def test_a_initialize(self):
"""Initialize the database for /rememberedforms tests."""
dbsession = self.dbsession
db = DBUtils(dbsession, self.settings)
self.create_db()
db.clear_all_models()
administrator = omb.generate_default_administrator(
settings=self.settings)
contributor = omb.generate_default_contributor(settings=self.settings)
viewer = omb.generate_default_viewer(settings=self.settings)
dbsession.add_all([administrator, contributor, viewer])
dbsession.commit()
_create_test_data(dbsession, db, self.n)
add_SEARCH_to_web_test_valid_methods()
# Create an application settings where the contributor is unrestricted
viewer, contributor, administrator = get_users(db)
application_settings = omb.generate_default_application_settings()
application_settings.unrestricted_users = [contributor]
dbsession.add(application_settings)
dbsession.commit()
def test_aaa_initialize(self):
"""Initialize the database using pseudo-data generated from random
lorem ipsum sentences.
These are located in ``old/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`` the MySQL 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...
"""
self.create_db()
dbsession = self.dbsession
db = DBUtils(dbsession, self.settings)
###################################################################
# 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
add_SEARCH_to_web_test_valid_methods()
# Add an application settings so that morpheme references will work
# out right.
application_settings = omb.generate_default_application_settings()
dbsession.add(application_settings)
dbsession.commit()
def create_model(line, categories, via_request=False):
"""Create a model (form or syncat) using the string in ``line``."""
model = 'Form'
elements = str(line).split('\t')
non_empty_elements = list(filter(None, elements))
try:
ol, mb, mg, ml, sc, sx = non_empty_elements
except ValueError:
try:
ol, mb, mg, ml, sc = non_empty_elements
sx = ''
except ValueError:
try:
model = 'SyntacticCategory'
n, t = non_empty_elements
except ValueError:
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(
'/{}/syntacticcategories'.format(self.old_name),
params, self.json_headers, self.extra_environ_admin)
cat_id = response.json_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': ''
}],
'syntax':
sx,
'syntactic_category':
categories.get(sc, '')
})
params = json.dumps(params)
self.app.post('/{}/forms'.format(self.old_name), params,
self.json_headers, self.extra_environ_admin)
else:
if model == 'SyntacticCategory':
syntactic_category = old_models.SyntacticCategory()
syntactic_category.name = n
syntactic_category.type = t
dbsession.add(syntactic_category)
categories[n] = syntactic_category.id
else:
form = old_models.Form()
form.transcription = ol
form.morpheme_break = mb
form.morpheme_gloss = mg
translation = old_models.Translation()
translation.transcription = ml
form.translations.append(translation)
form.syntax = sx
form.syntacticcategory_id = categories.get(sc, None)
dbsession.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: dbsession.commit()
LOGGER.debug('%d lines processed' % i)
i = i + 1
categories = create_model(l.replace('\n', ''), categories,
via_request)
dbsession.commit()
loremipsum_path_no_ext = os.path.splitext(loremipsum_path)[0]
sqlalchemy_URL = self.settings['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.startswith('mysql'):
RDBMS = 'mysql'
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]
# This is not an option anymore: too frustrated trying to load
# the dump file.
if False and os.path.exists(mysql_dump_path):
LOGGER.debug(
'The lorem ipsum MySQL dump file exists. Loading it...')
# Clear the current DB completely
db.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, 0o744)
# 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)
LOGGER.debug('Loaded.')
else:
LOGGER.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, 0o744)
# 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)
LOGGER.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_full_dump_path = '%s_full_dump_tmp_sqlite.sql' % (
loremipsum_path_no_ext, )
sqlite_schema_dump_path = '%s_schema_dump_tmp_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):
LOGGER.debug(
'The lorem ipsum SQLite dump file exists. Loading it...')
print(
'The lorem ipsum SQLite dump file exists. Loading it...')
# Destroy the sqlite db file
os.remove(dbpath)
# 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, 0o744)
# 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)
LOGGER.debug('Loaded.')
else:
LOGGER.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\n'
'sqlite3 {dbpath} .dump > {dump_path}\n'.format(
dbpath=dbpath, dump_path=sqlite_dump_path))
_shell_script = ('#!/bin/sh\n'
'sqlite3 {dbpath} .schema > {schema_path}\n'
'sqlite3 {dbpath} .dump > {full_dump_path}\n'
'grep -vx -f {schema_path} {full_dump_path} >'
' {dump_path}\n'
'rm {schema_path}\n'
'rm {full_dump_path}\n'.format(
dbpath=dbpath,
schema_path=sqlite_schema_dump_path,
full_dump_path=sqlite_full_dump_path,
dump_path=sqlite_dump_path))
# 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, 0o744)
# 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)
LOGGER.debug('Imported and dumped.')
forms = db.get_forms()
LOGGER.debug(
'Lorem ipsum data loaded. There are now %d forms in the db.' %
len(forms))
print('Lorem ipsum data loaded. There are now %d forms in the db.' %
len(forms))
# Restrict one sentential form in the db.
restricted_tag = omb.generate_restricted_tag()
dbsession.add(restricted_tag)
dbsession.commit()
a_form = dbsession.query(old_models.Form).\
filter(old_models.Form.syntactic_category.\
has(old_models.SyntacticCategory.name=='S')).first()
a_form_id = a_form.id
a_form.tags.append(restricted_tag)
dbsession.commit()
restricted_form = dbsession.query(old_models.Form).\
filter(old_models.Form.tags.any(
old_models.Tag.name=='restricted')).first()
assert a_form_id == restricted_form.id
def test_index(self):
"""Tests that GET & SEARCH /collectionbackups behave correctly.
"""
db = DBUtils(self.dbsession, self.settings)
# Add some test data to the database.
application_settings = omb.generate_default_application_settings()
source = omb.generate_default_source()
restricted_tag = omb.generate_restricted_tag()
file1 = omb.generate_default_file()
file1.name = 'file1'
file2 = omb.generate_default_file()
file2.name = 'file2'
speaker = omb.generate_default_speaker()
self.dbsession.add_all([
application_settings, source, restricted_tag, file1, file2, speaker
])
self.dbsession.flush()
speaker_id = speaker.id
restricted_tag_id = restricted_tag.id
file1_id = file1.id
file2_id = file2.id
self.dbsession.commit()
tag_ids = [restricted_tag_id]
file_ids = [file1_id, file2_id]
# Create a restricted collection (via request) as the default contributor
users = db.get_users()
contributor_id = [u for u in users if u.role == 'contributor'][0].id
administrator_id = [u for u in users
if u.role == 'administrator'][0].id
# Define some extra_environs
view = {
'test.authentication.role': 'viewer',
'test.application_settings': True
}
contrib = {
'test.authentication.role': 'contributor',
'test.application_settings': True
}
admin = {
'test.authentication.role': 'administrator',
'test.application_settings': True
}
params = self.collection_create_params.copy()
params.update({
'title': 'Created by the Contributor',
'elicitor': contributor_id,
'tags': [restricted_tag_id]
})
params = json.dumps(params)
response = self.app.post(coll_url('create'), params, self.json_headers,
contrib)
collection_count = self.dbsession.query(old_models.Collection).count()
resp = response.json_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': 'find/me/here',
'title': 'Updated by the Administrator',
'speaker': speaker_id,
'tags': tag_ids + [
None, ''
], # None and '' ('') will be ignored by collections.update_collection
'enterer': administrator_id # This should change nothing.
})
params = json.dumps(params)
response = self.app.put(coll_url('update', id=collection_id), params,
self.json_headers, admin)
resp = response.json_body
collection_count = self.dbsession.query(old_models.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': 'Updated by the Contributor',
'speaker': speaker_id,
'tags': tag_ids,
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(coll_url('update', id=collection_id), params,
self.json_headers, contrib)
resp = response.json_body
collection_count = self.dbsession.query(old_models.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('index'),
headers=self.json_headers,
extra_environ=contrib)
resp = response.json_body
assert len(resp) == 2
assert response.content_type == 'application/json'
# The admin should get them all too.
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=admin)
resp = response.json_body
assert len(resp) == 2
# The viewer should get none because they're all restricted.
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=view)
resp = response.json_body
assert len(resp) == 0
# Now update the collection and de-restrict it.
params = self.collection_create_params.copy()
params.update({
'title': 'Updated and de-restricted by the Contributor',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(coll_url('update', id=collection_id), params,
self.json_headers, contrib)
resp = response.json_body
collection_count = self.dbsession.query(old_models.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('index'),
headers=self.json_headers,
extra_environ=contrib)
resp = response.json_body
assert len(resp) == 3
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=admin)
resp = response.json_body
assert len(resp) == 3
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=view)
resp = response.json_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': 'Updated by the Contributor *again*',
'speaker': speaker_id,
'tags': [],
'files': file_ids
})
params = json.dumps(params)
response = self.app.put(coll_url('update', id=collection_id), params,
self.json_headers, contrib)
resp = response.json_body
collection_count = self.dbsession.query(old_models.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('index'),
headers=self.json_headers,
extra_environ=contrib)
resp = response.json_body
assert len(resp) == 4
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=admin)
resp = response.json_body
all_collection_backups = resp
assert len(resp) == 4
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=view)
resp = response.json_body
unrestricted_collection_backup = resp[0]
assert len(resp) == 1
assert resp[0][
'title'] == '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('index'),
paginator,
headers=self.json_headers,
extra_environ=admin)
resp = response.json_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('index'),
order_by_params,
headers=self.json_headers,
extra_environ=admin)
resp = response.json_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('index'),
params,
headers=self.json_headers,
extra_environ=admin)
resp = response.json_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('show',
id=restricted_collection_backups[0]['id']),
headers=self.json_headers,
extra_environ=admin)
resp = response.json_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('show',
id=restricted_collection_backups[0]['id']),
headers=self.json_headers,
extra_environ=view,
status=403)
resp = response.json_body
assert resp[
'error'] == '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('show',
id=unrestricted_collection_backup['id']),
headers=self.json_headers,
extra_environ=view)
resp = response.json_body
assert resp['title'] == unrestricted_collection_backup['title']
# A nonexistent cb id will return a 404 error
response = self.app.get(url('show', id=100987),
headers=self.json_headers,
extra_environ=view,
status=404)
resp = response.json_body
assert resp['error'] == 'There is no collection backup with id 100987'
assert response.content_type == 'application/json'
# Test the search action
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', '%Contributor%']
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, admin)
resp = response.json_body
result_set = [
cb for cb in all_collection_backups if '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', '%Administrator%']
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=admin)
resp = response.json_body
result_set = [
cb for cb in all_collection_backups
if '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', '%Contributor%']
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, view)
resp = response.json_body
result_set = [
cb for cb in [unrestricted_collection_backup]
if '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', '%Administrator%']
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=view)
resp = response.json_body
result_set = [
cb for cb in [unrestricted_collection_backup]
if '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', id=2232),
status=404,
extra_environ=admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.get(url('new', id=2232),
status=404,
extra_environ=admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.post(url('create'),
status=404,
extra_environ=admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.put(url('update', id=2232),
status=404,
extra_environ=admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.delete(url('delete', id=2232),
status=404,
extra_environ=admin)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
add_SEARCH_to_web_test_valid_methods()
def test_search(self):
"""Tests that SEARCH /formsearches (a.k.a. POST /formsearches/search) correctly returns an array of formsearches based on search criteria."""
dbsession = self.dbsession
db = DBUtils(dbsession, self.settings)
# Create some form_searches (and other models) to search and add SEARCH to the list of allowable methods
_create_test_data(db, dbsession, 100)
add_SEARCH_to_web_test_valid_methods()
RDBMSName = h.get_RDBMS_name(self.settings)
form_searches = json.loads(
json.dumps([
self.fix_formsearch(fs.get_dict())
for fs in db.get_form_searches(True)
]))
# Searching where values may be NULL
json_query = json.dumps(
{'query': {
'filter': ['FormSearch', 'search', 'like', '%2%']
}})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
result_set = [
fs for fs in form_searches if '2' in json.dumps(fs['search'])
]
assert resp
assert len(resp) == len(result_set)
assert set([s['id']
for s in resp]) == set([s['id'] for s in result_set])
assert response.content_type == 'application/json'
# A fairly complex search
json_query = json.dumps({
'query': {
'filter': [
'and',
[['FormSearch', 'name', 'regex', '[13456]'],
['not', ['FormSearch', 'name', 'like', '%F%']],
[
'or',
[['FormSearch', 'search', 'regex', '[1456]'],
[
'FormSearch', 'datetime_modified', '>',
yesterday_timestamp.isoformat()
]]
]]
]
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
mysql_engine = old_models.Model.__table_args__.get('mysql_engine')
if RDBMSName == 'mysql' and mysql_engine == 'InnoDB':
_yesterday_timestamp = h.round_datetime(yesterday_timestamp)
else:
_yesterday_timestamp = yesterday_timestamp
result_set = [
fs for fs in form_searches
if re.search('[13456]', fs['name']) and not 'F' in fs['name'] and (
re.search('[1456]', json.dumps(fs['search']))
or fs['datetime_modified'] > _yesterday_timestamp.isoformat())
]
assert resp
assert len(resp) == len(result_set)
assert set([s['id']
for s in resp]) == set([s['id'] for s in result_set])
# A basic search with a paginator provided.
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'like', '%3%']
},
'paginator': {
'page': 2,
'items_per_page': 5
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=self.extra_environ_admin)
resp = response.json_body
result_set = [
fs for fs in form_searches
if json.dumps(fs['search']) and '3' in json.dumps(fs['search'])
]
assert resp['paginator']['count'] == len(result_set)
assert len(resp['items']) == 5
assert resp['items'][0]['id'] == result_set[5]['id']
assert resp['items'][-1]['id'] == result_set[9]['id']
# An invalid paginator (here 'page' is less than 1) will result in formencode.Invalid
# being raised resulting in a response with a 400 status code and a JSON error msg.
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'like', '%3%']
},
'paginator': {
'page': 0,
'items_per_page': 10
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=self.extra_environ_admin,
status=400)
resp = response.json_body
assert resp['errors'][
'page'] == 'Please enter a number that is 1 or greater'
assert response.content_type == 'application/json'
# Some "invalid" paginators will silently fail. For example, if there is
# no 'pages' key, then SEARCH /formsearches will just assume there is no paginator
# and all of the results will be returned.
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'like', '%3%']
},
'paginator': {
'pages': 1,
'items_per_page': 10
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=self.extra_environ_admin)
resp = response.json_body
assert len(resp) == len([
fs for fs in form_searches
if json.dumps(fs['search']) and '3' in json.dumps(fs['search'])
])
# Adding a 'count' key to the paginator object in the request will spare
# the server from running query.count(). Note that the server will not
# attempt to verify the count (since that would defeat the purpose) but
# will simply pass it back. The server trusts that the client is passing
# in a factual count. Here we pass in an inaccurate count for demonstration.
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'like', '%3%']
},
'paginator': {
'page': 2,
'items_per_page': 4,
'count': 750
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=self.extra_environ_admin)
resp = response.json_body
assert resp['paginator']['count'] == 750
assert len(resp['items']) == 4
assert resp['items'][0]['id'] == result_set[4]['id']
assert resp['items'][-1]['id'] == result_set[7]['id']
# Test order by: order by name descending
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'regex', '.'],
'order_by': ['FormSearch', 'name', 'desc']
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
result_set = sorted(form_searches,
key=lambda fs: fs['name'].lower(),
reverse=True)
assert len(resp) == 100
rs_names = [fs['name'] for fs in result_set]
r_names = [fs['name'] for fs in resp]
assert rs_names == r_names
assert resp[0]['name'] == 'form search 99'
assert resp[-1]['name'] == 'form search 1'
# order by with missing direction defaults to 'asc'
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'regex', '.'],
'order_by': ['FormSearch', 'name']
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
assert len(resp) == 100
assert resp[0]['name'] == 'form search 1'
assert resp[-1]['name'] == 'form search 99'
assert response.content_type == 'application/json'
# order by with unknown direction defaults to 'asc'
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'regex', '.'],
'order_by': ['FormSearch', 'name', 'descending']
}
})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
assert len(resp) == 100
assert resp[0]['name'] == 'form search 1'
assert resp[-1]['name'] == 'form search 99'
# syntactically malformed order by
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'search', 'regex', '.'],
'order_by': ['FormSearch']
}
})
response = self.app.post(url('search_post'),
json_query,
self.json_headers,
self.extra_environ_admin,
status=400)
resp = response.json_body
assert resp['errors'][
'OrderByError'] == 'The provided order by expression was invalid.'
assert response.content_type == 'application/json'
# searches with lexically malformed order bys
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'name', 'regex', '.'],
'order_by': ['FormSearch', 'foo', 'desc']
}
})
response = self.app.post(url('search_post'),
json_query,
self.json_headers,
self.extra_environ_admin,
status=400)
resp = response.json_body
assert resp['errors'][
'FormSearch.foo'] == 'Searching on FormSearch.foo is not permitted'
assert resp['errors'][
'OrderByError'] == 'The provided order by expression was invalid.'
json_query = json.dumps({
'query': {
'filter': ['FormSearch', 'name', 'regex', '.'],
'order_by': ['Foo', 'id', 'desc']
}
})
response = self.app.post(url('search_post'),
json_query,
self.json_headers,
self.extra_environ_admin,
status=400)
resp = response.json_body
assert resp['errors'][
'Foo'] == 'Searching the FormSearch model by joining on the Foo model is not possible'
assert resp['errors'][
'Foo.id'] == 'Searching on Foo.id is not permitted'
assert resp['errors'][
'OrderByError'] == 'The provided order by expression was invalid.'
def test_index(self):
"""Tests that GET & SEARCH /corpusbackups behave correctly.
"""
dbsession = self.dbsession
db = DBUtils(dbsession, self.settings)
tag = old_models.Tag()
tag.name = 'random tag name'
dbsession.add(tag)
dbsession.flush()
tag_id = tag.id
dbsession.commit()
# Add 10 forms and use them to generate a valid value for ``test_corpus_content``
def create_form_from_index(index):
form = old_models.Form()
form.transcription = 'Form %d' % index
translation = old_models.Translation()
translation.transcription = 'Translation %d' % index
form.translation = translation
return form
forms = [create_form_from_index(i) for i in range(1, 10)]
dbsession.add_all(forms)
dbsession.commit()
forms = db.get_forms()
half_forms = forms[:5]
form_ids = [form.id for form in forms]
half_form_ids = [form.id for form in half_forms]
test_corpus_content = ','.join(map(str, form_ids))
test_corpus_half_content = ','.join(map(str, half_form_ids))
# Create a form search model
query = {'filter': ['Form', 'transcription', 'regex', '[a-zA-Z]{3,}']}
params = json.dumps({
'name': 'form search',
'description': 'This one\'s worth saving!',
'search': query
})
response = self.app.post(fs_url('create'), params, self.json_headers,
self.extra_environ_admin)
resp = response.json_body
form_search_id = resp['id']
# Generate some valid corpus creation input parameters.
params = self.corpus_create_params.copy()
params.update({
'name': 'Corpus',
'description': 'Covers a lot of the data.',
'content': test_corpus_content
})
params = json.dumps(params)
# Attempt to create a corpus as a viewer and expect to fail
response = self.app.post(crps_url('create'),
params,
self.json_headers,
self.extra_environ_view,
status=403)
resp = response.json_body
assert resp[
'error'] == 'You are not authorized to access this resource.'
assert response.content_type == 'application/json'
# Successfully create a corpus as the admin
assert os.listdir(self.corpora_path) == []
original_corpus_count = dbsession.query(Corpus).count()
response = self.app.post(crps_url('create'), params, self.json_headers,
self.extra_environ_admin)
resp = response.json_body
corpus_id = resp['id']
new_corpus_count = dbsession.query(Corpus).count()
corpus = dbsession.query(Corpus).get(corpus_id)
corpus_form_ids = sorted([f.id for f in corpus.forms])
corpus_dir = os.path.join(self.corpora_path, 'corpus_%d' % corpus_id)
corpus_dir_contents = os.listdir(corpus_dir)
assert new_corpus_count == original_corpus_count + 1
assert resp['name'] == 'Corpus'
assert resp['description'] == 'Covers a lot of the data.'
assert corpus_dir_contents == []
assert response.content_type == 'application/json'
assert resp['content'] == test_corpus_content
assert corpus_form_ids == sorted(form_ids)
# Update the corpus as the contributor -- now we should have one backup
dbsession.expire(corpus)
params = self.corpus_create_params.copy()
params.update({
'name': 'Corpus',
'description': 'Covers a little less data.',
'content': test_corpus_half_content
})
params = json.dumps(params)
response = self.app.put(crps_url('update', id=corpus_id), params,
self.json_headers, self.extra_environ_contrib)
resp = response.json_body
corpus_count = new_corpus_count
new_corpus_count = dbsession.query(Corpus).count()
corpus = dbsession.query(Corpus).get(corpus_id)
corpus_form_ids = sorted([f.id for f in corpus.forms])
assert new_corpus_count == corpus_count
assert resp['name'] == 'Corpus'
assert resp['description'] == 'Covers a little less data.'
assert response.content_type == 'application/json'
assert resp['content'] == test_corpus_half_content
assert corpus_form_ids == sorted(half_form_ids)
# Update the corpus again -- now we should have two backups
sleep(1)
params = self.corpus_create_params.copy()
params.update({
'name': 'Corpus',
'description': 'Covers a little less data.',
'content': test_corpus_half_content,
'tags': [tag_id]
})
params = json.dumps(params)
response = self.app.put(crps_url('update', id=corpus_id), params,
self.json_headers, self.extra_environ_admin)
resp = response.json_body
corpus_count = new_corpus_count
new_corpus_count = dbsession.query(Corpus).count()
corpus = dbsession.query(Corpus).get(corpus_id)
corpus_form_ids = sorted([f.id for f in corpus.forms])
assert new_corpus_count == corpus_count
assert resp['name'] == 'Corpus'
assert resp['description'] == 'Covers a little less data.'
assert response.content_type == 'application/json'
assert resp['content'] == test_corpus_half_content
assert corpus_form_ids == sorted(half_form_ids)
all_corpus_backups = dbsession.query(CorpusBackup).order_by(
CorpusBackup.id).all()
all_corpus_backup_ids = [cb.id for cb in all_corpus_backups]
all_corpus_backup_descriptions = [
cb.description for cb in all_corpus_backups
]
# Now request the corpus backups as either the contributor or the viewer and
# expect to get them all.
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=self.extra_environ_contrib)
resp = response.json_body
assert len(resp) == 2
assert response.content_type == 'application/json'
assert resp[0]['modifier']['role'] == 'administrator'
assert resp[1]['modifier']['role'] == 'contributor'
# The admin should get them all too.
response = self.app.get(url('index'),
headers=self.json_headers,
extra_environ=self.extra_environ_view)
resp = response.json_body
assert len(resp) == 2
assert [cb['id'] for cb in resp] == all_corpus_backup_ids
# Test the paginator GET params.
paginator = {'items_per_page': 1, 'page': 2}
response = self.app.get(url('index'),
paginator,
headers=self.json_headers,
extra_environ=self.extra_environ_admin)
resp = response.json_body
assert len(resp['items']) == 1
assert resp['paginator']['count'] == 2
assert response.content_type == 'application/json'
assert resp['items'][0]['id'] == all_corpus_backup_ids[1]
# Test the order_by GET params.
order_by_params = {
'order_by_model': 'CorpusBackup',
'order_by_attribute': 'id',
'order_by_direction': 'desc'
}
response = self.app.get(url('index'),
order_by_params,
headers=self.json_headers,
extra_environ=self.extra_environ_admin)
resp = response.json_body
result_set = list(reversed(all_corpus_backup_ids))
assert [cb['id'] for cb in resp] == result_set
# Test the order_by *with* paginator.
params = {
'order_by_model': 'CorpusBackup',
'order_by_attribute': 'id',
'order_by_direction': 'desc',
'items_per_page': 1,
'page': 1
}
response = self.app.get(url('index'),
params,
headers=self.json_headers,
extra_environ=self.extra_environ_admin)
resp = response.json_body
assert result_set[0] == resp['items'][0]['id']
# Now test the show action:
# Get a specific corpus backup.
response = self.app.get(url('show', id=all_corpus_backup_ids[0]),
headers=self.json_headers,
extra_environ=self.extra_environ_admin)
resp = response.json_body
assert resp['description'] == 'Covers a lot of the data.'
assert resp['content'] == test_corpus_content
assert response.content_type == 'application/json'
# A nonexistent cb id will return a 404 error
response = self.app.get(url('show', id=100987),
headers=self.json_headers,
extra_environ=self.extra_environ_view,
status=404)
resp = response.json_body
assert resp['error'] == 'There is no corpus backup with id 100987'
assert response.content_type == 'application/json'
# Test the search action
add_SEARCH_to_web_test_valid_methods()
# A search on corpus backup titles using POST /corpusbackups/search
json_query = json.dumps({
'query': {
'filter': ['CorpusBackup', 'description', 'like', '%less%']
}
})
response = self.app.post(url('search_post'),
json_query,
self.json_headers,
extra_environ=self.extra_environ_admin)
resp = response.json_body
result_set = [
name for name in all_corpus_backup_descriptions if 'less' in name
]
assert len(resp) == len(result_set) == 1
assert resp[0]['description'] == result_set[0]
assert response.content_type == 'application/json'
# A search on corpus backup titles using SEARCH /corpusbackups
json_query = json.dumps({
'query': {
'filter': ['CorpusBackup', 'description', 'like', '%less%']
}
})
response = self.app.request(url('search'),
method='SEARCH',
body=json_query.encode('utf8'),
headers=self.json_headers,
environ=self.extra_environ_admin)
resp = response.json_body
assert len(resp) == len(result_set) == 1
assert resp[0]['description'] == result_set[0]
assert response.content_type == 'application/json'
# Attempting to call edit/new/create/delete/update on a read-only resource
# will return a 404 response
response = self.app.get(url('edit', id=2232),
status=404,
extra_environ=self.extra_environ_admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.get(url('new', id=2232),
status=404,
extra_environ=self.extra_environ_admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.post(url('create'),
status=404,
extra_environ=self.extra_environ_admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.put(url('update', id=2232),
status=404,
extra_environ=self.extra_environ_admin)
assert response.json_body['error'] == 'This resource is read-only.'
response = self.app.delete(url('delete', id=2232),
status=404,
extra_environ=self.extra_environ_admin)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
def test_index(self):
"""Tests that GET & SEARCH /languages behave correctly.
NOTE: during testing, the language table contains only 8 records.
"""
dbsession = self.dbsession
db = DBUtils(dbsession, self.settings)
languages = dbsession.query(old_models.Language).all()
# GET the languages
response = self.app.get(url('index'), headers=self.json_headers, extra_environ=self.extra_environ_view)
resp = response.json_body
assert len(resp) == len(languages)
assert response.content_type == 'application/json'
# Test the paginator GET params.
paginator = {'items_per_page': 2, 'page': 2}
response = self.app.get(url('index'), paginator, headers=self.json_headers,
extra_environ=self.extra_environ_view)
resp = response.json_body
assert len(resp['items']) == 2
assert resp['items'][0]['Part2B'] == languages[2].Part2B
assert response.content_type == 'application/json'
# Test the order_by GET params.
order_by_params = {'order_by_model': 'Language', 'order_by_attribute': 'Ref_Name',
'order_by_direction': 'desc'}
response = self.app.get(url('index'), order_by_params,
headers=self.json_headers, extra_environ=self.extra_environ_view)
resp = response.json_body
result_set = sorted(languages, key=lambda l: l.Ref_Name, reverse=True)
assert [l['Id'] for l in resp] == [l.Id for l in result_set]
# Test the order_by *with* paginator.
params = {'order_by_model': 'Language', 'order_by_attribute': 'Ref_Name',
'order_by_direction': 'desc', 'items_per_page': 1, 'page': 3}
response = self.app.get(url('index'), params,
headers=self.json_headers, extra_environ=self.extra_environ_admin)
resp = response.json_body
assert result_set[2].Ref_Name == resp['items'][0]['Ref_Name']
assert response.content_type == 'application/json'
# Now test the show action:
response = self.app.get(url('show', id=languages[4].Id),
headers=self.json_headers, extra_environ=self.extra_environ_view)
resp = response.json_body
assert resp['Ref_Name'] == languages[4].Ref_Name
assert response.content_type == 'application/json'
# A nonexistent language Id will return a 404 error
response = self.app.get(url('show', id=100987),
headers=self.json_headers, extra_environ=self.extra_environ_view, status=404)
resp = response.json_body
assert resp['error'] == 'There is no language with Id 100987'
assert response.content_type == 'application/json'
# Test the search action
add_SEARCH_to_web_test_valid_methods()
# A search on language transcriptions using POST /languages/search
json_query = json.dumps({'query': {'filter':
['Language', 'Ref_Name', 'like', '%m%']}})
response = self.app.post(url('search_post'), json_query,
self.json_headers, self.extra_environ_view)
resp = response.json_body
result_set = [l for l in languages if 'm' in l.Ref_Name]
assert resp
assert set([l['Id'] for l in resp]) == set([l.Id for l in result_set])
assert response.content_type == 'application/json'
# A search on language Ref_Name using SEARCH /languages
json_query = json.dumps({'query': {'filter':
['Language', 'Ref_Name', 'like', '%l%']}})
response = self.app.request(url('search'), method='SEARCH', body=json_query.encode('utf8'),
headers=self.json_headers, environ=self.extra_environ_view)
resp = response.json_body
result_set = [l for l in languages if 'l' in l.Ref_Name]
assert resp
assert len(resp) == len(result_set)
assert set([l['Id'] for l in resp]) == set([l.Id for l in result_set])
# 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', id=2232), extra_environ=self.extra_environ_contrib,
status=404)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
response = self.app.get(
url('new', id=2232),
extra_environ=self.extra_environ_contrib, status=404)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
response = self.app.post(
url('create'), extra_environ=self.extra_environ_contrib,
status=404)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
response = self.app.put(
url('update', id=2232), extra_environ=self.extra_environ_contrib,
status=404)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
response = self.app.delete(
url('delete', id=2232), extra_environ=self.extra_environ_contrib,
status=404)
assert response.json_body['error'] == 'This resource is read-only.'
assert response.content_type == 'application/json'
