Merge branch '18574-pysdk-vocabulary-support'. Closes #18574
authorLucas Di Pentima <lucas.dipentima@curii.com>
Fri, 4 Mar 2022 20:33:40 +0000 (17:33 -0300)
committerLucas Di Pentima <lucas.dipentima@curii.com>
Fri, 4 Mar 2022 20:33:40 +0000 (17:33 -0300)
Arvados-DCO-1.1-Signed-off-by: Lucas Di Pentima <lucas.dipentima@curii.com>

doc/api/index.html.textile.liquid
doc/sdk/python/cookbook.html.textile.liquid
sdk/python/arvados/api.py
sdk/python/arvados/util.py
sdk/python/arvados/vocabulary.py [new file with mode: 0644]
sdk/python/tests/test_vocabulary.py [new file with mode: 0644]
services/api/app/controllers/arvados/v1/schema_controller.rb

index 8586a166d35eaa94d8e330a31748ccd9d87a2caf..3d69d02ea92379fcb11c17d78cae442112e8df84 100644 (file)
@@ -20,6 +20,10 @@ h2. Exported configuration
 
 The Controller exposes a subset of the cluster's configuration and makes it available to clients in JSON format. This public config includes valuable information like several service's URLs, timeout settings, etc. and it is available at @/arvados/v1/config@, for example @https://{{ site.arvados_api_host }}/arvados/v1/config@. The new Workbench is one example of a client using this information, as it's a client-side application and doesn't have access to the cluster's config file.
 
+h2. Exported vocabulary definition
+
+When configured, the Controller also exports the "metadata vocabulary definition":{{site.baseurl}}/admin/metadata-vocabulary.html in JSON format. This functionality is useful for clients like Workbench2 and the Python SDK to provide "identifier to human-readable labels" translations facilities for reading and writing objects on the system. This is available at @/arvados/v1/vocabulary@, for example @https://{{ site.arvados_api_host }}/arvados/v1/vocabulary@.
+
 h2. Workbench examples
 
 Many Arvados Workbench pages, under the *Advanced* tab, provide examples of API and SDK use for accessing the current resource .
index eda6563d7a1de99602eecada47eab021807a1782..f3186ebbb6d76d66221860f669960cb9737688eb 100644 (file)
@@ -298,3 +298,37 @@ api = arvados.api()
 for c in arvados.util.keyset_list_all(api.collections().list, filters=[["name", "like", "%sample123%"]]):
     print("got collection " + c["uuid"])
 {% endcodeblock %}
+
+h2. Querying the vocabulary definition
+
+The Python SDK provides facilities to interact with the "active metadata vocabulary":{{ site.baseurl }}/admin/metadata-vocabulary.html in the system. The developer can do key and value lookups in a case-insensitive manner:
+
+{% codeblock as python %}
+from arvados import api, vocabulary
+voc = vocabulary.load_vocabulary(api('v1'))
+
+[k.identifier for k in set(voc.key_aliases.values())]
+# Example output: ['IDTAGCOLORS', 'IDTAGFRUITS', 'IDTAGCOMMENT', 'IDTAGIMPORTANCES', 'IDTAGCATEGORIES', 'IDTAGSIZES', 'IDTAGANIMALS']
+voc['IDTAGSIZES'].preferred_label
+# Example output: 'Size'
+[v.preferred_label for v in set(voc['size'].value_aliases.values())]
+# Example output: ['S', 'M', 'L', 'XL', 'XS']
+voc['size']['s'].aliases
+# Example output: ['S', 'small']
+voc['size']['Small'].identifier
+# Example output: 'IDVALSIZES2'
+{% endcodeblock %}
+
+h2. Translating between vocabulary identifiers and labels
+
+Client software might need to present properties to the user in a human-readable form or take input from the user without requiring them to remember identifiers. For these cases, there're a couple of conversion methods that take a dictionary as input like this:
+
+{% codeblock as python %}
+from arvados import api, vocabulary
+voc = vocabulary.load_vocabulary(api('v1'))
+
+voc.convert_to_labels({'IDTAGIMPORTANCES': 'IDVALIMPORTANCES1'})
+# Example output: {'Importance': 'Critical'}
+voc.convert_to_identifiers({'creature': 'elephant'})
+# Example output: {'IDTAGANIMALS': 'IDVALANIMALS3'}
+{% endcodeblock %}
\ No newline at end of file
index 88596211d4e06b028e6974df2d31a9eee1cf7e19..e0d1c50f03a10fe6f3c0e0e5f45df4cb0f57aec9 100644 (file)
@@ -253,6 +253,7 @@ def api(version=None, cache=True, host=None, token=None, insecure=False,
     svc.insecure = insecure
     svc.request_id = request_id
     svc.config = lambda: util.get_config_once(svc)
+    svc.vocabulary = lambda: util.get_vocabulary_once(svc)
     kwargs['http'].max_request_size = svc._rootDesc.get('maxRequestSize', 0)
     kwargs['http'].cache = None
     kwargs['http']._request_id = lambda: svc.request_id or util.new_request_id()
index 2380e48b734005505f125a05f453e6b88c76265c..be8a03fc314d2cf599c16d5f44b1ab61cc9e885d 100644 (file)
@@ -491,3 +491,11 @@ def get_config_once(svc):
     if not hasattr(svc, '_cached_config'):
         svc._cached_config = svc.configs().get().execute()
     return svc._cached_config
+
+def get_vocabulary_once(svc):
+    if not svc._rootDesc.get('resources').get('vocabularies', False):
+        # Old API server version, no vocabulary export endpoint
+        return {}
+    if not hasattr(svc, '_cached_vocabulary'):
+        svc._cached_vocabulary = svc.vocabularies().get().execute()
+    return svc._cached_vocabulary
diff --git a/sdk/python/arvados/vocabulary.py b/sdk/python/arvados/vocabulary.py
new file mode 100644 (file)
index 0000000..3bb87c4
--- /dev/null
@@ -0,0 +1,127 @@
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
+import logging
+
+from . import api
+
+_logger = logging.getLogger('arvados.vocabulary')
+
+def load_vocabulary(api_client=None):
+    """Load the Arvados vocabulary from the API.
+    """
+    if api_client is None:
+        api_client = api('v1')
+    return Vocabulary(api_client.vocabulary())
+
+class VocabularyError(Exception):
+    """Base class for all vocabulary errors.
+    """
+    pass
+
+class VocabularyKeyError(VocabularyError):
+    pass
+
+class VocabularyValueError(VocabularyError):
+    pass
+
+class Vocabulary(object):
+    def __init__(self, voc_definition={}):
+        self.strict_keys = voc_definition.get('strict_tags', False)
+        self.key_aliases = {}
+
+        for key_id, val in (voc_definition.get('tags') or {}).items():
+            strict = val.get('strict', False)
+            key_labels = [l['label'] for l in val.get('labels', [])]
+            values = {}
+            for v_id, v_val in (val.get('values') or {}).items():
+                labels = [l['label'] for l in v_val.get('labels', [])]
+                values[v_id] = VocabularyValue(v_id, labels)
+            vk = VocabularyKey(key_id, key_labels, values, strict)
+            self.key_aliases[key_id.lower()] = vk
+            for alias in vk.aliases:
+                self.key_aliases[alias.lower()] = vk
+
+    def __getitem__(self, key):
+        return self.key_aliases[key.lower()]
+
+    def convert_to_identifiers(self, obj={}):
+        """Translate key/value pairs to machine readable identifiers.
+        """
+        return self._convert_to_what(obj, 'identifier')
+
+    def convert_to_labels(self, obj={}):
+        """Translate key/value pairs to human readable labels.
+        """
+        return self._convert_to_what(obj, 'preferred_label')
+
+    def _convert_to_what(self, obj={}, what=None):
+        if not isinstance(obj, dict):
+            raise ValueError("obj must be a dict")
+        if what not in ['preferred_label', 'identifier']:
+            raise ValueError("what attr must be 'preferred_label' or 'identifier'")
+        r = {}
+        for k, v in obj.items():
+            # Key validation & lookup
+            key_found = False
+            if not isinstance(k, str):
+                raise VocabularyKeyError("key '{}' must be a string".format(k))
+            k_what, v_what = k, v
+            try:
+                k_what = getattr(self[k], what)
+                key_found = True
+            except KeyError:
+                if self.strict_keys:
+                    raise VocabularyKeyError("key '{}' not found in vocabulary".format(k))
+
+            # Value validation & lookup
+            if isinstance(v, list):
+                v_what = []
+                for x in v:
+                    if not isinstance(x, str):
+                        raise VocabularyValueError("value '{}' for key '{}' must be a string".format(x, k))
+                    try:
+                        v_what.append(getattr(self[k][x], what))
+                    except KeyError:
+                        if self[k].strict:
+                            raise VocabularyValueError("value '{}' not found for key '{}'".format(x, k))
+                        v_what.append(x)
+            else:
+                if not isinstance(v, str):
+                    raise VocabularyValueError("{} value '{}' for key '{}' must be a string".format(type(v).__name__, v, k))
+                try:
+                    v_what = getattr(self[k][v], what)
+                except KeyError:
+                    if key_found and self[k].strict:
+                        raise VocabularyValueError("value '{}' not found for key '{}'".format(v, k))
+
+            r[k_what] = v_what
+        return r
+
+class VocabularyData(object):
+    def __init__(self, identifier, aliases=[]):
+        self.identifier = identifier
+        self.aliases = aliases
+
+    def __getattribute__(self, name):
+        if name == 'preferred_label':
+            return self.aliases[0]
+        return super(VocabularyData, self).__getattribute__(name)
+
+class VocabularyValue(VocabularyData):
+    def __init__(self, identifier, aliases=[]):
+        super(VocabularyValue, self).__init__(identifier, aliases)
+
+class VocabularyKey(VocabularyData):
+    def __init__(self, identifier, aliases=[], values={}, strict=False):
+        super(VocabularyKey, self).__init__(identifier, aliases)
+        self.strict = strict
+        self.value_aliases = {}
+        for v_id, v_val in values.items():
+            self.value_aliases[v_id.lower()] = v_val
+            for v_alias in v_val.aliases:
+                self.value_aliases[v_alias.lower()] = v_val
+
+    def __getitem__(self, key):
+        return self.value_aliases[key.lower()]
\ No newline at end of file
diff --git a/sdk/python/tests/test_vocabulary.py b/sdk/python/tests/test_vocabulary.py
new file mode 100644 (file)
index 0000000..aa2e739
--- /dev/null
@@ -0,0 +1,319 @@
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
+import arvados
+import unittest
+import mock
+
+from arvados import api, vocabulary
+
+class VocabularyTest(unittest.TestCase):
+    EXAMPLE_VOC = {
+        'tags': {
+            'IDTAGANIMALS': {
+                'strict': False,
+                'labels': [
+                    {'label': 'Animal'},
+                    {'label': 'Creature'},
+                ],
+                'values': {
+                    'IDVALANIMAL1': {
+                        'labels': [
+                            {'label': 'Human'},
+                            {'label': 'Homo sapiens'},
+                        ],
+                    },
+                    'IDVALANIMAL2': {
+                        'labels': [
+                            {'label': 'Elephant'},
+                            {'label': 'Loxodonta'},
+                        ],
+                    },
+                },
+            },
+            'IDTAGIMPORTANCES': {
+                'strict': True,
+                'labels': [
+                    {'label': 'Importance'},
+                    {'label': 'Priority'},
+                ],
+                'values': {
+                    'IDVALIMPORTANCE1': {
+                        'labels': [
+                            {'label': 'High'},
+                            {'label': 'High priority'},
+                        ],
+                    },
+                    'IDVALIMPORTANCE2': {
+                        'labels': [
+                            {'label': 'Medium'},
+                            {'label': 'Medium priority'},
+                        ],
+                    },
+                    'IDVALIMPORTANCE3': {
+                        'labels': [
+                            {'label': 'Low'},
+                            {'label': 'Low priority'},
+                        ],
+                    },
+                },
+            },
+            'IDTAGCOMMENTS': {
+                'strict': False,
+                'labels': [
+                    {'label': 'Comment'},
+                    {'label': 'Notes'},
+                ],
+                'values': None,
+            },
+        },
+    }
+
+    def setUp(self):
+        self.api = arvados.api('v1')
+        self.voc = vocabulary.Vocabulary(self.EXAMPLE_VOC)
+        self.api.vocabulary = mock.MagicMock(return_value=self.EXAMPLE_VOC)
+
+    def test_vocabulary_keys(self):
+        self.assertEqual(self.voc.strict_keys, False)
+        self.assertEqual(
+            self.voc.key_aliases.keys(),
+            set(['idtaganimals', 'creature', 'animal',
+                'idtagimportances', 'importance', 'priority',
+                'idtagcomments', 'comment', 'notes'])
+        )
+
+        vk = self.voc.key_aliases['creature']
+        self.assertEqual(vk.strict, False)
+        self.assertEqual(vk.identifier, 'IDTAGANIMALS')
+        self.assertEqual(vk.aliases, ['Animal', 'Creature'])
+        self.assertEqual(vk.preferred_label, 'Animal')
+        self.assertEqual(
+            vk.value_aliases.keys(),
+            set(['idvalanimal1', 'human', 'homo sapiens',
+                'idvalanimal2', 'elephant', 'loxodonta'])
+        )
+
+    def test_vocabulary_values(self):
+        vk = self.voc.key_aliases['creature']
+        vv = vk.value_aliases['human']
+        self.assertEqual(vv.identifier, 'IDVALANIMAL1')
+        self.assertEqual(vv.aliases, ['Human', 'Homo sapiens'])
+        self.assertEqual(vv.preferred_label, 'Human')
+
+    def test_vocabulary_indexing(self):
+        self.assertEqual(self.voc['creature']['human'].identifier, 'IDVALANIMAL1')
+        self.assertEqual(self.voc['Creature']['Human'].identifier, 'IDVALANIMAL1')
+        self.assertEqual(self.voc['CREATURE']['HUMAN'].identifier, 'IDVALANIMAL1')
+        with self.assertRaises(KeyError):
+            inexistant = self.voc['foo']
+
+    def test_empty_vocabulary(self):
+        voc = vocabulary.Vocabulary({})
+        self.assertEqual(voc.strict_keys, False)
+        self.assertEqual(voc.key_aliases, {})
+
+    def test_load_vocabulary_with_api(self):
+        voc = vocabulary.load_vocabulary(self.api)
+        self.assertEqual(voc['creature']['human'].identifier, 'IDVALANIMAL1')
+        self.assertEqual(voc['Creature']['Human'].identifier, 'IDVALANIMAL1')
+        self.assertEqual(voc['CREATURE']['HUMAN'].identifier, 'IDVALANIMAL1')
+
+    def test_convert_to_identifiers(self):
+        cases = [
+            {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1'},
+            {'IDTAGIMPORTANCES': 'High'},
+            {'importance': 'IDVALIMPORTANCE1'},
+            {'priority': 'high priority'},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_identifiers(case),
+                {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1'},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_identifiers_multiple_pairs(self):
+        cases = [
+            {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1', 'IDTAGANIMALS': 'IDVALANIMAL1', 'IDTAGCOMMENTS': 'Very important person'},
+            {'IDTAGIMPORTANCES': 'High', 'IDTAGANIMALS': 'IDVALANIMAL1', 'comment': 'Very important person'},
+            {'importance': 'IDVALIMPORTANCE1', 'animal': 'IDVALANIMAL1', 'notes': 'Very important person'},
+            {'priority': 'high priority', 'animal': 'IDVALANIMAL1', 'NOTES': 'Very important person'},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_identifiers(case),
+                {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1', 'IDTAGANIMALS': 'IDVALANIMAL1', 'IDTAGCOMMENTS': 'Very important person'},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_identifiers_value_lists(self):
+        cases = [
+            {'IDTAGIMPORTANCES': ['IDVALIMPORTANCE1', 'IDVALIMPORTANCE2']},
+            {'IDTAGIMPORTANCES': ['High', 'Medium']},
+            {'importance': ['IDVALIMPORTANCE1', 'IDVALIMPORTANCE2']},
+            {'priority': ['high', 'medium']},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_identifiers(case),
+                {'IDTAGIMPORTANCES': ['IDVALIMPORTANCE1', 'IDVALIMPORTANCE2']},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_identifiers_unknown_key(self):
+        # Non-strict vocabulary
+        self.assertEqual(self.voc.strict_keys, False)
+        self.assertEqual(self.voc.convert_to_identifiers({'foo': 'bar'}), {'foo': 'bar'})
+        # Strict vocabulary
+        strict_voc = arvados.vocabulary.Vocabulary(self.EXAMPLE_VOC)
+        strict_voc.strict_keys = True
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            strict_voc.convert_to_identifiers({'foo': 'bar'})
+
+    def test_convert_to_identifiers_invalid_key(self):
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_identifiers({42: 'bar'})
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_identifiers({None: 'bar'})
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_identifiers({('f', 'o', 'o'): 'bar'})
+
+    def test_convert_to_identifiers_unknown_value(self):
+        # Non-strict key
+        self.assertEqual(self.voc['animal'].strict, False)
+        self.assertEqual(self.voc.convert_to_identifiers({'Animal': 'foo'}), {'IDTAGANIMALS': 'foo'})
+        # Strict key
+        self.assertEqual(self.voc['priority'].strict, True)
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Priority': 'foo'})
+
+    def test_convert_to_identifiers_invalid_value(self):
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': 42})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': None})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': {'hello': 'world'}})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': [42]})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': [None]})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Animal': [{'hello': 'world'}]})
+
+    def test_convert_to_identifiers_unknown_value_list(self):
+        # Non-strict key
+        self.assertEqual(self.voc['animal'].strict, False)
+        self.assertEqual(
+            self.voc.convert_to_identifiers({'Animal': ['foo', 'loxodonta']}),
+            {'IDTAGANIMALS': ['foo', 'IDVALANIMAL2']}
+        )
+        # Strict key
+        self.assertEqual(self.voc['priority'].strict, True)
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_identifiers({'Priority': ['foo', 'bar']})
+
+    def test_convert_to_labels(self):
+        cases = [
+            {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1'},
+            {'IDTAGIMPORTANCES': 'High'},
+            {'importance': 'IDVALIMPORTANCE1'},
+            {'priority': 'high priority'},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_labels(case),
+                {'Importance': 'High'},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_labels_multiple_pairs(self):
+        cases = [
+            {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1', 'IDTAGANIMALS': 'IDVALANIMAL1', 'IDTAGCOMMENTS': 'Very important person'},
+            {'IDTAGIMPORTANCES': 'High', 'IDTAGANIMALS': 'IDVALANIMAL1', 'comment': 'Very important person'},
+            {'importance': 'IDVALIMPORTANCE1', 'animal': 'IDVALANIMAL1', 'notes': 'Very important person'},
+            {'priority': 'high priority', 'animal': 'IDVALANIMAL1', 'NOTES': 'Very important person'},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_labels(case),
+                {'Importance': 'High', 'Animal': 'Human', 'Comment': 'Very important person'},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_labels_value_lists(self):
+        cases = [
+            {'IDTAGIMPORTANCES': ['IDVALIMPORTANCE1', 'IDVALIMPORTANCE2']},
+            {'IDTAGIMPORTANCES': ['High', 'Medium']},
+            {'importance': ['IDVALIMPORTANCE1', 'IDVALIMPORTANCE2']},
+            {'priority': ['high', 'medium']},
+        ]
+        for case in cases:
+            self.assertEqual(
+                self.voc.convert_to_labels(case),
+                {'Importance': ['High', 'Medium']},
+                "failing test case: {}".format(case)
+            )
+
+    def test_convert_to_labels_unknown_key(self):
+        # Non-strict vocabulary
+        self.assertEqual(self.voc.strict_keys, False)
+        self.assertEqual(self.voc.convert_to_labels({'foo': 'bar'}), {'foo': 'bar'})
+        # Strict vocabulary
+        strict_voc = arvados.vocabulary.Vocabulary(self.EXAMPLE_VOC)
+        strict_voc.strict_keys = True
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            strict_voc.convert_to_labels({'foo': 'bar'})
+
+    def test_convert_to_labels_invalid_key(self):
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_labels({42: 'bar'})
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_labels({None: 'bar'})
+        with self.assertRaises(vocabulary.VocabularyKeyError):
+            self.voc.convert_to_labels({('f', 'o', 'o'): 'bar'})
+
+    def test_convert_to_labels_unknown_value(self):
+        # Non-strict key
+        self.assertEqual(self.voc['animal'].strict, False)
+        self.assertEqual(self.voc.convert_to_labels({'IDTAGANIMALS': 'foo'}), {'Animal': 'foo'})
+        # Strict key
+        self.assertEqual(self.voc['priority'].strict, True)
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': 'foo'})
+
+    def test_convert_to_labels_invalid_value(self):
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': {'high': True}})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': None})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': 42})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': False})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': [42]})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': [None]})
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': [{'high': True}]})
+
+    def test_convert_to_labels_unknown_value_list(self):
+        # Non-strict key
+        self.assertEqual(self.voc['animal'].strict, False)
+        self.assertEqual(
+            self.voc.convert_to_labels({'IDTAGANIMALS': ['foo', 'IDVALANIMAL1']}),
+            {'Animal': ['foo', 'Human']}
+        )
+        # Strict key
+        self.assertEqual(self.voc['priority'].strict, True)
+        with self.assertRaises(vocabulary.VocabularyValueError):
+            self.voc.convert_to_labels({'IDTAGIMPORTANCES': ['foo', 'bar']})
+
+    def test_convert_roundtrip(self):
+        initial = {'IDTAGIMPORTANCES': 'IDVALIMPORTANCE1', 'IDTAGANIMALS': 'IDVALANIMAL1', 'IDTAGCOMMENTS': 'Very important person'}
+        converted = self.voc.convert_to_labels(initial)
+        self.assertNotEqual(converted, initial)
+        self.assertEqual(self.voc.convert_to_identifiers(converted), initial)
index 59ac639baf929dd2c06c9352159f33288be1d792..100b9168179e04b4baa5d6b51a438f00e2e42104 100644 (file)
@@ -37,7 +37,7 @@ class Arvados::V1::SchemaController < ApplicationController
         # format is YYYYMMDD, must be fixed width (needs to be lexically
         # sortable), updated manually, may be used by clients to
         # determine availability of API server features.
-        revision: "20210628",
+        revision: "20220222",
         source_version: AppVersion.hash,
         sourceVersion: AppVersion.hash, # source_version should be deprecated in the future
         packageVersion: AppVersion.package_version,
@@ -427,6 +427,27 @@ class Arvados::V1::SchemaController < ApplicationController
         }
       }
 
+      discovery[:resources]['vocabularies'] = {
+        methods: {
+          get: {
+            id: "arvados.vocabularies.get",
+            path: "vocabulary",
+            httpMethod: "GET",
+            description: "Get vocabulary definition",
+            parameters: {
+            },
+            parameterOrder: [
+            ],
+            response: {
+            },
+            scopes: [
+              "https://api.arvados.org/auth/arvados",
+              "https://api.arvados.org/auth/arvados.readonly"
+            ]
+          },
+        }
+      }
+
       discovery[:resources]['sys'] = {
         methods: {
           get: {