Select Git revision
test_converters.py
Code owners
Assign users and groups as approvers for specific file changes. Learn more.
test_tool.py 15.53 KiB
#!/bin/python
# Tests for the tool using pytest
# Adapted from check-sfs
# A. Schlemmer, 06/2021
from newcrawler import Crawler
from newcrawler.converters import MarkdownFileConverter
from newcrawler.structure_elements import File, DictTextElement, DictListElement
from newcrawler.identifiable_adapters import IdentifiableAdapter, LocalStorageIdentifiableAdapter
from functools import partial
from copy import deepcopy
from unittest.mock import MagicMock, Mock
from os.path import join, dirname, basename
import yaml
import caosdb as db
from caosdb.apiutils import compare_entities
import pytest
from pytest import raises
def rfp(*pathcomponents):
"""
Return full path.
Shorthand convenience function.
"""
return join(dirname(__file__), *pathcomponents)
def dircheckstr(*pathcomponents):
"""
Return the debug tree identifier for a given path.
"""
return "newcrawler.structure_elements.Directory: " + basename(join(*pathcomponents)) + ", " + rfp("test_directories", "examples_article", *pathcomponents)
@pytest.fixture
def crawler():
crawler = Crawler(debug=True)
crawler.crawl_directory(rfp("test_directories", "examples_article"),
rfp("scifolder_cfood.yml"))
return crawler
@pytest.fixture
def ident(crawler):
ident = LocalStorageIdentifiableAdapter()
crawler.identifiableAdapter = ident
# The records.xml file is constructed as follows:
# To a full run of the crawler, resolve all identifiables and insert all resulting entities.
# See: test-setup/datamodel/generate_test_data.py for details.
ident.restore_state(rfp("records.xml"))
ident.register_identifiable(
"Person", db.RecordType()
.add_parent(name="Person")
.add_property(name="first_name")
.add_property(name="last_name"))
ident.register_identifiable(
"Measurement", db.RecordType()
.add_parent(name="Measurement")
.add_property(name="identifier")
.add_property(name="date")
.add_property(name="project"))
ident.register_identifiable(
"Project", db.RecordType()
.add_parent(name="Project")
.add_property(name="date")
.add_property(name="identifier"))
return ident
def test_crawler(crawler):
subd = crawler.debug_tree[dircheckstr("DataAnalysis")]
subc = crawler.debug_metadata["copied"][dircheckstr("DataAnalysis")]
assert len(subd) == 2
assert len(subd[0]) == 0
assert len(subd[1]) == 0
assert len(subc) == 2
assert len(subc[0]) == 0
assert len(subc[1]) == 0
subd = crawler.debug_tree[dircheckstr("DataAnalysis", "2020_climate-model-predict")]
subc = crawler.debug_metadata["copied"][dircheckstr(
"DataAnalysis", "2020_climate-model-predict")]
assert len(subd[1]) == 1
assert len(subd[1]["Project"].get_parents()) == 1
assert subd[1]["Project"].get_parents()[0].name == "Project"
assert subd[1]["Project"].get_property("date").value == "2020"
assert subd[1]["Project"].get_property("identifier").value == "climate-model-predict"
assert len(subd[0]) == 3
assert subd[0]["date"] == "2020"
assert subd[0]["identifier"] == "climate-model-predict"
assert subd[0]["Project"].__class__ == db.Record
# Check the copy flags for the first level in the hierarchy:
assert len(subc[0]) == 3
assert len(subc[1]) == 1
assert subc[1]["Project"] is False
assert subc[0]["Project"] is False
assert subc[0]["date"] is False
assert subc[0]["identifier"] is False
subd = crawler.debug_tree[dircheckstr("DataAnalysis",
"2020_climate-model-predict",
"2020-02-08_prediction-errors")]
subc = crawler.debug_metadata["copied"][dircheckstr("DataAnalysis",
"2020_climate-model-predict",
"2020-02-08_prediction-errors")]
assert len(subd[0]) == 4
assert subd[0]["date"] == "2020-02-08"
assert subd[0]["identifier"] == "prediction-errors"
assert subd[0]["Project"].__class__ == db.Record
assert subd[0]["Measurement"].__class__ == db.Record
assert len(subd[1]) == 2
assert len(subd[1]["Project"].get_parents()) == 1
assert subd[1]["Project"].get_parents()[0].name == "Project"
assert subd[1]["Project"].get_property("date").value == "2020"
assert subd[1]["Project"].get_property("identifier").value == "climate-model-predict"
assert len(subd[1]["Measurement"].get_parents()) == 1
assert subd[1]["Measurement"].get_parents()[0].name == "Measurement"
assert subd[1]["Measurement"].get_property("date").value == "2020-02-08"
assert subd[1]["Measurement"].get_property("identifier").value == "prediction-errors"
assert subd[1]["Measurement"].get_property("project").value != "$Project"
assert subd[1]["Measurement"].get_property("project").value.__class__ == db.Record
assert subd[1]["Measurement"].get_property("project").value == subd[0]["Project"]
# Check the copy flags for the second level in the hierarchy:
assert subc[1]["Project"] is True
assert subc[0]["Project"] is True
assert subc[1]["Measurement"] is False
assert subc[0]["Measurement"] is False
assert subc[0]["date"] is False
assert subc[0]["identifier"] is False
def test_markdown_converter():
test_readme = File("README.md", rfp(
"test_directories", "examples_article", "DataAnalysis",
"2020_climate-model-predict", "2020-02-08_prediction-errors", "README.md"))
converter = MarkdownFileConverter({
"match": "(.*)"
}, "TestMarkdownFileConverter")
m = converter.match(File("test_tool.py", rfp(
"test_tool.py")))
assert m is None
m = converter.match(test_readme)
assert m is not None
assert m.__class__ == dict
assert len(m) == 0
converter = MarkdownFileConverter({
"match": "README.md"
}, "TestMarkdownFileConverter")
m = converter.match(test_readme)
assert m is not None
assert len(m) == 0
children = converter.create_children(None, test_readme)
assert len(children) == 5
assert children[1].__class__ == DictTextElement
assert children[1].name == "description"
assert children[1].value.__class__ == str
assert children[0].__class__ == DictTextElement
assert children[0].name == "responsible"
assert children[0].value.__class__ == str
test_readme2 = File("README.md", rfp("test_directories", "examples_article",
"ExperimentalData", "2020_SpeedOfLight", "2020-01-01_TimeOfFlight", "README.md"))
m = converter.match(test_readme2)
assert m is not None
assert len(m) == 0
children = converter.create_children(None, test_readme2)
assert len(children) == 2
assert children[1].__class__ == DictTextElement
assert children[1].name == "description"
assert children[1].value.__class__ == str
assert children[0].__class__ == DictListElement
assert children[0].name == "responsible"
assert children[0].value.__class__ == list
# def prepare_test_record_file():
# ident = LocalStorageIdentifiableAdapter()
# crawler = Crawler(debug=True, identifiableAdapter=ident)
# crawler.crawl_directory(rfp("test_directories", "examples_article"),
# rfp("scifolder_cfood.yml"))
# # clean record list:
# recordlist = ident.get_records()
# for i in range(len(recordlist)-1, 1, -1):
# if recordlist[i].parents[0].name == "Person":
# del recordlist[i]
# ident.store_state(rfp("records.xml"))
def test_ambigious_records(crawler, ident):
ident.get_records().clear()
ident.get_records().extend(crawler.updateList)
r = ident.get_records()
id_r0 = ident.get_identifiable(r[0])
with raises(RuntimeError, match=".*unambigiously.*"):
ident.retrieve_identified_record(id_r0)
def test_crawler_update_list(crawler, ident):
# If the following assertions fail, that is a hint, that the test file records.xml is
# incorrect:
assert len(ident.get_records()) == 18
assert len([r for r in ident.get_records() if r.parents[0].name == "Person"]) == 5
assert len([r for r in ident.get_records() if r.parents[0].name == "Measurement"]) == 11
assert len([r for r in ident.get_records() if r.parents[0].name == "Project"]) == 2
# The crawler contains lots of duplicates, because identifiables have not been resolved yet:
assert len(ident.get_records()) != len(crawler.updateList)
# Check consistency:
# Check whether identifiables retrieved from current identifiable store return the same results.
# take the first person in the list of records:
for r in ident.get_records():
if r.parents[0].name == "Person":
r_cur = r
break
id_r0 = ident.get_identifiable(r_cur)
assert r_cur.parents[0].name == id_r0.parents[0].name
assert r_cur.get_property("first_name").value == id_r0.get_property("first_name").value
assert r_cur.get_property("last_name").value == id_r0.get_property("last_name").value
assert len(r_cur.parents) == 1
assert len(id_r0.parents) == 1
assert len(r_cur.properties) == 2
assert len(id_r0.properties) == 2
idr_r0_test = ident.retrieve_identified_record(id_r0)
idr_r0 = ident.retrieve_identifiable(r_cur)
assert idr_r0 == idr_r0_test
# take the first measurement in the list of records:
for r in ident.get_records():
if r.parents[0].name == "Measurement":
r_cur = r
break
id_r1 = ident.get_identifiable(r_cur)
assert r_cur.parents[0].name == id_r1.parents[0].name
assert r_cur.get_property("identifier").value == id_r1.get_property("identifier").value
assert r_cur.get_property("date").value == id_r1.get_property("date").value
assert r_cur.get_property("project").value == id_r1.get_property("project").value
assert len(r_cur.parents) == 1
assert len(id_r1.parents) == 1
assert len(r_cur.properties) == 5
assert len(id_r1.properties) == 3
idr_r1_test = ident.retrieve_identified_record(id_r1)
idr_r1 = ident.retrieve_identifiable(r_cur)
assert idr_r1 == idr_r1_test
assert idr_r1 != idr_r0
assert idr_r1_test != idr_r0_test
assert len(idr_r1.properties) == 5
assert r_cur.get_property("responsible").value == idr_r1.get_property("responsible").value
assert r_cur.get_property("description").value == idr_r1.get_property("description").value
# test whether compare_entites function works in this context:
comp = compare_entities(r_cur, id_r1)
assert len(comp[0]["parents"]) == 0
assert len(comp[1]["parents"]) == 0
assert len(comp[0]["properties"]) == 2
assert len(comp[1]["properties"]) == 0
assert "responsible" in comp[0]["properties"]
assert "description" in comp[0]["properties"]
comp = compare_entities(r_cur, idr_r1)
assert len(comp[0]["parents"]) == 0
assert len(comp[1]["parents"]) == 0
assert len(comp[0]["properties"]) == 0
assert len(comp[1]["properties"]) == 0
insl, updl = crawler.synchronize()
assert len(insl) == 0
assert len(updl) == 0
def test_identifiable_update(crawler, ident):
# change one value in updateList and then run the synchronization:
meas = [r for r in crawler.updateList if r.parents[0].name == "Measurement"][0]
meas.get_property("responsible").value = []
insl, updl = crawler.synchronize()
assert len(updl) == 1
def test_identifiable_update2(crawler, ident):
# change one unit in updateList and then run the synchronization:
meas = [r for r in crawler.updateList if r.parents[0].name == "Measurement"][0]
meas.get_property("description").unit = "cm"
insl, updl = crawler.synchronize()
assert len(updl) == 1
def test_identifiable_update3(crawler, ident):
# change values of multiple records in updateList and then run the synchronization:
meas = [r for r in crawler.updateList if r.parents[0].name == "Measurement"]
meas[0].get_property("responsible").value = []
meas[3].get_property("responsible").value = []
insl, updl = crawler.synchronize()
assert len(updl) == 2
def test_identifiable_adapter():
query = IdentifiableAdapter.create_query_for_identifiable(
db.Record().add_parent("Person")
.add_property("first_name", value="A")
.add_property("last_name", value="B"))
assert query.lower() == "find record person with 'first_name'='a' and 'last_name'='b' "
@pytest.mark.xfail
def test_identifiable_adapter_no_identifiable(crawler, ident):
del ident._registered_identifiables["Person"]
insl, updl = crawler.synchronize()
assert len(updl) == 0
pers = [r for r in crawler.updateList if r.parents[0].name == "Person"]
# All persons are inserted, because they are not identifiable:
assert len(insl) == len(pers)
def test_provenance_debug_data(crawler):
crawler.save_debug_data(rfp("provenance.yml"))
with open(rfp("provenance.yml"), "r") as f:
provenance = yaml.load(f, Loader=yaml.SafeLoader)
pr = provenance["provenance"]
def check_key_count(prefix):
return sum([1 for key in pr.keys() if key.startswith(prefix)])
assert check_key_count("Measurement") == 11
assert check_key_count("Project") == 5
assert check_key_count("Person") == 14
def test_split_into_inserts_and_updates(crawler):
# Try trivial argument
crawler.split_into_inserts_and_updates([])
# simulate remote server content by using the names to identify records
def base_mocked_lookup(rec, known):
if rec.name in known:
return known[rec.name]
else:
return None
cache = []
def trivial_cache_loockup(stuff):
print("current cache", cache)
if stuff.name in cache:
return stuff
else:
return None
def trivial_cache_add(stuff):
cache.append(stuff.name)
crawler.get_identifiable_from_local_cache = Mock(side_effect=trivial_cache_loockup)
crawler.add_identifiable_to_local_cache = Mock(side_effect=trivial_cache_add)
crawler.copy_attributes = Mock()
# a record that is found remotely and should be added to the update list and one that is not
# found and should be added to the insert one
remote_known = {"A": db.Record(id=1111, name="A")}
entlist = [db.Record(name="A"), db.Record(name="B")]
crawler.identifiableAdapter.retrieve_identifiable = Mock(side_effect=partial(
base_mocked_lookup, known=remote_known))
insert, update = crawler.split_into_inserts_and_updates(entlist)
print(crawler.identifiableAdapter.retrieve_identifiable.call_args_list)
print(entlist)
crawler.identifiableAdapter.retrieve_identifiable.assert_any_call(entlist[0])
crawler.identifiableAdapter.retrieve_identifiable.assert_any_call(entlist[1])
assert len(insert) == 1
assert insert[0].name == "B"
assert len(update) == 1
assert update[0].name == "A"
# reset cache
cache.clear()
# try it with a reference
a = db.Record(name="A")
b = db.Record(name="B")
b.add_property("A", a)
entlist = [a, b]
crawler.identifiableAdapter.retrieve_identifiable = Mock(side_effect=partial(
base_mocked_lookup, known=remote_known))
insert, update = crawler.split_into_inserts_and_updates(entlist)
assert len(insert) == 1
assert insert[0].name == "B"
assert len(update) == 1
assert update[0].name == "A"
# reset cache
cache.clear()
# try circular
a = db.Record(name="A")
b = db.Record(name="B")
b.add_property("A", a)
a.add_property("B", b)
entlist = [a, b]
with raises(RuntimeError):
crawler.split_into_inserts_and_updates(entlist)