s

README.md

@@ -36,7 +36,7 @@ with Scythe("https://zenodo.org/oai2d") as scythe:
 `oaipmh-scythe` is built with:
 
 - [httpx](https://github.com/encode/httpx) for issuing HTTP requests
-- [lxml](https://github.com/lxml/lxml) for parsing XML responses
+- [xsdata](https://github.com/tefra/xsdata) for parsing XML responses
 
 ## Installation
 

pyproject.toml

@@ -201,7 +201,7 @@ ignore = [
 minversion = "8.0"
 addopts = ["-ra", "--showlocals", "--strict-markers", "--strict-config"]
 filterwarnings = ["error"]
-log_cli_level = "INFO"
+# log_cli_level = "INFO"
 xfail_strict = true
 testpaths = "tests"
 

src/oaipmh_scythe/client.py

@@ -34,7 +34,6 @@
 logger = logging.getLogger(__name__)
 
 USER_AGENT: str = f"oaipmh-scythe/{__version__}"
-OAI_NAMESPACE: str = "{http://www.openarchives.org/OAI/2.0/}"
 
 
 class Scythe:
@@ -86,7 +85,6 @@ def __init__(
         self.max_retries = max_retries
         self.retry_status_codes = retry_status_codes or (503,)
         self.default_retry_after = default_retry_after
-        self.oai_namespace = OAI_NAMESPACE
         self.encoding = encoding
         self.auth = auth
         self.timeout = timeout
@@ -158,8 +156,9 @@ def harvest(self, query: dict[str, str]) -> Response:
                 logger.warning("HTTP %d! Retrying after %d seconds...", http_response.status_code, retry_after)
                 time.sleep(retry_after)
                 http_response = self._request(query)
-        http_response.raise_for_status()
-        return _build_response(http_response)
+        # http_response.raise_for_status()
+        metadata_prefix = query.get("metadataPrefix")
+        return _build_response(http_response, metadata_prefix)
 
     def _request(self, query: dict[str, str]) -> httpx.Response:
         """Send an HTTP request to the OAI server using the configured HTTP method and given query parameters.
@@ -255,7 +254,6 @@ def list_identifiers(
             cannotDisseminateFormat: If the specified metadata_prefix is not supported by the OAI server.
             noRecordsMatch: If no records match the provided criteria.
             noSetHierarchy: If set-based harvesting is requested but the OAI server does not support sets.
-
         """
         _query = {
             "verb": Verb.LIST_IDENTIFIERS.value,
@@ -287,7 +285,6 @@ def list_sets(self, resumption_token: str | None = None) -> Iterator[Response |
         Raises:
             badResumptionToken: If the provided resumption token is invalid or expired.
             noSetHierarchy: If the OAI server does not support sets or has no set hierarchy available.
-
         """
         _query = {
             "verb": Verb.LIST_SETS.value,
@@ -308,10 +305,7 @@ def identify(self) -> Identify:
         Returns:
             Identify: An object encapsulating the server's identify response, which contains various pieces of information
             about the OAI server.
-
         """
-        # return Identify(self.harvest(query))
-
         query = {"verb": Verb.IDENTIFY.value}
         response = self.harvest(query)
         return response.parsed.identify
@@ -337,14 +331,14 @@ def get_record(self, identifier: str, metadata_prefix: str = "oai_dc") -> Respon
             cannotDisseminateFormat: If the specified metadata_prefix is not supported by the OAI server
                 for the requested record.
             idDoesNotExist: If the specified identifier does not correspond to any record in the OAI server.
-
         """
         query = {
             "verb": Verb.GET_RECORD.value,
             "identifier": identifier,
             "metadataPrefix": metadata_prefix,
         }
-        return next(iter(self.iterator(self, query)))
+        response = self.harvest(query)
+        return response.parsed.get_record.record
 
     def list_metadata_formats(self, identifier: str | None = None) -> Iterator[Response | MetadataFormat]:
         """Issue a ListMetadataFormats request to the OAI server.
@@ -367,7 +361,6 @@ def list_metadata_formats(self, identifier: str | None = None) -> Iterator[Respo
         Raises:
             idDoesNotExist: If the specified identifier does not correspond to any record in the OAI server.
             noMetadataFormats: If there are no metadata formats available for the requested record or repository.
-
         """
         _query = {
             "verb": Verb.LIST_METADATA_FORMATS.value,

src/oaipmh_scythe/iterator.py

@@ -19,7 +19,7 @@
 from __future__ import annotations
 
 from abc import ABC, abstractmethod
-from enum import StrEnum
+from enum import Enum
 from operator import attrgetter
 from typing import TYPE_CHECKING
 
@@ -29,40 +29,41 @@
     from collections.abc import Iterator
 
     from oaipmh_scythe import Scythe
-    from oaipmh_scythe.models import Header, MetadataFormat, Record, ResumptionToken, Set
+    from oaipmh_scythe.models import Header, Item, MetadataFormat, Record, ResumptionToken, Set
     from oaipmh_scythe.response import Response
 
 
-class GetRecord(StrEnum):
+class GetRecord(Enum):
     attribute = "get_record"
     element = "record"
 
 
-class ListIdentifiers(StrEnum):
+class ListIdentifiers(Enum):
     attribute = "list_identifiers"
     element = "header"
 
 
-class ListRecords(StrEnum):
+class ListRecords(Enum):
     attribute = "list_records"
     element = "record"
 
 
-class ListSets(StrEnum):
+class ListSets(Enum):
     attribute = "list_sets"
     element = "set"
 
 
-class ListMetadataFormats(StrEnum):
-    attribute = "list_metadataformats"
-    element = "metadataformat"
+class ListMetadataFormats(Enum):
+    attribute = "list_metadata_formats"
+    element = "metadata_format"
 
 
 MAPPING = {
     Verb.LIST_IDENTIFIERS.value: ListIdentifiers,
     Verb.GET_RECORD.value: GetRecord,
     Verb.LIST_RECORDS.value: ListRecords,
     Verb.LIST_SETS.value: ListSets,
+    Verb.LIST_METADATA_FORMATS.value: ListMetadataFormats,
 }
 
 
@@ -112,11 +113,14 @@ def _get_resumption_token(self) -> ResumptionToken | None:
         Returns:
             A ResumptionToken instance if a token is found in the response, otherwise None.
         """
-        if self.response is not None:
+        if self.response is None:
+            return None
+        try:
             attribute = MAPPING[self.verb].attribute.value
             parsed_data = getattr(self.response.parsed, attribute)
             return parsed_data.resumption_token
-        return None
+        except AttributeError:
+            return None
 
     def _next_response(self) -> None:
         """Request the next batch of data from the OAI server using the current resumption token.
@@ -132,7 +136,7 @@ def _next_response(self) -> None:
         self.resumption_token = self._get_resumption_token()
 
 
-class OAIResponseIterator(BaseOAIIterator):
+class ResponseIterator(BaseOAIIterator):
     """An iterator class for iterating over OAI responses obtained via the OAI-PMH protocol.
 
     This iterator specifically handles the iteration of OAIResponse objects, allowing for seamless
@@ -176,7 +180,6 @@ def __init__(self, scythe: Scythe, query: dict[str, str], ignore_deleted: bool =
         attribute = MAPPING[query["verb"]].attribute.value
         element = MAPPING[query["verb"]].element.value
         self.items_getter = attrgetter(f"{attribute}.{element}")
-        print(self.items_getter)
         super().__init__(scythe, query, ignore_deleted)
 
     def _next_response(self) -> None:
@@ -191,7 +194,7 @@ def _next_response(self) -> None:
         else:
             self._items = iter(())
 
-    def __iter__(self) -> Iterator[Header | Record | Set | MetadataFormat]:
+    def __iter__(self) -> Iterator[Item]:
         """Iterate over individual OAI items from the response.
 
         Go through the items in the OAI-PMH response, applying any necessary mapping and handling
@@ -202,13 +205,7 @@ def __iter__(self) -> Iterator[Header | Record | Set | MetadataFormat]:
         """
         while True:
             for item in self._items:
-                status = None
-                if isinstance(item, Header):
-                    status = item.status
-                elif isinstance(item, Record):
-                    status = item.header.status
-
-                if self.ignore_deleted and status and status.DELETED:
+                if self.ignore_deleted and item.deleted:
                     continue
                 yield item
             if self.resumption_token:

src/oaipmh_scythe/models/.xsdata.xml

@@ -30,10 +30,12 @@
     <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap/" replace="soap"/>
     <Substitution type="package" search="http://schemas.xmlsoap.org/wsdl/soap12/" replace="soap12"/>
     <Substitution type="package" search="http://schemas.xmlsoap.org/soap/envelope/" replace="soapenv"/>
-
-    <!-- # TODO: replace "ElementType" with "Element" -->
-    <!-- <Substitution type="class" search="(.*)(T|t)ype$" replace="\1"/> -->
+    <!-- replaces e.g. "ElementType" with "Element" -->
     <Substitution type="class" search="(.*)Type$" replace="\1"/>
   </Substitutions>
-  <Extensions/>
+  <Extensions>
+    <!-- add mixin classes -->
+    <Extension type="class" class="Header" import="oaipmh_scythe.models.mixins.HeaderMixin" prepend="false" applyIfDerived="false"/>
+    <Extension type="class" class="Record" import="oaipmh_scythe.models.mixins.RecordMixin" prepend="false" applyIfDerived="false"/>
+  </Extensions>
 </Config>