Adjust lib to resolve function

Author: mielvds

lib/pyld/context_resolver.py

@@ -10,7 +10,7 @@
 
 from frozendict import frozendict
 from c14n.Canonicalize import canonicalize
-from pyld import jsonld
+from pyld import jsonld, iri_resolver
 from .resolved_context import ResolvedContext
 
 MAX_CONTEXT_URLS = 10
@@ -104,7 +104,7 @@ def _cache_resolved_context(self, key, resolved, tag):
 
     def _resolve_remote_context(self, active_ctx, url, base, cycles):
         # resolve relative URL and fetch context
-        url = jsonld.prepend_base(base, url)
+        url = iri_resolver.resolve(url, base)
         context, remote_doc = self._fetch_context(active_ctx, url, cycles)
 
         # update base according to remote document and resolve any relative URLs
@@ -194,13 +194,13 @@ def _resolve_context_urls(self, context, base):
         ctx = context.get('@context')
 
         if isinstance(ctx, str):
-            context['@context'] = jsonld.prepend_base(base, ctx)
+            context['@context'] = iri_resolver.resolve(ctx, base)
             return
 
         if isinstance(ctx, list):
             for num, element in enumerate(ctx):
                 if isinstance(element, str):
-                    ctx[num] = jsonld.prepend_base(base, element)
+                    ctx[num] = iri_resolver.resolve(element, base)
                 elif isinstance(element, dict) or isinstance(element, frozendict):
                     self. _resolve_context_urls({'@context': element}, base)
             return

lib/pyld/documentloader/aiohttp.py

@@ -13,6 +13,7 @@
 import threading
 import urllib.parse as urllib_parse
 
+from pyld import iri_resolver
 from pyld.jsonld import (JsonLdError, parse_link_header, LINK_HEADER_REL)
 
 
@@ -114,7 +115,7 @@ async def async_loader(url, headers):
                                 linked_alternate.get('type') == 'application/ld+json' and
                                 not re.match(r'^application\/(\w*\+)?json$', content_type)):
                             doc['contentType'] = 'application/ld+json'
-                            doc['documentUrl'] = jsonld.prepend_base(url, linked_alternate['target'])
+                            doc['documentUrl'] = iri_resolver.resolve(linked_alternate['target'], url)
 
                     return doc
         except JsonLdError as e:

lib/pyld/documentloader/requests.py

@@ -9,9 +9,11 @@
 .. moduleauthor:: Tim McNamara <tim.mcnamara@okfn.org>
 .. moduleauthor:: Olaf Conradi <olaf@conradi.org>
 """
+import re
 import string
 import urllib.parse as urllib_parse
 
+from pyld import iri_resolver
 from pyld.jsonld import (JsonLdError, parse_link_header, LINK_HEADER_REL)
 
 
@@ -92,7 +94,7 @@ def loader(url, options={}):
                         linked_alternate.get('type') == 'application/ld+json' and
                         not re.match(r'^application\/(\w*\+)?json$', content_type)):
                     doc['contentType'] = 'application/ld+json'
-                    doc['documentUrl'] = jsonld.prepend_base(url, linked_alternate['target'])
+                    doc['documentUrl'] = iri_resolver.resolve(linked_alternate['target'], url)
             return doc
         except JsonLdError as e:
             raise e

lib/pyld/jsonld.py

@@ -31,6 +31,7 @@
 from numbers import Integral, Real
 from frozendict import frozendict
 from pyld.__about__ import (__copyright__, __license__, __version__)
+from .iri_resolver import resolve 
 
 def cmp(a, b):
     return (a > b) - (a < b)
@@ -443,80 +444,6 @@ def unregister_rdf_parser(content_type):
         del _rdf_parsers[content_type]
 
 
-def prepend_base(base, iri):
-    """
-    Prepends a base IRI to the given relative IRI.
-
-    :param base: the base IRI.
-    :param iri: the relative IRI.
-
-    :return: the absolute IRI.
-    """
-    # skip IRI processing
-    if base is None:
-        return iri
-
-    # already an absolute iri
-    if _is_absolute_iri(iri):
-        return iri
-
-    # parse IRIs
-    base = parse_url(base)
-    rel = parse_url(iri)
-
-    # per RFC3986 5.2.2
-    transform = {
-        'scheme': base.scheme
-    }
-
-    if rel.authority is not None:
-        transform['authority'] = rel.authority
-        transform['path'] = rel.path
-        transform['query'] = rel.query
-    else:
-        transform['authority'] = base.authority
-
-        if rel.path == '':
-            transform['path'] = base.path
-            if rel.query is not None:
-                transform['query'] = rel.query
-            else:
-                transform['query'] = base.query
-        else:
-            if rel.path.startswith('/'):
-                # IRI represents an absolute path
-                transform['path'] = rel.path
-            else:
-                # merge paths
-                path = base.path
-
-                # append relative path to the end of the last directory from
-                # base
-                path = path[0:path.rfind('/') + 1]
-                if (len(path) > 0 or base.authority) and not path.endswith('/'):
-                    path += '/'
-                path += rel.path
-
-                transform['path'] = path
-
-            transform['query'] = rel.query
-
-    if rel.path != '':
-        # normalize path
-        transform['path'] = remove_dot_segments(transform['path'])
-
-    transform['fragment'] = rel.fragment
-
-    # construct URL
-    rval = unparse_url(transform)
-
-    # handle empty base case
-    if rval == '':
-        rval = './'
-
-    return rval
-
-
 def remove_base(base, iri):
     """
     Removes a base IRI from the given absolute IRI.
@@ -3188,10 +3115,10 @@ def _process_context(self, active_ctx, local_ctx, options,
                 base = ctx['@base']
                 if base is None:
                     base = None
-                elif _is_absolute_iri(base):
+                elif _is_absolute_iri(base) or (_is_relative_iri(base) and active_ctx.get('@base') is None):
                     base = base
-                elif _is_relative_iri(base):
-                    base = prepend_base(active_ctx.get('@base'), base)
+                elif _is_relative_iri(base) and active_ctx.get('@base') is not None:
+                    base = resolve(base, active_ctx.get('@base'))
                 else:
                     raise JsonLdError(
                         'Invalid JSON-LD syntax; the value of "@base" in a '
@@ -3295,7 +3222,7 @@ def _process_context(self, active_ctx, local_ctx, options,
                     process = True
 
                     if _is_string(key_ctx):
-                        url = prepend_base(options['base'], key_ctx)
+                        url = resolve(key_ctx, options['base'])
                         if url in cycles:
                             process = False
                         else:
@@ -4887,7 +4814,7 @@ def _compact_iri(
                 if active_ctx['@base'] is None:
                     return iri
                 else:
-                    return remove_base(prepend_base(base, active_ctx['@base']), iri)
+                    return remove_base(resolve(active_ctx['@base'], base), iri)
             else:
                 return remove_base(base, iri)
 
@@ -5510,9 +5437,10 @@ def _expand_iri(
         if base and '@base' in active_ctx:
             # The None case preserves rval as potentially relative
             if active_ctx['@base'] is not None:
-                rval = prepend_base(prepend_base(base, active_ctx['@base']), rval)
+                resolved_base = active_ctx['@base'] if _is_absolute_iri(active_ctx['@base']) else resolve(active_ctx['@base'], base)
+                rval = resolve(rval, resolved_base)
         elif base:
-            rval = prepend_base(base, rval)
+            rval = resolve(rval, base)
 
         return rval
 
@@ -6642,7 +6570,7 @@ def load_html(input, url, profile, options):
         # use either specified base, or document location
         effective_base = options.get('base', url)
         if effective_base:
-            html_base = prepend_base(effective_base, html_base[0])
+            html_base = resolve(html_base[0], effective_base)
         options['base'] = html_base
 
     url_elements = parse_url(url)

tests/runtests.py

@@ -20,7 +20,7 @@
 from unittest import TextTestResult
 
 sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', 'lib'))
-from pyld import jsonld
+from pyld import jsonld, iri_resolver
 
 __copyright__ = 'Copyright (c) 2011-2013 Digital Bazaar, Inc.'
 __license__ = 'New BSD license'
@@ -97,9 +97,9 @@ def main(self):
             # default to find known sibling test dirs
             test_targets = []
             sibling_dirs = [
-                '../json-ld-api/tests/',
-                '../json-ld-framing/tests/',
-                '../normalization/tests/',
+                './specifications/json-ld-api/tests/',
+                './specifications/json-ld-framing/tests/',
+                './specifications/normalization/tests/',
             ]
             for dir in sibling_dirs:
                 if os.path.exists(dir):
@@ -528,7 +528,7 @@ def load_locally(url):
                         linked_alternate.get('type') == 'application/ld+json' and
                         not re.match(r'^application\/(\w*\+)?json$', content_type)):
                     doc['contentType'] = 'application/ld+json'
-                    doc['documentUrl'] = jsonld.prepend_base(url, linked_alternate['target'])
+                    doc['documentUrl'] = iri_resolver.resolve(linked_alternate['target'], url)
         global ROOT_MANIFEST_DIR
         if doc['documentUrl'].find(':') == -1:
             filename = os.path.join(ROOT_MANIFEST_DIR, doc['documentUrl'])