ricgraph 3.4__tar.gz → 3.5__tar.gz

{ricgraph-3.4/ricgraph.egg-info → ricgraph-3.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ricgraph
-Version: 3.4
+Version: 3.5
 Summary: Ricgraph - Research in context graph
 Author: Rik D.T. Janssen
 License-Expression: MIT
@@ -67,19 +67,19 @@ adapted to new application areas.
 
 Ricgraph is unique because it allows you to:
 
-* Explore collaborations between (sub-)organizations with Ricgraph;
-* Explore Open science with Ricgraph;
-* Enhance research information systems with Ricgraph;
+* Explore collaborations between (sub-)organizations;
+* Explore open science;
+* Enhance research information systems.
 
-The nice thing
-about Ricgraph is that anyone with Python knowledge can easily extend it,
-so you can use Ricgraph to test your own ideas or to create your own applications.
+One of the other strengths of Ricgraph is that anyone with
+Python knowledge can easily extend it, allowing you to test
+your own ideas or create your own applications.
 
 You can use Ricgraph either as a standalone tool,
 or as a component that provides a graph based data structure with
 a number of handy tools.
 Ricgraph has been developed for the application area
-research information
+research information,
 and it can be used for other application areas that require
 relations between items from various source systems.
 

{ricgraph-3.4 → ricgraph-3.5}/README-PyPI.md

@@ -39,19 +39,19 @@ adapted to new application areas.
 
 Ricgraph is unique because it allows you to:
 
-* Explore collaborations between (sub-)organizations with Ricgraph;
-* Explore Open science with Ricgraph;
-* Enhance research information systems with Ricgraph;
+* Explore collaborations between (sub-)organizations;
+* Explore open science;
+* Enhance research information systems.
 
-The nice thing
-about Ricgraph is that anyone with Python knowledge can easily extend it,
-so you can use Ricgraph to test your own ideas or to create your own applications.
+One of the other strengths of Ricgraph is that anyone with
+Python knowledge can easily extend it, allowing you to test
+your own ideas or create your own applications.
 
 You can use Ricgraph either as a standalone tool,
 or as a component that provides a graph based data structure with
 a number of handy tools.
 Ricgraph has been developed for the application area
-research information
+research information,
 and it can be used for other application areas that require
 relations between items from various source systems.
 

{ricgraph-3.4 → ricgraph-3.5}/README.md

@@ -59,14 +59,13 @@ adapted to new application areas.
 
 Ricgraph is unique because it allows you to:
 
-* [Explore collaborations between (sub-)organizations with
-  Ricgraph](https://www.ricgraph.eu/collaborations-with-ricgraph.html).
-* [Explore Open science with Ricgraph](https://www.ricgraph.eu/open-science-with-ricgraph.html).
-* [Enhance research information systems with Ricgraph](https://www.ricgraph.eu/enhancing-with-ricgraph.html).
+* [Explore collaborations between (sub-)organizations](https://www.ricgraph.eu/collaborations-with-ricgraph.html).
+* [Explore open science](https://www.ricgraph.eu/open-science-with-ricgraph.html).
+* [Enhance research information systems](https://www.ricgraph.eu/enhancing-with-ricgraph.html).
 
-The nice thing
-about Ricgraph is that anyone with Python knowledge can easily extend it,
-so you can use Ricgraph to test your own ideas or to create your own applications.
+One of the other strengths of Ricgraph is that anyone with
+Python knowledge can easily extend it, allowing you to test
+your own ideas or create your own applications.
 
 You can use Ricgraph either as a standalone tool,
 or as a component that provides a graph based data structure with
@@ -177,4 +176,4 @@ please let me know by creating a GitHub *Issue* at the top left of that page.
 <!-- Do not change the lines below. The version number will be
      modified by script maintenance/update_version.bash -->
 This README.md file is part of the 
-documentation for version 3.4 of Ricgraph - Research in context graph.
+documentation for version 3.5 of Ricgraph - Research in context graph.

{ricgraph-3.4 → ricgraph-3.5}/pyproject.toml

@@ -9,7 +9,7 @@ packages = ["ricgraph"]
 
 [project]
 name = "ricgraph"
-version = "3.4"
+version = "3.5"
 authors = [
   { name="Rik D.T. Janssen" },
 ]

{ricgraph-3.4 → ricgraph-3.5}/ricgraph/ricgraph.py

@@ -120,7 +120,7 @@ __package__ = 'Ricgraph'
 # __version__ should be placed here instead of in __init__.py.
 # If not here, it cannot be found when executing something like
 # PYTHONPATH=../ricgraph ../bin/python ricgraph_explorer.py
-__version__ = '3.4'
+__version__ = '3.5'
 
 
 # ##############################################################################
@@ -147,9 +147,9 @@ def update_nodes_df(nodes: DataFrame) -> None:
     """
     nodes_clean = nodes.copy(deep=True)
     # Ensure that all '' values are NaN, so that those rows can be easily removed with dropna()
-    nodes_clean.replace('', nan, inplace=True)
-    nodes_clean.dropna(axis=0, how='any', inplace=True)
-    nodes_clean.drop_duplicates(keep='first', inplace=True, ignore_index=True)
+    nodes_clean = nodes_clean.replace('', nan)
+    nodes_clean = nodes_clean.dropna(axis=0, how='any')
+    nodes_clean = nodes_clean.drop_duplicates(keep='first', ignore_index=True)
 
     print('There are ' + str(len(nodes_clean)) + ' nodes ('
           + timestamp() + '), updating node:')
@@ -248,7 +248,7 @@ def create_nodepairs_and_edges_params(name1: dict | str, category1: dict | str,
                                                     index=[0])
     else:
         left_and_right_nodepairs = DataFrame(data=value1).copy(deep=True)
-        left_and_right_nodepairs.rename(columns={value1.name: 'value1'}, inplace=True)
+        left_and_right_nodepairs = left_and_right_nodepairs.rename(columns={value1.name: 'value1'})
 
     # This makes programming more efficient
     allowed_properties = get_ricgraph_properties_additional() + get_ricgraph_properties_standard()
@@ -268,14 +268,14 @@ def create_nodepairs_and_edges_params(name1: dict | str, category1: dict | str,
                     left_and_right_nodepairs[prop_name + '1'] = other_value
                 else:
                     left_and_right_nodepairs = concat([left_and_right_nodepairs, other_value], axis='columns')
-                    left_and_right_nodepairs.rename(columns={other_value.name: prop_name + '1'}, inplace=True)
+                    left_and_right_nodepairs = left_and_right_nodepairs.rename(columns={other_value.name: prop_name + '1'})
 
             if prop_name + '2' == other_name:
                 if isinstance(other_value, str):
                     left_and_right_nodepairs[prop_name + '2'] = other_value
                 else:
                     left_and_right_nodepairs = concat([left_and_right_nodepairs, other_value], axis='columns')
-                    left_and_right_nodepairs.rename(columns={other_value.name: prop_name + '2'}, inplace=True)
+                    left_and_right_nodepairs = left_and_right_nodepairs.rename(columns={other_value.name: prop_name + '2'})
 
     # Reorder a bit, starting with the last, only the name-category-value property, since they will surely exist
     left_and_right_nodepairs.insert(0, 'value2', left_and_right_nodepairs.pop('value2'))
@@ -314,10 +314,10 @@ def unify_personal_identifiers(personal_identifiers: DataFrame,
             identifiers = personal_identifiers[[column1, column2]].copy(deep=True)
 
             # Ensure that all '' values are NaN, so that those rows can be easily removed with dropna()
-            identifiers.replace('', nan, inplace=True)
+            identifiers = identifiers.replace('', nan)
             # dropna(how='any'): drop row if one or more row values contain NaN
-            identifiers.dropna(axis=0, how='any', inplace=True)
-            identifiers.drop_duplicates(keep='first', inplace=True, ignore_index=True)
+            identifiers = identifiers.dropna(axis=0, how='any')
+            identifiers = identifiers.drop_duplicates(keep='first', ignore_index=True)
 
             if get_ricgraph_nodeadd_mode() == 'strict':
                 # Remove rows which have a cell value that occurs more than once.
@@ -342,11 +342,11 @@ def unify_personal_identifiers(personal_identifiers: DataFrame,
                 # For more explanation, see file docs/ricgraph_install_configure.md,
                 # section RICGRAPH_NODEADD_MODE.
                 if column1 != 'FULL_NAME' and column2 != 'FULL_NAME':
-                    identifiers.drop_duplicates(subset=column1, keep=False, inplace=True, ignore_index=True)
-                    identifiers.drop_duplicates(subset=column2, keep=False, inplace=True, ignore_index=True)
+                    identifiers = identifiers.drop_duplicates(subset=column1, keep=False, ignore_index=True)
+                    identifiers = identifiers.drop_duplicates(subset=column2, keep=False, ignore_index=True)
 
-            identifiers.rename(columns={column1: 'value1',
-                                        column2: 'value2'}, inplace=True)
+            identifiers = identifiers.rename(columns={column1: 'value1',
+                                                      column2: 'value2'})
             new_identifiers_columns = {'name1': column1,
                                        'category1': PERSON_CATEGORY_PERSON,
                                        'name2': column2,

{ricgraph-3.4 → ricgraph-3.5}/ricgraph/ricgraph_cache.py

@@ -37,31 +37,34 @@
 # Original version Rik D.T. Janssen, December 2022.
 # Updated Rik D.T. Janssen, February, March, September to December 2023.
 # Updated Rik D.T. Janssen, February to June, September to December  2024.
-# Updated Rik D.T. Janssen, January to June, October 2025.
+# Updated Rik D.T. Janssen, January to June, October 2025, July 2026.
 #
 # ########################################################################
 
 
 from sys import getsizeof
-from typing import Optional, Tuple
-from pymemcache.client.base import Client
-from .ricgraph_constants import MAX_NODES_CACHE_KEY_ID
+from typing import Optional, Tuple, Any
+from pymemcache.client.base import PooledClient
+from .ricgraph_constants import MAX_NODES_CACHE_KEY_ID, MEMCACHED_MAX_POOL_SIZE
 from .ricgraph_utils import (get_configfile_key_memcached_parameters,
                              serialize_value, deserialize_value)
 
 
-# This dict is used as a cache for node id's. If we have a node id, we can
+# This dict is used as a cache. 
+# It may be used for node ids. If we have a node id, we can
 # do a direct lookup for a node in O(1) in the graph database,
 # instead of a search in O(log n).
-# The dict has the format: [Ricgraph _key]: [Node element_id].
-_nodes_cache_key_id = {}
+# For nodes, the dict has the format: [Ricgraph _key]: [Node element_id].
+# For all other types, it uses JSON, and it has the 
+# format: [key]: [value].
+_ricgraph_cache = {}
 
 # Global indicating whether Memcached is available.
 _memcached_available = False
 
 # Global for connection to Memcached.
 # Type hint necessary to avoid PyCharm warning.
-_memcached_client: Optional[Client] = None
+_memcached_client: Optional[PooledClient] = None
 
 
 def memcached_open_connection() -> None:
@@ -84,8 +87,10 @@ def memcached_open_connection() -> None:
         return
 
     try:
-        memcached_client = Client(server=(memcached_host, memcached_port),
-                                  allow_unicode_keys=True)
+        # Do not use Client(), it is not thread safe.
+        memcached_client = PooledClient(server=(memcached_host, memcached_port),
+                                        max_pool_size=MEMCACHED_MAX_POOL_SIZE,
+                                        allow_unicode_keys=True)
 
         # Test connection by setting and deleting a test key.
         memcached_client.set(key='test_key', value=b'test', expire=1)
@@ -118,125 +123,119 @@ def memcached_check_available() -> bool:
         return True
 
 
-def nodes_cache_key_id_create(key: str, elementid: str) -> None:
-    """Create an entry in the cache for node id's.
-    'id' in this sentence is the id assigned by de graph database.
+def ricgraph_cache_item_create(key: str, value: Any) -> None:
+    """Create an entry in the cache.
 
-    :param key: _key of a node.
-    :param elementid: id of a node.
+    :param key: key of a cache item.
+    :param value: value of a cache item.
     :return: None.
     """
-    global _nodes_cache_key_id, _memcached_available, _memcached_client
+    global _ricgraph_cache, _memcached_available, _memcached_client
 
-    if key == '' or elementid == '':
+    if key == '' or value == '':
         return
 
+    key_clean = key.replace(' ', '_')
+    value_serialized = serialize_value(value=value)
     if memcached_check_available():
         if _memcached_client is None:
             return
-        # The docs are unclear whether I need to serialize the key if I
-        # use flag allow_unicode_keys=True in Client() above.
-        # So I do serialize to be sure. Since keys cannot have spaces,
-        # these are replaced first.
-        key_serialized = serialize_value(value=key.replace(' ', '_'))
-        elementid_serialized = serialize_value(value=elementid)
         try:
-            _memcached_client.set(key=key_serialized,
-                                  value=elementid_serialized,
+            _memcached_client.set(key=key_clean,
+                                  value=value_serialized,
                                   expire=0)
         except:
-            print('nodes_cache_key_id_create(): Warning, connection to Memcached lost, continuing...')
+            print('ricgraph_cache_item_create(): Warning, connection to Memcached lost, continuing...')
             # Continue, hopefully the connection will come back soon.
         return
 
-    if len(_nodes_cache_key_id) > MAX_NODES_CACHE_KEY_ID:
-        nodes_cache_key_id_empty()
+    if len(_ricgraph_cache) > MAX_NODES_CACHE_KEY_ID:
+        ricgraph_cache_empty()
 
     # We use a 'dict', which does not allow for duplicates, so we
     # do not need to check for duplicates.
     # https://docs.python.org/3/library/stdtypes.html#mapping-types-dict
     # tells that Python dict keys and values can have _almost_ any type.
-    # So we do not need to serialize and deserialize.
-    _nodes_cache_key_id[key] = elementid
-    # print('Create: ' + str(key) + ' -- ' + _nodes_cache_key_id[key])
+    # We serialize for symmetry with Memcached.
+    # print('Create: ' + str(key_clean) + ' -- ' + str(value_serialized))
+    _ricgraph_cache[key_clean] = value_serialized
     return
 
 
-def nodes_cache_key_id_read(key: str) -> str:
-    """Read an entry from the cache for node id's.
-    'id' in this sentence is the id assigned by de graph database.
+def ricgraph_cache_item_read(key: str) -> Any:
+    """Read an entry from the cache.
 
-    :param key: _key of a node.
-    :return: The id of the node, or '' if not present.
+    :param key: key of a cache item.
+    :return: The value of the cache item, or '' if not present.
     """
-    global _nodes_cache_key_id, _memcached_available, _memcached_client
+    global _ricgraph_cache, _memcached_available, _memcached_client
 
     if key == '':
         return ''
 
+    key_clean = key.replace(' ', '_')
     if memcached_check_available():
         if _memcached_client is None:
             return ''
-        key_serialized = serialize_value(value=key.replace(' ', '_'))
         try:
-            elementid_serialized = _memcached_client.get(key=key_serialized)
+            value_serialized = _memcached_client.get(key=key_clean)
         except:
-            print('nodes_cache_key_id_read(): Warning, connection to Memcached lost, continuing...')
+            print('ricgraph_cache_item_read(): Warning, connection to Memcached lost, continuing...')
             # Continue, hopefully the connection will come back soon.
             return ''
 
-        if elementid_serialized is None:
+        if value_serialized is None:
             # Key not found.
             return ''
         else:
-            elementid = deserialize_value(serialized=elementid_serialized)
-            return elementid
-
-    if key in _nodes_cache_key_id:
-        # print('Read: ' + str(key) + ' -- ' + _nodes_cache_key_id[key])
-        return _nodes_cache_key_id[key]
+            value = deserialize_value(serialized=value_serialized)
+            return value
+
+    if key_clean in _ricgraph_cache:
+        value_serialized = _ricgraph_cache[key_clean]
+        # print('Read: ' + str(key_clean) + ' -- ' + str(value_serialized))
+        value = deserialize_value(serialized=value_serialized)
+        return value
     return ''
 
 
-def nodes_cache_key_id_delete_key(key: str) -> None:
-    """Delete a key 'key' from the cache for node id's.
-    'id' in this sentence is the id assigned by de graph database.
+def ricgraph_cache_item_delete_key(key: str) -> None:
+    """Delete a key 'key' from the cache.
 
-    :param key: _key of a node.
+    :param key: key of a cache item.
     :return: None.
     """
-    global _nodes_cache_key_id, _memcached_available, _memcached_client
+    global _ricgraph_cache, _memcached_available, _memcached_client
 
     if key == '':
         return
 
+    key_clean = key.replace(' ', '_')
     if memcached_check_available():
         if _memcached_client is None:
             return
-        key_serialized = serialize_value(value=key.replace(' ', '_'))
         try:
             # Works regardless whether the key is present or not.
-            _memcached_client.delete(key=key_serialized)
+            _memcached_client.delete(key=key_clean)
         except:
-            print('nodes_cache_key_id_delete(): Error, connection to Memcached lost, exiting...')
+            print('nodes_cache_item_delete_key(): Error, connection to Memcached lost, exiting...')
             # Exit because the cache gets in an unexpected state:
             # an element remains in the cache while it should not.
             exit(1)
         return
 
-    if key in _nodes_cache_key_id:
-        # print('Delete: ' + str(key) + ' -- ' + _nodes_cache_key_id[key])
-        _nodes_cache_key_id.pop(key)
+    if key_clean in _ricgraph_cache:
+        # print('Delete: ' + str(key_clean) + ' -- ' + _ricgraph_cache[key_clean])
+        _ricgraph_cache.pop(key_clean)
     return
 
 
-def nodes_cache_key_id_empty() -> None:
-    """Empty the cache for node id's.
-    'id' in this sentence is the id assigned by de graph database.
+def ricgraph_cache_empty() -> None:
+    """Empty the cache.
 
     :return: None.
     """
-    global _nodes_cache_key_id, _memcached_available, _memcached_client
+    global _ricgraph_cache, _memcached_available, _memcached_client
 
     if memcached_check_available():
         if _memcached_client is None:
@@ -244,24 +243,23 @@ def nodes_cache_key_id_empty() -> None:
         try:
             _memcached_client.flush_all()
         except:
-            print('nodes_cache_key_id_empty(): Error, connection to Memcached lost, exiting...')
+            print('ricgraph_cache_empty(): Error, connection to Memcached lost, exiting...')
             # Exit because the cache gets in an unexpected state:
             # it is not emptied while it should have been.
             exit(1)
         return
 
     # print('Clear cache.')
-    _nodes_cache_key_id.clear()
+    _ricgraph_cache.clear()
     return
 
 
-def nodes_cache_key_id_size() -> Tuple[int, float]:
-    """Return the size of the cache for node id's.
-    'id' in this sentence is the id assigned by de graph database.
+def ricgraph_cache_size() -> Tuple[int, float]:
+    """Return the size of the cache.
 
     :return: a tuple with number of items and size in kB.
     """
-    global _nodes_cache_key_id, _memcached_client
+    global _ricgraph_cache, _memcached_client
 
     if memcached_check_available():
         if _memcached_client is None:
@@ -277,18 +275,18 @@ def nodes_cache_key_id_size() -> Tuple[int, float]:
             nr_items = 0
             size_kb = 0.0
     else:
-        nr_items = len(_nodes_cache_key_id)
-        size_kb = round(getsizeof(_nodes_cache_key_id) / 1000, 1)
+        nr_items = len(_ricgraph_cache)
+        size_kb = round(getsizeof(_ricgraph_cache) / 1000, 1)
     return nr_items, size_kb
 
 
-def nodes_cache_key_id_type_size() -> str:
+def ricgraph_cache_size_text() -> str:
     """Return the size of the cache for node id's.
     'id' in this sentence is the id assigned by de graph database.
 
     :return: a sentence with the length of the cache and its size.
     """
-    nr_items, size_kb = nodes_cache_key_id_size()
+    nr_items, size_kb = ricgraph_cache_size()
     if memcached_check_available():
         result = 'Ricgraph uses Memcached as cache. '
     else:

{ricgraph-3.4 → ricgraph-3.5}/ricgraph/ricgraph_constants.py

@@ -69,6 +69,8 @@ class QueryParams(TypedDict):
     start_orgs: str
     collab_orgs: str
     max_nr_items: int
+    skip_nr_nodes: int
+
 
 # This TypedDict is filled in ricgraph_utils.py/create_empty_page_params(),
 # and in ricgraph_explorer_utils.py/get_url_page_params().
@@ -84,6 +86,29 @@ class PageParams(TypedDict):
     view_mode: str
 
 
+# This TypedDict is a buffer/cache for the elements retrieved
+# from the graph database. Since retrieving elements may be
+# expensive, we get a batch_size of them and then return them
+# in pages of page_size. This is useful for the e.g. REST API.
+# graph_function_name is the function we call, and it has
+# graph_function_kwargs as parameters.
+# first_element and last_element are the 'real' indexes of
+# elements. E.g., if elements has a length of 100, and first_element
+# is 200, then the first element of elements is the 200-th element
+# returned from graph_function_name.
+# If exhausted is True, there are no more elements to retrieve
+# from graph_function_name, so it is not necessary to recall it.
+class GraphResultCache(TypedDict):
+    first_element: int      # Counts from 0, not from 1.
+    last_element: int       # So with 10 elems this will be 9.
+    elements: list[dict]    # In the dict, the elements are 0..9.
+    exhausted: bool
+    page_size: int
+    batch_size: int
+    graph_function_name: str
+    graph_function_kwargs: dict
+
+
 RICGRAPH_INI_FILENAME = 'ricgraph.ini'
 
 # This one separates value & name in property _key of a node.
@@ -114,12 +139,17 @@ HTTP_RESPONSE_OK = 200
 HTTP_RESPONSE_NOTHING_FOUND = 250
 HTTP_RESPONSE_INVALID_SEARCH = 251
 
-# The dict '_nodes_cache_key_id' is used to cache IDs to nodes. This is the cache size.
+# The dict '_ricgraph_cache' is used to cache IDs to nodes. This is the cache size.
 # The dict itself is defined in ricgraph_cache.py.
 # This dict will be emptied if it reaches this number of elements.
 # A cache entry is approx. 20 - 30 bytes. 30 x 2000000 ~ 60MB.
 MAX_NODES_CACHE_KEY_ID = 2000000
 
+# This is the maximum pool size for Memcached PooledClient().
+# max_pool_size is a per-worker setting, not a global one.
+# Various sources suggest that value 2 seems fine.
+MEMCACHED_MAX_POOL_SIZE = 2
+
 # In merge_two_nodes(node_merge_from, node_merge_to), we are constructing a list of
 # all the nodes that are to be merged with node_merge_to, to add to property _history
 # in that node. This value restricts the length of that list (to prevent property
@@ -444,29 +474,51 @@ NODELABELS_CATEGORY = [
 # ########################################################################
 # License related constants.
 # ########################################################################
-# For software the names and spellings from https://opensource.org/licenses
-# are used, column 'SPDX ID'.
+# For the full SPDX license list, see https://spdx.org/licenses.
+# Their names and spellings are used.
+# For software licenses: https://opensource.org/licenses.
+
+# This one should have '-only' or '-or-later' added, but
+# the source systems do not specify.
 LICENSE_AGPL_30 = 'AGPL-3.0'
+
 LICENSE_APACHE_20 = 'Apache-2.0'
 LICENSE_BSD_2 = 'BSD-2-Clause'
 LICENSE_BSD_3 = 'BSD-3-Clause'
-LICENSE_CC0 = 'CC0'
-LICENSE_CC_BY = 'CC BY'
-LICENSE_CC_BY_40 = 'CC BY-4.0'
-LICENSE_CC_BY_NC = 'CC BY-NC'
-LICENSE_CC_BY_NC_ND = 'CC BY-NC-ND'
-LICENSE_CC_BY_NC_SA = 'CC BY-NC-SA'
-LICENSE_CC_BY_ND = 'CC BY-ND'
-LICENSE_CC_BY_SA = 'CC BY-SA'
+LICENSE_CC0_10 = 'CC0-1.0'
+
+# A distinction is made between CC licenses that specify a
+# version number and CC licenses that do not.
+LICENSE_CC_BY = 'CC-BY'
+LICENSE_CC_BY_40 = 'CC-BY-4.0'
+LICENSE_CC_BY_NC = 'CC-BY-NC'
+LICENSE_CC_BY_NC_ND = 'CC-BY-NC-ND'
+LICENSE_CC_BY_NC_SA = 'CC-BY-NC-SA'
+LICENSE_CC_BY_ND = 'CC-BY-ND'
+LICENSE_CC_BY_SA = 'CC-BY-SA'
+
+# These should have '-only' or '-or-later' added, but
+# the source systems do not specify.
 LICENSE_GPL_20 = 'GPL-2.0'
 LICENSE_GPL_30 = 'GPL-3.0'
 LICENSE_LGPL_21 = 'LGPL-2.1'
 LICENSE_LGPL_30 = 'LGPL-3.0'
+
 LICENSE_MIT = 'MIT'
-LICENSE_MULTIPLE = 'Multiple'
+
+# This one is not an official SPDX license name, but it
+# appears in our source systems.
+LICENSE_MULTIPLE = 'multiple'
+
+LICENSE_ODBL_10 = 'ODbL-1.0'
 LICENSE_ODC_BY_10 = 'ODC-By-1.0'
-LICENSE_OPEN_ACCESS = 'Open access'
+
+# These are not official SPDX license names, but they
+# appear in our source systems.
+LICENSE_OPEN_ACCESS = 'open access'
 LICENSE_TAVERNE = 'Taverne (not for re-use)'
+
+# Fall back if no license is obtained from a source system.
 LICENSE_UNKNOWN = RICGRAPH_UNKNOWN
 
 LICENSE_ALL = [
@@ -474,7 +526,7 @@ LICENSE_ALL = [
     LICENSE_APACHE_20,
     LICENSE_BSD_2,
     LICENSE_BSD_3,
-    LICENSE_CC0,
+    LICENSE_CC0_10,
     LICENSE_CC_BY,
     LICENSE_CC_BY_40,
     LICENSE_CC_BY_NC,
@@ -488,6 +540,7 @@ LICENSE_ALL = [
     LICENSE_LGPL_30,
     LICENSE_MIT,
     LICENSE_MULTIPLE,
+    LICENSE_ODBL_10,
     LICENSE_ODC_BY_10,
     LICENSE_OPEN_ACCESS,
     LICENSE_TAVERNE,