ipsum dataset from the text file and create the SQLite dump file.')
# Populate the database from the loremipusm text file and dump it
add_loremipsum_to_db(loremipsum_path, via_request=via_request)
# Write the DB dump shell script
shell_script = '#!/bin/sh\nsqlite3 %s ".dump" | grep -v "^CREATE" > %s' % (
dbpath, sqlite_dump_path)
with open(olddump_script_path, 'w') as f:
f.write(shell_script)
os.chmod(olddump_script_path, 0744)
# Dump the DB
with open(os.devnull, 'w') as f:
call([olddump_script_path], stdout=f, stderr=f)
# Destroy the dump script
os.remove(olddump_script_path)
log.debug('Imported and dumped.')
forms = h.get_forms()
log.debug('Lorem ipsum data loaded. There are now %d forms in the db.' % len(forms))
# Restrict one sentential form in the db.
restricted_tag = h.generate_restricted_tag()
Session.add(restricted_tag)
Session.commit()
a_form = Session.query(model.Form).\
filter(model.Form.syntactic_category.\
has(model.SyntacticCategory.name==u'S')).first()
a_form_id = a_form.id
a_form.tags.append(restricted_tag)
Session.commit()
restricted_form = Session.query(model.Form).\
filter(model.Form.tags.any(model.Tag.name==u'restricted')).first()
assert a_form_id == restricted_form.id
def test_index(self):
"""Tests that GET & SEARCH /collectionbackups behave correctly.
"""
# Add some test data to the database.
application_settings = h.generate_default_application_settings()
source = h.generate_default_source()
restricted_tag = h.generate_restricted_tag()
file1 = h.generate_default_file()
file1.name = u'file1'
file2 = h.generate_default_file()
file2.name = u'file2'
speaker = h.generate_default_speaker()
Session.add_all([application_settings, source, restricted_tag, file1,
file2, speaker])
Session.commit()
speaker_id = speaker.id
restricted_tag_id = restricted_tag.id
tag_ids = [restricted_tag_id]
file1_id = file1.id
file2_id = file2.id
file_ids = [file1_id, file2_id]
# Create a restricted collection (via request) as the default contributor
users = h.get_users()
contributor_id = [u for u in users if u.role==u'contributor'][0].id
administrator_id = [u for u in users if u.role==u'administrator'][0].id
# Define some extra_environs
view = {'test.authentication.role': u'viewer', 'test.application_settings': True}
contrib = {'test.authentication.role': u'contributor', 'test.application_settings': True}
admin = {'test.authentication.role': u'administrator', 'test.application_settings': True}
params = self.collection_create_params.copy()
params.update({
'title': u'Created by the Contributor',
'elicitor': contributor_id,
'tags': [restricted_tag_id]
})
params = json.dumps(params)
response = self.app.post(url('collections'), params, self.json_headers, contrib)
collection_count = Session.query(model.Collection).count()
resp = json.loads(response.body)
collection_id = resp['id']
assert response.content_type == 'application/json'
assert collection_count == 1
# Update our collection (via request) as the default administrator; this
# will create one collection backup.
params = self.collection_create_params.copy()
params.update({
'url': u'find/me/here',
'title': u'Updated by the Administrator',
'speaker': speaker_id,
'tags': tag_ids + [None, u''], # None and u'' ('') will be ignored by collections.update_collection
'enterer': administrator_id # This should change nothing.
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, admin)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert response.content_type == 'application/json'
assert collection_count == 1
# Finally, update our collection (via request) as the default contributor.
# Now we will have two collection backups.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated by the Contributor',
'speaker': speaker_id,
'tags': tag_ids,
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups as the restricted enterer of the original
# collection and expect to get them all.
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 2
assert response.content_type == 'application/json'
# The admin should get them all too.
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
assert len(resp) == 2
# The viewer should get none because they're all restricted.
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=view)
resp = json.loads(response.body)
assert len(resp) == 0
# Now update the collection and de-restrict it.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated and de-restricted by the Contributor',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups. Admin and contrib should see 3 but the
# viewer should still see none.
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 3
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
assert len(resp) == 3
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=view)
resp = json.loads(response.body)
assert len(resp) == 0
assert response.content_type == 'application/json'
# Finally, update our collection in some trivial way.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated by the Contributor *again*',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups. Admin and contrib should see 4 and the
# viewer should see 1
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 4
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
all_collection_backups = resp
assert len(resp) == 4
response = self.app.get(url('collectionbackups'), headers=self.json_headers, extra_environ=view)
resp = json.loads(response.body)
unrestricted_collection_backup = resp[0]
assert len(resp) == 1
assert resp[0]['title'] == u'Updated and de-restricted by the Contributor'
restricted_collection_backups = [cb for cb in all_collection_backups
if cb != unrestricted_collection_backup]
assert len(restricted_collection_backups) == 3
# Test the paginator GET params.
paginator = {'items_per_page': 1, 'page': 2}
response = self.app.get(url('collectionbackups'), paginator, headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert len(resp['items']) == 1
assert resp['items'][0]['title'] == all_collection_backups[1]['title']
assert response.content_type == 'application/json'
# Test the order_by GET params.
order_by_params = {'order_by_model': 'CollectionBackup',
'order_by_attribute': 'id', 'order_by_direction': 'desc'}
response = self.app.get(url('collectionbackups'), order_by_params,
headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
result_set = sorted(all_collection_backups, key=lambda cb: cb['id'], reverse=True)
assert [cb['id'] for cb in resp] == [cb['id'] for cb in result_set]
# Test the order_by *with* paginator.
params = {'order_by_model': 'CollectionBackup', 'order_by_attribute': 'id',
'order_by_direction': 'desc', 'items_per_page': 1, 'page': 3}
response = self.app.get(url('collectionbackups'), params,
headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
assert result_set[2]['title'] == resp['items'][0]['title']
# Now test the show action:
# Admin should be able to GET a particular restricted collection backup
response = self.app.get(url('collectionbackup', id=restricted_collection_backups[0]['id']),
headers=self.json_headers, extra_environ=admin)
resp = json.loads(response.body)
assert resp['title'] == restricted_collection_backups[0]['title']
assert response.content_type == 'application/json'
# Viewer should receive a 403 error when attempting to do so.
response = self.app.get(url('collectionbackup', id=restricted_collection_backups[0]['id']),
headers=self.json_headers, extra_environ=view, status=403)
resp = json.loads(response.body)
assert resp['error'] == u'You are not authorized to access this resource.'
assert response.content_type == 'application/json'
# Viewer should be able to GET the unrestricted collection backup
response = self.app.get(url('collectionbackup', id=unrestricted_collection_backup['id']),
headers=self.json_headers, extra_environ=view)
resp = json.loads(response.body)
assert resp['title'] == unrestricted_collection_backup['title']
# A nonexistent cb id will return a 404 error
response = self.app.get(url('collectionbackup', id=100987),
headers=self.json_headers, extra_environ=view, status=404)
resp = json.loads(response.body)
assert resp['error'] == u'There is no collection backup with id 100987'
assert response.content_type == 'application/json'
# Test the search action
self._add_SEARCH_to_web_test_valid_methods()
# A search on collection backup titles using POST /collectionbackups/search
json_query = json.dumps({'query': {'filter':
['CollectionBackup', 'title', 'like', u'%Contributor%']}})
response = self.app.post(url('/collectionbackups/search'), json_query,
self.json_headers, admin)
resp = json.loads(response.body)
result_set = [cb for cb in all_collection_backups if u'Contributor' in cb['title']]
assert len(resp) == len(result_set) == 3
assert set([cb['id'] for cb in resp]) == set([cb['id'] for cb in result_set])
assert response.content_type == 'application/json'
# A search on collection backup titles using SEARCH /collectionbackups
json_query = json.dumps({'query': {'filter':
['CollectionBackup', 'title', 'like', u'%Administrator%']}})
response = self.app.request(url('collectionbackups'), method='SEARCH', body=json_query,
headers=self.json_headers, environ=admin)
resp = json.loads(response.body)
result_set = [cb for cb in all_collection_backups if u'Administrator' in cb['title']]
assert len(resp) == len(result_set) == 1
assert set([cb['id'] for cb in resp]) == set([cb['id'] for cb in result_set])
# Perform the two previous searches as a restricted viewer to show that
# the restricted tag is working correctly.
json_query = json.dumps({'query': {'filter':
['CollectionBackup', 'title', 'like', u'%Contributor%']}})
response = self.app.post(url('/collectionbackups/search'), json_query,
self.json_headers, view)
resp = json.loads(response.body)
result_set = [cb for cb in [unrestricted_collection_backup]
if u'Contributor' in cb['title']]
assert len(resp) == len(result_set) == 1
assert set([cb['id'] for cb in resp]) == set([cb['id'] for cb in result_set])
json_query = json.dumps({'query': {'filter':
['CollectionBackup', 'title', 'like', u'%Administrator%']}})
response = self.app.request(url('collectionbackups'), method='SEARCH', body=json_query,
headers=self.json_headers, environ=view)
resp = json.loads(response.body)
result_set = [cb for cb in [unrestricted_collection_backup]
if u'Administrator' in cb['title']]
assert len(resp) == len(result_set) == 0
# I'm just going to assume that the order by and pagination functions are
# working correctly since the implementation is essentially equivalent
# to that in the index action already tested above.
# Attempting to call edit/new/create/delete/update on a read-only resource
# will return a 404 response
response = self.app.get(url('edit_collectionbackup', id=2232), status=404)
assert json.loads(response.body)['error'] == u'This resource is read-only.'
response = self.app.get(url('new_collectionbackup', id=2232), status=404)
assert json.loads(response.body)['error'] == u'This resource is read-only.'
response = self.app.post(url('collectionbackups'), status=404)
assert json.loads(response.body)['error'] == u'This resource is read-only.'
response = self.app.put(url('collectionbackup', id=2232), status=404)
assert json.loads(response.body)['error'] == u'This resource is read-only.'
response = self.app.delete(url('collectionbackup', id=2232), status=404)
assert json.loads(response.body)['error'] == u'This resource is read-only.'
assert response.content_type == 'application/json'
def test_index(self):
"""Tests that GET & SEARCH /collectionbackups behave correctly.
"""
# Add some test data to the database.
application_settings = h.generate_default_application_settings()
source = h.generate_default_source()
restricted_tag = h.generate_restricted_tag()
file1 = h.generate_default_file()
file1.name = u'file1'
file2 = h.generate_default_file()
file2.name = u'file2'
speaker = h.generate_default_speaker()
Session.add_all([
application_settings, source, restricted_tag, file1, file2, speaker
])
Session.commit()
speaker_id = speaker.id
restricted_tag_id = restricted_tag.id
tag_ids = [restricted_tag_id]
file1_id = file1.id
file2_id = file2.id
file_ids = [file1_id, file2_id]
# Create a restricted collection (via request) as the default contributor
users = h.get_users()
contributor_id = [u for u in users if u.role == u'contributor'][0].id
administrator_id = [u for u in users
if u.role == u'administrator'][0].id
# Define some extra_environs
view = {
'test.authentication.role': u'viewer',
'test.application_settings': True
}
contrib = {
'test.authentication.role': u'contributor',
'test.application_settings': True
}
admin = {
'test.authentication.role': u'administrator',
'test.application_settings': True
}
params = self.collection_create_params.copy()
params.update({
'title': u'Created by the Contributor',
'elicitor': contributor_id,
'tags': [restricted_tag_id]
})
params = json.dumps(params)
response = self.app.post(url('collections'), params, self.json_headers,
contrib)
collection_count = Session.query(model.Collection).count()
resp = json.loads(response.body)
collection_id = resp['id']
assert response.content_type == 'application/json'
assert collection_count == 1
# Update our collection (via request) as the default administrator; this
# will create one collection backup.
params = self.collection_create_params.copy()
params.update({
'url': u'find/me/here',
'title': u'Updated by the Administrator',
'speaker': speaker_id,
'tags': tag_ids + [
None, u''
], # None and u'' ('') will be ignored by collections.update_collection
'enterer': administrator_id # This should change nothing.
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, admin)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert response.content_type == 'application/json'
assert collection_count == 1
# Finally, update our collection (via request) as the default contributor.
# Now we will have two collection backups.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated by the Contributor',
'speaker': speaker_id,
'tags': tag_ids,
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups as the restricted enterer of the original
# collection and expect to get them all.
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 2
assert response.content_type == 'application/json'
# The admin should get them all too.
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert len(resp) == 2
# The viewer should get none because they're all restricted.
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=view)
resp = json.loads(response.body)
assert len(resp) == 0
# Now update the collection and de-restrict it.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated and de-restricted by the Contributor',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups. Admin and contrib should see 3 but the
# viewer should still see none.
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 3
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert len(resp) == 3
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=view)
resp = json.loads(response.body)
assert len(resp) == 0
assert response.content_type == 'application/json'
# Finally, update our collection in some trivial way.
params = self.collection_create_params.copy()
params.update({
'title': u'Updated by the Contributor *again*',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(url('collection', id=collection_id), params,
self.json_headers, contrib)
resp = json.loads(response.body)
collection_count = Session.query(model.Collection).count()
assert collection_count == 1
# Now GET the collection backups. Admin and contrib should see 4 and the
# viewer should see 1
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=contrib)
resp = json.loads(response.body)
assert len(resp) == 4
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
all_collection_backups = resp
assert len(resp) == 4
response = self.app.get(url('collectionbackups'),
headers=self.json_headers,
extra_environ=view)
resp = json.loads(response.body)
unrestricted_collection_backup = resp[0]
assert len(resp) == 1
assert resp[0][
'title'] == u'Updated and de-restricted by the Contributor'
restricted_collection_backups = [
cb for cb in all_collection_backups
if cb != unrestricted_collection_backup
]
assert len(restricted_collection_backups) == 3
# Test the paginator GET params.
paginator = {'items_per_page': 1, 'page': 2}
response = self.app.get(url('collectionbackups'),
paginator,
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert len(resp['items']) == 1
assert resp['items'][0]['title'] == all_collection_backups[1]['title']
assert response.content_type == 'application/json'
# Test the order_by GET params.
order_by_params = {
'order_by_model': 'CollectionBackup',
'order_by_attribute': 'id',
'order_by_direction': 'desc'
}
response = self.app.get(url('collectionbackups'),
order_by_params,
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
result_set = sorted(all_collection_backups,
key=lambda cb: cb['id'],
reverse=True)
assert [cb['id'] for cb in resp] == [cb['id'] for cb in result_set]
# Test the order_by *with* paginator.
params = {
'order_by_model': 'CollectionBackup',
'order_by_attribute': 'id',
'order_by_direction': 'desc',
'items_per_page': 1,
'page': 3
}
response = self.app.get(url('collectionbackups'),
params,
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert result_set[2]['title'] == resp['items'][0]['title']
# Now test the show action:
# Admin should be able to GET a particular restricted collection backup
response = self.app.get(url('collectionbackup',
id=restricted_collection_backups[0]['id']),
headers=self.json_headers,
extra_environ=admin)
resp = json.loads(response.body)
assert resp['title'] == restricted_collection_backups[0]['title']
assert response.content_type == 'application/json'
# Viewer should receive a 403 error when attempting to do so.
response = self.app.get(url('collectionbackup',
id=restricted_collection_backups[0]['id']),
headers=self.json_headers,
extra_environ=view,
status=403)
resp = json.loads(response.body)
assert resp[
'error'] == u'You are not authorized to access this resource.'
assert response.content_type == 'application/json'
# Viewer should be able to GET the unrestricted collection backup
response = self.app.get(url('collectionbackup',
id=unrestricted_collection_backup['id']),
headers=self.json_headers,
extra_environ=view)
resp = json.loads(response.body)
assert resp['title'] == unrestricted_collection_backup['title']
# A nonexistent cb id will return a 404 error
response = self.app.get(url('collectionbackup', id=100987),
headers=self.json_headers,
extra_environ=view,
status=404)
resp = json.loads(response.body)
assert resp['error'] == u'There is no collection backup with id 100987'
assert response.content_type == 'application/json'
# Test the search action
self._add_SEARCH_to_web_test_valid_methods()
# A search on collection backup titles using POST /collectionbackups/search
json_query = json.dumps({
'query': {
'filter':
['CollectionBackup', 'title', 'like', u'%Contributor%']
}
})
response = self.app.post(url('/collectionbackups/search'), json_query,
self.json_headers, admin)
resp = json.loads(response.body)
result_set = [
cb for cb in all_collection_backups
if u'Contributor' in cb['title']
]
assert len(resp) == len(result_set) == 3
assert set([cb['id']
for cb in resp]) == set([cb['id'] for cb in result_set])
assert response.content_type == 'application/json'
# A search on collection backup titles using SEARCH /collectionbackups
json_query = json.dumps({
'query': {
'filter':
['CollectionBackup', 'title', 'like', u'%Administrator%']
}
})
response = self.app.request(url('collectionbackups'),
method='SEARCH',
body=json_query,
headers=self.json_headers,
environ=admin)
resp = json.loads(response.body)
result_set = [
cb for cb in all_collection_backups
if u'Administrator' in cb['title']
]
assert len(resp) == len(result_set) == 1
assert set([cb['id']
for cb in resp]) == set([cb['id'] for cb in result_set])
# Perform the two previous searches as a restricted viewer to show that
# the restricted tag is working correctly.
json_query = json.dumps({
'query': {
'filter':
['CollectionBackup', 'title', 'like', u'%Contributor%']
}
})
response = self.app.post(url('/collectionbackups/search'), json_query,
self.json_headers, view)
resp = json.loads(response.body)
result_set = [
cb for cb in [unrestricted_collection_backup]
if u'Contributor' in cb['title']
]
assert len(resp) == len(result_set) == 1
assert set([cb['id']
for cb in resp]) == set([cb['id'] for cb in result_set])
json_query = json.dumps({
'query': {
'filter':
['CollectionBackup', 'title', 'like', u'%Administrator%']
}
})
response = self.app.request(url('collectionbackups'),
method='SEARCH',
body=json_query,
headers=self.json_headers,
environ=view)
resp = json.loads(response.body)
result_set = [
cb for cb in [unrestricted_collection_backup]
if u'Administrator' in cb['title']
]
assert len(resp) == len(result_set) == 0
# I'm just going to assume that the order by and pagination functions are
# working correctly since the implementation is essentially equivalent
# to that in the index action already tested above.
# Attempting to call edit/new/create/delete/update on a read-only resource
# will return a 404 response
response = self.app.get(url('edit_collectionbackup', id=2232),
status=404)
assert json.loads(
response.body)['error'] == u'This resource is read-only.'
response = self.app.get(url('new_collectionbackup', id=2232),
status=404)
assert json.loads(
response.body)['error'] == u'This resource is read-only.'
response = self.app.post(url('collectionbackups'), status=404)
assert json.loads(
response.body)['error'] == u'This resource is read-only.'
response = self.app.put(url('collectionbackup', id=2232), status=404)
assert json.loads(
response.body)['error'] == u'This resource is read-only.'
response = self.app.delete(url('collectionbackup', id=2232),
status=404)
assert json.loads(
response.body)['error'] == u'This resource is read-only.'
assert response.content_type == 'application/json'
def test_aaa_initialize(self):
"""Initialize the database using pseudo-data generated from random lorem ipsum sentences.
These are located in ``onlinelinguisticdatabase/tests/data/corpora``.
The data contain morphologically analyzed sentences, their component
morphemes, and syntactic categories. The sentences have phrase
structure trees in bracket notation.
The test will try to load the lorem ipsum dataset from a MySQL/SQLite
dump file in ``onlinelinguisticdatabase/tests/data/corpora``. If the
dump file corresponding to ``loremipsum_path`` does not exist, it will
import the lorem ipsum data directly from the text files and create
the dump file so that future tests can run more speedily. The
``loremipsum100_path``, ``loremipsum1000_path``, ``loremipsum10000_path``
and ``loremipsum30000_path`` files are available and contain 100, 1000
and 10,000 sentences, respectively.
Setting the ``via_request`` variable to ``True`` will cause all of the
forms to be created via request, i.e., via
``self.app.post(url('forms))...``. This is much slower but may be
desirable since values for the morphological analysis attributes
will be generated.
.. note::
In order to run ``mysqldump`` with the MySQL user listed in
``test.ini``, that user must have permission to lock and update
tables (alter and file privileges may also be required ...)::
mysql -u root -p<root_password>
grant lock tables, update on old_test.* to 'old'@'localhost';
.. warning::
Loading the .txt or .sql files with the ``via_request`` option set to
``True`` will take a very long time. This might be an argument for
separating the interface and logic components of the controllers so
that a "core" HTTP-less OLD application could be exposed. This
would facilitate the creation of models with system-generated data
and validation but without the HTTP overhead...
"""
########################################################################
# Configure lorem ipsum data set import
########################################################################
# Set ``loremipsum_path`` this to ``self.loremipsum100_path``,
# ``self.loremipsum1000_path`` or ``self.loremipsum10000_path``.
# WARNING: the larger ones will take a long time.
# Use the 10,000-sentence lorem ipsum dataset to ensure that
# very large corpora are handled correctly.
loremipsum_path = self.loremipsum100_path
# Set ``via_request`` to ``True`` to create all forms via HTTP requests.
via_request = True
self._add_SEARCH_to_web_test_valid_methods()
# Add an application settings so that morpheme references will work out right.
application_settings = h.generate_default_application_settings()
Session.add(application_settings)
Session.commit()
def create_model(line, categories, via_request=False):
"""Create a model (form or syncat) using the string in ``line``."""
model = 'Form'
elements = unicode(line).split('\t')
non_empty_elements = filter(None, elements)
try:
ol, mb, mg, ml, sc, sx = non_empty_elements
except Exception:
try:
ol, mb, mg, ml, sc = non_empty_elements
sx = u''
except Exception:
try:
model = 'SyntacticCategory'
n, t = non_empty_elements
except Exception:
return categories
if via_request:
if model == 'SyntacticCategory':
params = self.syntactic_category_create_params.copy()
params.update({'name': n, 'type': t})
params = json.dumps(params)
response = self.app.post(url('syntacticcategories'),
params, self.json_headers,
self.extra_environ_admin)
cat_id = json.loads(response.body)['id']
categories[n] = cat_id
else:
params = self.form_create_params.copy()
params.update({
'transcription':
ol,
'morpheme_break':
mb,
'morpheme_gloss':
mg,
'translations': [{
'transcription': ml,
'grammaticality': u''
}],
'syntax':
sx,
'syntactic_category':
categories.get(sc, u'')
})
params = json.dumps(params)
self.app.post(url('forms'), params, self.json_headers,
self.extra_environ_admin)
else:
if model == 'SyntacticCategory':
syntactic_category = model.SyntacticCategory()
syntactic_category.name = n
syntactic_category.type = t
Session.add(syntactic_category)
categories[n] = syntactic_category.id
else:
form = model.Form()
form.transcription = ol
form.morpheme_break = mb
form.morpheme_gloss = mg
translation = model.Translation()
translation.transcription = ml
form.translations.append(translation)
form.syntax = sx
form.syntacticcategory_id = categories.get(sc, None)
Session.add(form)
return categories
def add_loremipsum_to_db(loremipsum_path, via_request=False):
"""Add the contents of the file at ``loremipsum_path`` to the database."""
categories = {}
with open(loremipsum_path, 'r') as f:
i = 0
for l in f:
if i % 100 == 0:
if not via_request: Session.commit()
log.debug('%d lines processed' % i)
i = i + 1
categories = create_model(l.replace('\n', ''), categories,
via_request)
Session.commit()
loremipsum_path_no_ext = os.path.splitext(loremipsum_path)[0]
sqlalchemy_URL = self.config['sqlalchemy.url']
sqlalchemy_URL_list = sqlalchemy_URL.split(':')
olddump_script_path = os.path.join(self.test_scripts_path,
'olddump.sh')
oldload_script_path = os.path.join(self.test_scripts_path,
'oldload.sh')
RDBMS = sqlalchemy_URL_list[0]
if RDBMS == 'mysql':
mysql_dump_path = '%s_mysql.sql' % loremipsum_path_no_ext
username = sqlalchemy_URL_list[1][2:]
password = sqlalchemy_URL_list[2].split('@')[0]
dbname = sqlalchemy_URL_list[3].split('/')[1]
if os.path.exists(mysql_dump_path):
log.debug(
'The lorem ipsum MySQL dump file exists. Loading it...')
# Clear the current DB completely
h.clear_all_models(retain=[])
# Load the dump file to the DB
shell_script = '#!/bin/sh\nmysql -u %s -p%s %s < %s' % (
username, password, dbname, mysql_dump_path)
with open(oldload_script_path, 'w') as f:
f.write(shell_script)
os.chmod(oldload_script_path, 0744)
# Load the DB
with open(os.devnull, 'w') as f:
call([oldload_script_path], stdout=f, stderr=f)
# Destroy the load script
os.remove(oldload_script_path)
log.debug('Loaded.')
else:
log.debug(
'Have to import the lorem ipsum dataset from the text file and create the MySQL dump file.'
)
# Populate the database from the loremipusm text file and dump it
add_loremipsum_to_db(loremipsum_path, via_request=via_request)
# Write the DB dump shell script
# Note: the --single-transaction option seems to be required (on Mac MySQL 5.6 using InnoDB tables ...)
# see http://forums.mysql.com/read.php?10,108835,112951#msg-112951
shell_script = '#!/bin/sh\nmysqldump -u %s -p%s --single-transaction --no-create-info --result-file=%s %s' % (
username, password, mysql_dump_path, dbname)
with open(olddump_script_path, 'w') as f:
f.write(shell_script)
os.chmod(olddump_script_path, 0744)
# Dump the DB
with open(os.devnull, 'w') as f:
call([olddump_script_path], stdout=f, stderr=f)
# Destroy the dump script
os.remove(olddump_script_path)
log.debug('Imported and dumped.')
elif RDBMS == 'sqlite' and h.command_line_program_installed('sqlite3'):
sqlite_dump_path = '%s_sqlite.sql' % loremipsum_path_no_ext
sqlite_db = sqlalchemy_URL.split('/')[-1]
dbpath = os.path.join(self.here, sqlite_db)
if os.path.exists(sqlite_dump_path):
log.debug(
'The lorem ipsum SQLite dump file exists. Loading it...')
# Clear the current DB completely
h.clear_all_models(retain=[])
# Load the dump file to the DB
shell_script = '#!/bin/sh\nsqlite3 %s < %s' % (
dbpath, sqlite_dump_path)
with open(oldload_script_path, 'w') as f:
f.write(shell_script)
os.chmod(oldload_script_path, 0744)
# Load the DB
with open(os.devnull, 'w') as f:
call([oldload_script_path], stdout=f, stderr=f)
# Destroy the load script
os.remove(oldload_script_path)
log.debug('Loaded.')
else:
log.debug(
'Have to import the lorem ipsum dataset from the text file and create the SQLite dump file.'
)
# Populate the database from the loremipusm text file and dump it
add_loremipsum_to_db(loremipsum_path, via_request=via_request)
# Write the DB dump shell script
shell_script = '#!/bin/sh\nsqlite3 %s ".dump" | grep -v "^CREATE" > %s' % (
dbpath, sqlite_dump_path)
with open(olddump_script_path, 'w') as f:
f.write(shell_script)
os.chmod(olddump_script_path, 0744)
# Dump the DB
with open(os.devnull, 'w') as f:
call([olddump_script_path], stdout=f, stderr=f)
# Destroy the dump script
os.remove(olddump_script_path)
log.debug('Imported and dumped.')
forms = h.get_forms()
log.debug(
'Lorem ipsum data loaded. There are now %d forms in the db.' %
len(forms))
# Restrict one sentential form in the db.
restricted_tag = h.generate_restricted_tag()
Session.add(restricted_tag)
Session.commit()
a_form = Session.query(model.Form).\
filter(model.Form.syntactic_category.\
has(model.SyntacticCategory.name==u'S')).first()
a_form_id = a_form.id
a_form.tags.append(restricted_tag)
Session.commit()
restricted_form = Session.query(model.Form).\
filter(model.Form.tags.any(model.Tag.name==u'restricted')).first()
assert a_form_id == restricted_form.id
def setup_app(command, conf, vars):
"""Commands to setup onlinelinguisticdatabase."""
config = load_environment(conf.global_conf, conf.local_conf)
log.info('Environment loaded.')
Base.metadata.create_all(bind=Session.bind)
filename = os.path.split(conf.filename)[
-1] # e.g., production.ini, development.ini, test.ini, ...
# Create the ``store`` directory and those for file, analysis and corpora
# objects and their subdirectories. See ``lib.utils.py`` for details.
h.create_OLD_directories(config=config)
# ISO-639-3 Language data for the languages table
log.info("Retrieving ISO-639-3 languages data.")
languages = h.get_language_objects(filename, config)
# Get default users.
log.info("Creating a default administrator, contributor and viewer.")
administrator = h.generate_default_administrator(config_filename=filename)
contributor = h.generate_default_contributor(config_filename=filename)
viewer = h.generate_default_viewer(config_filename=filename)
# If we are running tests, make sure the test db contains only language data.
if filename == 'test.ini':
# Permanently drop any existing tables
Base.metadata.drop_all(bind=Session.bind, checkfirst=True)
log.info("Existing tables dropped.")
# Create the tables if they don't already exist
Base.metadata.create_all(bind=Session.bind, checkfirst=True)
log.info('Tables created.')
Session.add_all(languages + [administrator, contributor, viewer])
Session.commit()
# Not a test: add a bunch of nice defaults.
else:
# Create the _requests_tests.py script
requests_tests_path = os.path.join(config['pylons.paths']['root'],
'tests', 'scripts',
'_requests_tests.py')
# This line is problematic in production apps because the
# _requests_tests.py file is not included in the build. So, I'm
# commenting it out by default.
# copyfile(requests_tests_path, '_requests_tests.py')
# Create the tables if they don't already exist
Base.metadata.create_all(bind=Session.bind, checkfirst=True)
log.info('Tables created.')
# Get default home & help pages.
log.info("Creating default home and help pages.")
homepage = h.generate_default_home_page()
helppage = h.generate_default_help_page()
# Get default application settings.
log.info("Generating default application settings.")
application_settings = h.generate_default_application_settings()
# Get default tags and categories
log.info("Creating some useful tags and categories.")
restricted_tag = h.generate_restricted_tag()
foreign_word_tag = h.generate_foreign_word_tag()
S = h.generate_s_syntactic_category()
N = h.generate_n_syntactic_category()
V = h.generate_v_syntactic_category()
# Initialize the database
log.info("Adding defaults.")
data = [
administrator, contributor, viewer, homepage, helppage,
application_settings, restricted_tag, foreign_word_tag
]
if config['add_language_data'] != '0':
data += languages
if config['empty_database'] == '0':
Session.add_all(data)
Session.commit()
log.info("OLD successfully set up.")
