idlib 0.0.1.dev21__tar.gz → 0.0.1.dev22__tar.gz

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: idlib
-Version: 0.0.1.dev21
+Version: 0.0.1.dev22
 Summary: A library for working with identifiers of all kinds.
 Home-page: https://github.com/tgbugs/idlib
 Author: Tom Gillespie
@@ -9,21 +9,22 @@ License: MIT
 Keywords: python persistent identifiers
 Classifier: Development Status :: 4 - Beta
 Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3.6
 Classifier: Programming Language :: Python :: 3.7
 Classifier: Programming Language :: Python :: 3.8
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy
 Classifier: Operating System :: POSIX :: Linux
 Classifier: Operating System :: MacOS :: MacOS X
 Classifier: Operating System :: Microsoft :: Windows
-Requires-Python: >=3.6
+Requires-Python: >=3.7
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: orthauth[sxpr,yaml]>=0.0.13
+Requires-Dist: orthauth[sxpr,yaml]>=0.0.18
 Requires-Dist: requests
 Provides-Extra: test
 Requires-Dist: pytest; extra == "test"

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/__init__.py

@@ -55,4 +55,4 @@ def get_right_id(uri):
     return pi
 
 
-__version__ = '0.0.1.dev21'
+__version__ = '0.0.1.dev22'

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/conventions/type.py

@@ -13,19 +13,21 @@ class ConvTypeBytesHeader(ConventionsType):
         self.stop = stop
         self.sentinel = sentinel
 
-    def _findElement(self, element, chunk):
-        m = re.search(element, chunk)
+    def _findElement(self, element, chunk, prev):
+        to_search = prev + chunk
+        m = re.search(element, to_search)
         if m is not None:
+            lp = len(prev)
             # FIXME icky api
-            return m.start(), m.end()
+            return m.start() - lp, m.end() -lp
         else:
             return None, None
 
-    def findStart(self, chunk):
-        return self._findElement(self.start, chunk)
+    def findStart(self, chunk, prev):
+        return self._findElement(self.start, chunk, prev)
 
-    def findStop(self, chunk):
-        return self._findElement(self.stop, chunk)
+    def findStop(self, chunk, prev):
+        return self._findElement(self.stop, chunk, prev)
 
-    def findSentinel(self, chunk):
-        return self._findElement(self.sentinel, chunk)
+    def findSentinel(self, chunk, prev):
+        return self._findElement(self.sentinel, chunk, prev)

idlib-0.0.1.dev22/idlib/core.py

@@ -0,0 +1,77 @@
+import requests
+from . import exceptions as exc
+from .utils import log
+
+
+def resolution_chain(iri):
+    for head in resolution_chain_responses(iri):
+        yield head.url
+
+
+def try_get(s, head):
+    # see whether the server has a bad/broken support for head requests
+    # sometimes they return e.g. 400 instead of 405, and really they should
+    # just work because they work correctly with get request ...
+    head = s.get(head.url, stream=True)
+    if head.ok:
+        log.info(f'bad HEAD implementation {head.url}')
+    else:
+        content = head.content
+        if content:
+            if head.headers['Content-Type'] == 'application/json':
+                try:
+                    j = head.json()
+                    log.error(f'{head.url} error json {j}')
+                except Exception as e:
+                    log.error(f'{head.url} error text {head.text}')
+
+    head.close()
+    return head
+
+
+def resolution_chain_responses(iri, raise_on_final=True):
+    #doi = doi  # TODO
+    s = requests.Session()
+    head = s.head(iri, allow_redirects=False)
+    if head.status_code < 400:
+        yield head
+    else:
+        head = try_get(s, head)
+        yield head
+
+    while head.is_redirect and head.status_code < 400:  # FIXME redirect loop issue
+        yield head.next
+        try:
+            head = s.send(head.next)
+        except requests.exceptions.SSLError as e:
+            raise exc.InbetweenError(msg) from e
+
+        if head.status_code < 400:
+            yield head
+        else:
+            head = try_get(s, head)
+            yield head
+
+        if not head.is_redirect:
+            break
+
+    if raise_on_final:  # we still want the chain ... null pointer error comes later?
+        if head.status_code == 404:
+            head.raise_for_status()  # probably a permissions issue
+        elif head.status_code >= 500:
+            msg = f'Remote in error due to {head.status_code} at {head.url}\n'
+            raise exc.RemoteError(msg)
+        elif head.status_code >= 400:
+            msg = f'Nothing found due to {head.status_code} at {head.url}\n'
+            raise exc.ResolutionError(msg)
+
+    if head.status_code >= 400:
+        # XXX this seems like the "right" thing to do, but it will
+        # probably break a bunch of stuff, but at least this way it
+        # will be visible now, obviously yeilding None as the final
+        # element in the reference chain conflates all the possible
+        # reasons so raising is a better solution and is the default
+        # for a reason but at least this way the user knows that the
+        # final element of the chain dereference to nothing instead of
+        # assuming that the last element succeeded
+        yield None

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/exceptions.py

@@ -36,7 +36,7 @@ class CouldNotReachReferentError(CouldNotReachError):
 
         Note that in a system that is not based on redirects
         this is an error that could be returned by the index
-        but it would appear as an referent unknown error. """
+        but it would appear as a referent unknown error. """
 
     # but its supertype is ReferentUnknown or ReferentUnreachable
 
@@ -59,6 +59,14 @@ class ResolutionError(IdlibError):
     # we should split into NetworkError/TimeoutError
     # and then have RemoteError, because Resolution error is
     # ambiguous, it is the superclass for all of them I think
+    # XXX as used, this is ClientError, and for all but 404
+    # and 410 the response means we can't actually say whether
+    # it is a true null pointer, the naming is thus very bad
+    # because as used this is not a network error, not a timeout
+    # error, it is the server saying i do not/cannot understand
+    # that is MessageReceivedButError which happens before we
+    # know if the id exists and probably actually covers most
+    # of the IdExistsButError cases
 
 
 class MalformedIdentifierError(IdlibError):

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/streams.py

@@ -606,7 +606,7 @@ class StreamUri(Stream):
         if not self.identifier_actionable in dc:
             try:
                 dc[self.identifier_actionable] = tuple(
-                    StringProgenitor(resp.url, progenitor=resp)
+                    resp if resp is None else StringProgenitor(resp.url, progenitor=resp)
                     for resp in
                     self._resolution_chain_responses(self.identifier_actionable,
                                                      raise_on_final=False))
@@ -668,7 +668,8 @@ class StreamUri(Stream):
     @cache_result
     def headers(self):
         uri = self.dereference()
-        return uri.progenitor().headers
+        if uri is not None:
+            return uri.progenitor().headers
 
     def data(self, mimetype_accept=None):
         # FIXME TODO should the data associated with the doi

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/systems/doi.py

@@ -26,8 +26,8 @@ class DoiId(oq.OntId, idlib.Identifier, idlib.Stream):  # also _technically_ a s
     _namespaces = _DoiPrefixes
     _local_conventions = _namespaces
     _id_class = str  # eventually we have to drop down to something
-    local_regex = '^10\.[0-9]{4,9}[-._;()/:a-zA-Z0-9]+$'
-    canonical_regex = '^https:\/\/doi.org/10\.[0-9]{4,9}[-._;()/:a-zA-Z0-9]+$'
+    local_regex = '^10\\.[0-9]{4,9}[-._;()/:a-zA-Z0-9]+$'
+    canonical_regex = '^https://doi.org/10\\.[0-9]{4,9}[-._;()/:a-zA-Z0-9]+$'
 
     def __new__(cls, doi_in_various_states_of_mangling=None, iri=None):
         if doi_in_various_states_of_mangling is None and iri is not None:

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/systems/orcid.py

@@ -1,3 +1,4 @@
+import random
 from pathlib import Path
 import ontquery as oq  # temporary implementation detail
 import idlib
@@ -13,6 +14,22 @@ from idlib.config import auth
 # but I need more information ...
 
 
+def genorcid():
+    """ generate a random orcid-like id suffix """
+    _09 = '09'
+    _01 = ''.join([str(_) for _ in range(2)])
+    _0_9 = ''.join([str(_) for _ in range(10)])
+    base = (
+        '000' + random.choice(_09) + '-' +
+        '00' + random.choice(_01) + random.choice(_0_9) + '-' +
+        ''.join(random.choice(_0_9) for _ in range(4))
+        + '-' +
+        ''.join(random.choice(_0_9) for _ in range(3)))
+    cs = OrcidId._digit_checksum(base.replace('-', ''))
+    fd = 'X' if cs == 10 else str(cs)
+    return base + fd
+
+
 class OrcidPrefixes(conv.QnameAsLocalHelper, oq.OntCuries):
     # set these manually since, sigh, factory patterns
     _dict = {}
@@ -68,6 +85,16 @@ class OrcidId(oq.OntId, idlib.Identifier):
 
         return self
 
+    @staticmethod
+    def _digit_checksum(digits):
+        total = 0
+        for digit_string in digits:
+            total = (total + int(digit_string)) * 2
+
+        remainder = total % 11
+        result = (12 - remainder) % 11
+        return result
+
     @property
     def checksumValid(self):
         """ see
@@ -77,12 +104,7 @@ class OrcidId(oq.OntId, idlib.Identifier):
         try:
             *digits, check_string = self.suffix.replace('-', '')
             check = 10 if check_string == 'X' else int(check_string)
-            total = 0
-            for digit_string in digits:
-                total = (total + int(digit_string)) * 2
-
-            remainder = total % 11
-            result = (12 - remainder) % 11
+            result = self._digit_checksum(digits)
             return result == check
         except ValueError as e:
             raise self.OrcidChecksumError(self) from e
@@ -186,3 +208,25 @@ class Orcid(idlib.HelperNoData, idlib.Stream):
         return (self.identifier.iri
                 if asType is None else
                 asType(self.identifier.iri))
+
+
+class OrcidSandboxPrefixes(conv.QnameAsLocalHelper, oq.OntCuries):
+    # set these manually since, sigh, factory patterns
+    _dict = {}
+    _n_to_p = {}
+    _strie = {}
+    _trie = {}
+
+
+OrcidSandboxPrefixes({'orcid': 'https://sandbox.orcid.org/',
+                      'ORCID': 'https://sandbox.orcid.org/',
+                      'orcid.pub.3': 'https://pub.sandbox.orcid.org/v3.0/',})
+
+
+class OrcidIdSandbox(OrcidId):
+    _namespaces = OrcidSandboxPrefixes
+    canonical_regex = (f'^https://sandbox.orcid.org/{OrcidId._common}$')
+
+
+class OrcidSandbox(Orcid):
+    _id_class = OrcidIdSandbox

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib/systems/ror.py

@@ -167,6 +167,7 @@ class Ror(formats.Rdf, idlib.HelperNoData, idlib.Stream):
         'Healthcare': 'Institution',
         'Facility':   'CoreFacility',
         'Nonprofit':  'Nonprofit',
+        'Funder':     'Funder',  # FIXME funder is a role ...
         'Other':      'Institution',
     }
     @property

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: idlib
-Version: 0.0.1.dev21
+Version: 0.0.1.dev22
 Summary: A library for working with identifiers of all kinds.
 Home-page: https://github.com/tgbugs/idlib
 Author: Tom Gillespie
@@ -9,21 +9,22 @@ License: MIT
 Keywords: python persistent identifiers
 Classifier: Development Status :: 4 - Beta
 Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3.6
 Classifier: Programming Language :: Python :: 3.7
 Classifier: Programming Language :: Python :: 3.8
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy
 Classifier: Operating System :: POSIX :: Linux
 Classifier: Operating System :: MacOS :: MacOS X
 Classifier: Operating System :: Microsoft :: Windows
-Requires-Python: >=3.6
+Requires-Python: >=3.7
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: orthauth[sxpr,yaml]>=0.0.13
+Requires-Dist: orthauth[sxpr,yaml]>=0.0.18
 Requires-Dist: requests
 Provides-Extra: test
 Requires-Dist: pytest; extra == "test"

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/idlib.egg-info/requires.txt

@@ -1,4 +1,4 @@
-orthauth[sxpr,yaml]>=0.0.13
+orthauth[sxpr,yaml]>=0.0.18
 requests
 
 [oauth]

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/setup.py

@@ -37,12 +37,13 @@ setup(name='idlib',
       classifiers=[
           'Development Status :: 4 - Beta',
           'License :: OSI Approved :: MIT License',
-          'Programming Language :: Python :: 3.6',
           'Programming Language :: Python :: 3.7',
           'Programming Language :: Python :: 3.8',
           'Programming Language :: Python :: 3.9',
           'Programming Language :: Python :: 3.10',
           'Programming Language :: Python :: 3.11',
+          'Programming Language :: Python :: 3.12',
+          'Programming Language :: Python :: 3.13',
           'Programming Language :: Python :: Implementation :: CPython',
           'Programming Language :: Python :: Implementation :: PyPy',
           'Operating System :: POSIX :: Linux',
@@ -57,9 +58,9 @@ setup(name='idlib',
           'idlib.formats',
           'idlib.systems',
       ],
-      python_requires='>=3.6',
+      python_requires='>=3.7',
       tests_require=tests_require,
-      install_requires=['orthauth[yaml,sxpr]>=0.0.13', 'requests'],
+      install_requires=['orthauth[yaml,sxpr]>=0.0.18', 'requests'],
       extras_require={'test': tests_require,
                       'org': org_require,
                       'rdf': rdf_require,

{idlib-0.0.1.dev21 → idlib-0.0.1.dev22}/test/test_identifiers.py

@@ -73,12 +73,27 @@ class TestOrcid(HelpTestStreams, unittest.TestCase):
     def test_asType(self):
         import rdflib
         from idlib.formats import rdf as _bind_rdf
-        o = idlib.Orcid(self.ids[0])
+        o = self.stream(self.ids[0])
         nt = o.asType(rdflib.URIRef)
         # still haven't decided if this is a good idea or not
         assert str(o) != str(nt), 'string representation of streams should not match id ???'
 
 
+class TestOrcidSandbox(TestOrcid):
+    stream = idlib.systems.orcid.OrcidSandbox
+    ids = [
+        'https://sandbox.orcid.org/0009-0007-0266-5208',
+        'https://sandbox.orcid.org/0000-0002-9383-616X',
+        'https://sandbox.orcid.org/0000-0002-9267-5456',
+        'https://sandbox.orcid.org/0000-0002-0391-3177',
+        'https://sandbox.orcid.org/0009-0002-1173-8142',
+    ]
+    ids_bad = []
+    for __i in ids:  # LOL PYTHON class scope list comprehensions
+        ids_bad.append(__i.replace('https', 'http'))
+    del __i
+
+
 class TestRor(HelpTestStreams, unittest.TestCase):
     stream = idlib.Ror
     ids = [

idlib-0.0.1.dev21/idlib/core.py

@@ -1,33 +0,0 @@
-import requests
-from . import exceptions as exc
-
-
-def resolution_chain(iri):
-    for head in resolution_chain_responses(iri):
-        yield head.url
-
-
-def resolution_chain_responses(iri, raise_on_final=True):
-    #doi = doi  # TODO
-    s = requests.Session()
-    head = requests.head(iri)
-    yield head
-    while head.is_redirect and head.status_code < 400:  # FIXME redirect loop issue
-        yield head.next
-        try:
-            head = s.send(head.next)
-        except requests.exceptions.SSLError as e:
-            raise exc.InbetweenError(msg) from e
-
-        yield head
-        if not head.is_redirect:
-            break
-
-    if raise_on_final:  # we still want the chain ... null pointer error comes later?
-        if head.status_code == 404:
-            head.raise_for_status()  # probably a permissions issue
-        elif head.status_code >= 500:
-            raise exc.RemoteError()
-        elif head.status_code >= 400:
-            msg = f'Nothing found due to {head.status_code} at {head.url}\n'
-            raise exc.ResolutionError(msg)