Use a per-documentLoader resolved context cache.

Fixes an issue where multiple document loaders are used which each have different values for static contexts. A WeakMap is used for caches and is cleaned up using WeakMap semantics based on the lifetime of the documentLoader keys.
digitalbazaar · May 31, 2024 · 0664018 · 0664018
1 parent 5367858
commit 0664018
Show file tree

Hide file tree

Showing 3 changed files with 112 additions and 21 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,12 @@
 # jsonld ChangeLog
 
+## 8.4.0 - 2024-xx-xx
+
+### Fixed
+- Use a per-`documentLoader` resolved context cache. Fixes an issue where
+  multiple document loaders are used which each have different values for
+  static contexts.
+
 ## 8.3.2 - 2023-12-06
 
 ### Fixed

diff --git a/lib/jsonld.js b/lib/jsonld.js
@@ -97,10 +97,28 @@ const wrapper = function(jsonld) {
 /** Registered RDF dataset parsers hashed by content-type. */
 const _rdfParsers = {};
 
-// resolved context cache
-// TODO: consider basing max on context size rather than number
+// resolved context caches
+// TODO: add controls for cache resource usage
+// cache size per document loader
 const RESOLVED_CONTEXT_CACHE_MAX_SIZE = 100;
-const _resolvedContextCache = new LRU({max: RESOLVED_CONTEXT_CACHE_MAX_SIZE});
+// caches are created and indexed per documentLoader
+// resources are cleaned up with WeakMap semantics for the documentLoaders
+const _resolvedContextCaches = new WeakMap();
+// default key to use when no documentLoader used
+const _defaultDocumentLoaderKey = {};
+
+// make a ContextResolver using a per-documentLoader shared cache
+function _makeContextResolver({documentLoader = _defaultDocumentLoaderKey}) {
+  let cache = _resolvedContextCaches.get(documentLoader);
+  if(!cache) {
+    // TODO: consider basing max on context size rather than number
+    cache = new LRU({max: RESOLVED_CONTEXT_CACHE_MAX_SIZE});
+    _resolvedContextCaches.set(documentLoader, cache);
+  }
+  return new ContextResolver({
+    sharedCache: cache
+  });
+}
 
 /* Core API */
 
@@ -152,8 +170,9 @@ jsonld.compact = async function(input, ctx, options) {
     skipExpansion: false,
     link: false,
     issuer: new IdentifierIssuer('_:b'),
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
   if(options.link) {
     // force skip expansion when linking, "link" is not part of the public
@@ -269,8 +288,9 @@ jsonld.expand = async function(input, options) {
   // set default options
   options = _setDefaults(options, {
     keepFreeFloatingNodes: false,
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // build set of objects that may have @contexts to resolve
@@ -368,8 +388,9 @@ jsonld.flatten = async function(input, ctx, options) {
   // set default options
   options = _setDefaults(options, {
     base: _isString(input) ? input : '',
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // expand input
@@ -423,8 +444,9 @@ jsonld.frame = async function(input, frame, options) {
     requireAll: false,
     omitDefault: false,
     bnodesToClear: [],
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // if frame is a string, attempt to dereference remote document
@@ -565,8 +587,9 @@ jsonld.normalize = jsonld.canonize = async function(input, options) {
     algorithm: 'URDNA2015',
     skipExpansion: false,
     safe: true,
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
   if('inputFormat' in options) {
     if(options.inputFormat !== 'application/n-quads' &&
@@ -674,8 +697,9 @@ jsonld.toRDF = async function(input, options) {
   options = _setDefaults(options, {
     base: _isString(input) ? input : '',
     skipExpansion: false,
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // TODO: support toRDF custom map?
@@ -726,8 +750,9 @@ jsonld.createNodeMap = async function(input, options) {
   // set default options
   options = _setDefaults(options, {
     base: _isString(input) ? input : '',
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // expand input
@@ -774,8 +799,9 @@ jsonld.merge = async function(docs, ctx, options) {
 
   // set default options
   options = _setDefaults(options, {
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // expand all documents
@@ -926,8 +952,9 @@ jsonld.processContext = async function(
   // set default options
   options = _setDefaults(options, {
     base: '',
-    contextResolver: new ContextResolver(
-      {sharedCache: _resolvedContextCache})
+    contextResolver: _makeContextResolver({
+      documentLoader: options ? options.documentLoader : undefined
+    })
   });
 
   // return initial context early for null context

diff --git a/tests/misc.js b/tests/misc.js
@@ -299,6 +299,63 @@ describe('loading multiple levels of contexts', () => {
   });
 });
 
+// check that internal caching is unique for each document loader
+describe('unique document loaders caching', () => {
+  const documentLoader0 = url => {
+    if(url === 'https://example.com/context') {
+      return {
+        document: {
+          "@context": {
+            "ex": "https://example.com/0#"
+          }
+        },
+        // must be marked static to get into the shared cache
+        tag: 'static',
+        contextUrl: null,
+        documentUrl: url
+      };
+    }
+  };
+  const documentLoader1 = url => {
+    if(url === 'https://example.com/context') {
+      return {
+        document: {
+          "@context": {
+            "ex": "https://example.com/1#"
+          }
+        },
+        contextUrl: null,
+        documentUrl: url
+      };
+    }
+  };
+  const doc = {
+    "@context": "https://example.com/context",
+    "ex:test": "test"
+  };
+  const expected0 = [{
+    "https://example.com/0#test": [{
+      "@value": "test"
+    }]
+  }];
+  const expected1 = [{
+    "https://example.com/1#test": [{
+      "@value": "test"
+    }]
+  }];
+
+  it('unique document loader caches', async () => {
+    const expanded0 = await jsonld.expand(doc, {
+      documentLoader: documentLoader0
+    });
+    assert.deepEqual(expanded0, expected0);
+    const expanded1 = await jsonld.expand(doc, {
+      documentLoader: documentLoader1
+    });
+    assert.deepEqual(expanded1, expected1);
+  });
+});
+
 describe('url tests', () => {
   it('should detect absolute IRIs', done => {
     // absolute IRIs