idlib 0.0.1.dev20__py2.py3-none-any.whl → 0.0.1.dev22__py2.py3-none-any.whl

idlib/__init__.py

@@ -55,4 +55,4 @@ def get_right_id(uri):
     return pi
 
 
-__version__ = '0.0.1.dev20'
+__version__ = '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/core.py

@@ -1,5 +1,6 @@
 import requests
 from . import exceptions as exc
+from .utils import log
 
 
 def resolution_chain(iri):
@@ -7,11 +8,37 @@ def resolution_chain(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 = requests.head(iri)
-    yield head
+    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:
@@ -19,7 +46,12 @@ def resolution_chain_responses(iri, raise_on_final=True):
         except requests.exceptions.SSLError as e:
             raise exc.InbetweenError(msg) from e
 
-        yield head
+        if head.status_code < 400:
+            yield head
+        else:
+            head = try_get(s, head)
+            yield head
+
         if not head.is_redirect:
             break
 
@@ -27,7 +59,19 @@ def resolution_chain_responses(iri, raise_on_final=True):
         if head.status_code == 404:
             head.raise_for_status()  # probably a permissions issue
         elif head.status_code >= 500:
-            raise exc.RemoteError()
+            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/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/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/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/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 = {}
@@ -29,10 +46,9 @@ OrcidPrefixes({'orcid': 'https://orcid.org/',
 class OrcidId(oq.OntId, idlib.Identifier):
     _namespaces = OrcidPrefixes
     _local_conventions = _namespaces
-    local_regex = ('^0000-000(1-[5-9]|2-[0-9]|3-[0-4])'
-                   '[0-9][0-9][0-9]-[0-9][0-9][0-9]([0-9]|X)$')
-    canonical_regex = ('^https://orcid.org/0000-000(1-[5-9]|2-[0-9]|3-'
-                       '[0-4])[0-9][0-9][0-9]-[0-9][0-9][0-9]([0-9]|X)$')
+    _common = '000[09]-00[01][0-9]-[0-9]{4}-[0-9]{3}([0-9]|X)'
+    local_regex = (f'^{_common}$')
+    canonical_regex = (f'^https://orcid.org/{_common}$')
 
     class OrcidMalformedError(exc.IdlibError):
         """ WHAT HAVE YOU DONE!? """
@@ -69,6 +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
@@ -78,12 +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
@@ -187,3 +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/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.dev20.dist-info → idlib-0.0.1.dev22.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: idlib
-Version: 0.0.1.dev20
+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
@@ -9,34 +9,35 @@ 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: oauth
-Requires-Dist: google-auth-oauthlib ; extra == 'oauth'
+Provides-Extra: test
+Requires-Dist: pytest; extra == "test"
+Requires-Dist: joblib>=1.1.0; extra == "test"
+Requires-Dist: beautifulsoup4[html5lib]; extra == "test"
+Requires-Dist: pyontutils>=0.1.28; extra == "test"
+Requires-Dist: google-auth-oauthlib; extra == "test"
 Provides-Extra: org
-Requires-Dist: beautifulsoup4[html5lib] ; extra == 'org'
+Requires-Dist: beautifulsoup4[html5lib]; extra == "org"
 Provides-Extra: rdf
-Requires-Dist: pyontutils >=0.1.28 ; extra == 'rdf'
-Provides-Extra: test
-Requires-Dist: pytest ; extra == 'test'
-Requires-Dist: joblib >=1.1.0 ; extra == 'test'
-Requires-Dist: beautifulsoup4[html5lib] ; extra == 'test'
-Requires-Dist: pyontutils >=0.1.28 ; extra == 'test'
-Requires-Dist: google-auth-oauthlib ; extra == 'test'
+Requires-Dist: pyontutils>=0.1.28; extra == "rdf"
+Provides-Extra: oauth
+Requires-Dist: google-auth-oauthlib; extra == "oauth"
 
 # idlib
 [![PyPI version](https://badge.fury.io/py/idlib.svg)](https://pypi.org/project/idlib/)

{idlib-0.0.1.dev20.dist-info → idlib-0.0.1.dev22.dist-info}/RECORD

@@ -1,35 +1,35 @@
-idlib/__init__.py,sha256=qnrq_qOQKd6AyZuvFQHZzYTLz8YGVe1TBdGtZmt89dM,1696
+idlib/__init__.py,sha256=FdowWsUq9bwbPqZPU2sTU1iWMxm1KetrH4kknNBlMmQ,1696
 idlib/auth-config.py,sha256=tYM2aYCYJFG5NB4Af1dSn6YxhafttOp7W7aSc-QbVV8,460
 idlib/cache.py,sha256=lR1I3iEYvBhNeqoptS-Sj7SpIHt1K65nYlBrTK-ROVQ,3875
 idlib/config.py,sha256=Gli2tZlbKxBos8sXDoDK1g9mngfDw1ZWpBt_7MDZH-g,76
-idlib/core.py,sha256=zLqHrSwtgbQll-04J2B5j8W69AXpWT_Ql5n0eBHp2pk,1055
+idlib/core.py,sha256=DfJoPe_xqJJducri1ub7e_anJYuhJ37-ZxvZCuuOpaY,2704
 idlib/edit.py,sha256=s3yRtvNelQUA8fziu_igZeoFsW4xCK_kgfNVv3TS0iE,3635
-idlib/exceptions.py,sha256=4qUTUrGkCMu2jE0TO4_hUCP-7XDCrk6FE08Ff-c1N94,4242
+idlib/exceptions.py,sha256=QC3Jq98iwezcqB7SmDr-UEOqdNCnvEXVkKB0eJLIwZs,4714
 idlib/from_oq.py,sha256=HlgepGbgSUbOt5R79ALRhgiLOEvU6TSu3evB8zq7N84,94170
 idlib/identifiers.py,sha256=s39bAS-fCQrlwAbMtaa7tjSnqKq4YPlfSbfJ6ju2o7I,3861
-idlib/streams.py,sha256=3L5JnSDzbodHET66Mua4zlfJEZbsXACIsWtVd-iDUZM,26381
+idlib/streams.py,sha256=HGcs_S7bOiUYfiSjZNGLXo5bJI7yd0bYjc8BKkdqECI,26439
 idlib/utils.py,sha256=zItGz-lXovZytSENUZ9q7A7gLr5uK_9VMawokGSskbk,5127
 idlib/apis/__init__.py,sha256=q2-iSMUDPlYbjqGWa1mtbe2G95N7wbjWHGj30BUdVDM,88
 idlib/apis/protocols_io.py,sha256=3PGKBwe9P_DxhScIG-L3ZFtnEHxgHaccLhRGVPZuYHg,3956
 idlib/conventions/__init__.py,sha256=ug6W8eNHuq9JZy5bwrZ6Rjfu8r81Rrnm0nzApmvvMzw,204
 idlib/conventions/core.py,sha256=g0T1VoNpoC9afvUd0wp6Mu7tRxbPDFK59uwVXV-zTtA,940
 idlib/conventions/local.py,sha256=H1WnXTY06aLq0RB_SsTLVhMZiSP9GGvWuWm_zrFUNmQ,129
-idlib/conventions/type.py,sha256=2ATKqoWNaQrc4pirJ2-WucBTsJud7RC4g59Z62DSK7o,898
+idlib/conventions/type.py,sha256=PM_BkxP5a7BTOSsb77kMmMRF2u5xMdlptIa4QYIaE6Y,1013
 idlib/formats/__init__.py,sha256=yS6l0mbDwa2xjyxhgVzlqIf0LPqJmuhNVg8rfKSsh7M,528
 idlib/formats/rdf.py,sha256=yIKHPaPn-SUuUMkR4VD7KffVgQD8oNaZNxcqDBilxtg,466
 idlib/systems/__init__.py,sha256=3lWSK2DyzYX0d61DJo-gGytc-HoiOlNJl5so4s7Grrg,203
 idlib/systems/ark.py,sha256=i9JqmuTWsJABrLHR0-_FX_PBbbpeFa7X4CO65QhyOD8,111
-idlib/systems/doi.py,sha256=GVL_uiLmU-Xa9odi7pGkbIniU7tTDesKO1pNbsYm2Fs,13100
+idlib/systems/doi.py,sha256=PQAstSUocPpK9Ug7L-eKvJrQhn3jRU9WuKITGyZoL9g,13100
 idlib/systems/handle.py,sha256=Gg97IuysdAdCX9j6EhuNfIddb1GDq0CmwW1SmuLkpgM,95
 idlib/systems/isni.py,sha256=TgiSTlDcSBLRro3gg_YfQJQkfaJ3saw0Au1OMhE23wU,89
-idlib/systems/orcid.py,sha256=3j17X59Fi-Yk_xhlJ6oPjb8BdzZD6_2BvbkIs1HeCPE,6745
+idlib/systems/orcid.py,sha256=aTVu_Q6fnaQvqmd6jeXdst6yKRTO-BEUMs7_4fdQK5E,7886
 idlib/systems/pmid.py,sha256=Na-4_n7qmJ3VAbG0RUwnNTg7o4Z8OkuL_Hluq_A3IO0,181
-idlib/systems/ror.py,sha256=q_NyvVP-Xe9DrzQMFZAC2WeQU6QyxSa6RzrqQ0P7gUA,8067
+idlib/systems/ror.py,sha256=xIS30mxKd0IcjwalBjnyUU7GKeeU5pxM3i9hSofG6u8,8129
 idlib/systems/rrid.py,sha256=XVHyHyiGC2OvbOSXmfU3_wswS6yUw65qZ3Pct0sjbfU,6886
 idlib/systems/urn.py,sha256=p222KjHgfF6y9srCm7kFWuWA9NK4lI-wEQUtSeuBqYk,113
 idlib/systems/web.py,sha256=vrSbQYWnxhe9pSgZcOqwLM3V6XQDLwv1QqWlhWWxbL0,1291
-idlib-0.0.1.dev20.dist-info/LICENSE,sha256=nEuJBsoHfpXGgf6ZyFGUhQJqxFWwhSYF-45pRAqTSb4,1080
-idlib-0.0.1.dev20.dist-info/METADATA,sha256=KrxuH4b2cYtMO0scJJCocd0lI8aIczeKhfIvOQLRToM,2012
-idlib-0.0.1.dev20.dist-info/WHEEL,sha256=DZajD4pwLWue70CAfc7YaxT1wLUciNBvN_TTcvXpltE,110
-idlib-0.0.1.dev20.dist-info/top_level.txt,sha256=xIk87gJPbeCKih9BVMP8PlEFdpcsGplenz0VQhx5sP8,6
-idlib-0.0.1.dev20.dist-info/RECORD,,
+idlib-0.0.1.dev22.dist-info/LICENSE,sha256=nEuJBsoHfpXGgf6ZyFGUhQJqxFWwhSYF-45pRAqTSb4,1080
+idlib-0.0.1.dev22.dist-info/METADATA,sha256=kjQ8xltmifA_qoV2GmEQAFeO5P3HzWVMc0MjiINC38o,2052
+idlib-0.0.1.dev22.dist-info/WHEEL,sha256=pxeNX5JdtCe58PUSYP9upmc7jdRPgvT0Gm9kb1SHlVw,109
+idlib-0.0.1.dev22.dist-info/top_level.txt,sha256=xIk87gJPbeCKih9BVMP8PlEFdpcsGplenz0VQhx5sP8,6
+idlib-0.0.1.dev22.dist-info/RECORD,,

{idlib-0.0.1.dev20.dist-info → idlib-0.0.1.dev22.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.43.0)
+Generator: setuptools (75.6.0)
 Root-Is-Purelib: true
 Tag: py2-none-any
 Tag: py3-none-any