From 0c84af603894049dd8edd95da18d8990ab0516d1 Mon Sep 17 00:00:00 2001 From: Martin Czygan Date: Fri, 5 Nov 2021 17:19:07 +0100 Subject: turn "match_release_fuzzy" into a class Goal of this refactoring was to make the matching process a bit more configurable by using a class and a cascade of queries. For a limited test set: `FuzzyReleaseMatcher.match` is works the same as `match_release_fuzzy`. --- tests/test_matching.py | 123 ++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 116 insertions(+), 7 deletions(-) (limited to 'tests/test_matching.py') diff --git a/tests/test_matching.py b/tests/test_matching.py index ad971a5..ca94c2a 100644 --- a/tests/test_matching.py +++ b/tests/test_matching.py @@ -1,13 +1,14 @@ +import collections import logging import warnings import elasticsearch import pytest import requests -from fatcat_openapi_client import ReleaseEntity +from fatcat_openapi_client import ReleaseEntity, ReleaseContrib -from fuzzycat.entities import entity_from_dict -from fuzzycat.matching import anything_to_entity, match_release_fuzzy +from fuzzycat.entities import entity_from_dict, entity_from_json +from fuzzycat.matching import anything_to_entity, match_release_fuzzy, FuzzyReleaseMatcher warnings.filterwarnings( "ignore") # InsecureRequestWarning: Unverified HTTPS request is being made to host ... @@ -18,6 +19,9 @@ from fatcat_openapi_client import ReleaseEntity import pytest import elasticsearch import logging +import yaml +import glob +import json logger = logging.getLogger('test_matching') logger.setLevel(logging.DEBUG) @@ -40,19 +44,35 @@ def is_reachable(url, timeout=3): return False +def yaml_to_cases(klass, + files="tests/files/fuzzy_release_match_release_exact_title_exact_contrib/*.yaml"): + """ + Turn yaml files into a collection of named tuple test cases. The glob is + relative to the project root (i.e. where you usually run `pytest` from). + """ + cases = [] + for path in glob.glob(files): + with open(path) as f: + doc = yaml.load(f, Loader=yaml.Loader) + cases.append(klass(**doc)) + return cases + + @pytest.fixture def es_client(): return elasticsearch.Elasticsearch([FATCAT_SEARCH_URL]) -@pytest.mark.skipif( - is_not_reachable(FATCAT_SEARCH_URL), - reason="{} not reachable, use e.g. FUZZYCAT_FATCAT_SEARCH_URL=localhost:9200 to override". - format(FATCAT_SEARCH_URL)) +# @pytest.mark.skipif( +# is_not_reachable(FATCAT_SEARCH_URL), +# reason="{} not reachable, use e.g. FUZZYCAT_FATCAT_SEARCH_URL=localhost:9200 to override". +# format(FATCAT_SEARCH_URL)) def test_match_release_fuzzy(es_client, caplog): """ This test is tied to the current index contents, so if that changes, this test may fail as well. + + Note: Deprecated. We want to get rid of this. """ cases = ( ("wtv64ahbdzgwnan7rllwr3nurm", 1), @@ -106,3 +126,92 @@ def test_match_release_fuzzy(es_client, caplog): logging.info("[{}] given title '{}', found {}, {}".format(i, entity.title, len(result), [v.title for v in result])) assert len(result) == count, doc + + +def test_matcher_match_release(es_client, caplog): + cases = ( + ("wtv64ahbdzgwnan7rllwr3nurm", 1), + ("eqcgtpav3na5jh56o5vjsvb4ei", 1), + ) + matcher = FuzzyReleaseMatcher(es=es_client, size=5) + for i, (ident, count) in enumerate(cases): + entity = anything_to_entity(ident, ReleaseEntity) + result = matcher.match(entity) + logger.info("[{}] given {}, found {}".format(i, entity.title, len(result))) + assert len(result) == count + + # Partial data. + cases = ( + ({ + "title": "digital libraries", + "ext_ids": {} + }, 5), + ({ + "title": "unlikelytitle", + "ext_ids": {} + }, 0), + ({ + "title": "Imminent dystopia", + "ext_ids": {} + }, 5), + ({ + "title": "", + "contribs": [{ + "raw_name": "Aristoteles" + }], + "ext_ids": {} + }, 5), + # ({ + # "title": "Letter", + # "contribs": [{"raw_name": "Claudel"}], + # "ext_ids": {} + # }, 1), + # ({ + # "title": "The Future of Digital Scholarship", + # "contribs": [{ + # "raw_name": "Costantino Thanos" + # }], + # "ext_ids": {} + # }, 5), + ) + for i, (doc, count) in enumerate(cases): + entity = entity_from_dict(doc, ReleaseEntity) + result = matcher.match(entity) + with caplog.at_level(logging.INFO): + logging.info("[{}] given title '{}', found {}, {}".format(i, entity.title, len(result), + [v.title for v in result])) + assert len(result) == count, doc + + +def test_fuzzy_release_matcher_match_release_by_id(es_client, caplog): + matcher = FuzzyReleaseMatcher(es=es_client) + cases = ( + ("wtv64ahbdzgwnan7rllwr3nurm", 1), + ("eqcgtpav3na5jh56o5vjsvb4ei", 1), + ) + for i, (ident, count) in enumerate(cases): + entity = anything_to_entity(ident, ReleaseEntity) + result = matcher.match_release_by_id(entity) + assert len(result) == count + + +def test_fuzzy_release_match_release_exact_title_exact_contrib(es_client, caplog): + matcher = FuzzyReleaseMatcher(es=es_client) + Case = collections.namedtuple("Case", "title date input expected") + cases = yaml_to_cases( + Case, "tests/files/fuzzy_release_match_release_exact_title_exact_contrib/*.yaml") + for i, c in enumerate(cases): + entity = entity_from_json(c.input, ReleaseEntity) + result = matcher.match_release_exact_title_exact_contrib(entity) + assert len(result) == c.expected, "[{}] {}".format(c.title, c.input) + + +def test_fuzzy_release_match_release_exact_title_partial_contrib(es_client, caplog): + matcher = FuzzyReleaseMatcher(es=es_client) + Case = collections.namedtuple("Case", "title date input jaccard_index_threshold expected") + cases = yaml_to_cases( + Case, "tests/files/fuzzy_release_match_release_exact_title_partial_contrib/*.yaml") + for i, c in enumerate(cases): + entity = entity_from_json(c.input, ReleaseEntity) + result = matcher.match_release_exact_title_partial_contrib(entity) + assert len(result) == c.expected, "[{}] {}".format(c.title, c.input) -- cgit v1.2.3