ds-caselaw-marklogic-api-client 39.0.0__py3-none-any.whl → 39.1.0__py3-none-any.whl

caselawclient/Client.py

@@ -7,11 +7,12 @@ import warnings
 from datetime import datetime, time, timedelta
 from pathlib import Path
 from typing import Any, Optional, Type, Union
-from xml.etree import ElementTree
-from xml.etree.ElementTree import Element, ParseError, fromstring
+from xml.etree.ElementTree import Element
 
 import environ
 import requests
+from defusedxml import ElementTree
+from defusedxml.ElementTree import ParseError, fromstring
 from ds_caselaw_utils.types import NeutralCitationString
 from lxml import etree
 from requests.auth import HTTPBasicAuth
@@ -258,10 +259,12 @@ class MarklogicApiClient:
             return "Unknown error, Marklogic returned a null or empty response"
         try:
             xml = fromstring(content_as_xml)
-            return xml.find(
-                "message-code",
-                namespaces={"": "http://marklogic.com/xdmp/error"},
-            ).text  # type: ignore
+            return str(
+                xml.find(
+                    "message-code",
+                    namespaces={"": "http://marklogic.com/xdmp/error"},
+                ).text
+            )
         except (ParseError, TypeError, AttributeError):
             return "Unknown error, Marklogic returned a null or empty response"
 
@@ -682,10 +685,12 @@ class MarklogicApiClient:
         if content == "":
             return None
         response_xml = ElementTree.fromstring(content)
-        return response_xml.find(
-            "dls:annotation",
-            namespaces={"dls": "http://marklogic.com/xdmp/dls"},
-        ).text  # type: ignore
+        return str(
+            response_xml.find(
+                "dls:annotation",
+                namespaces={"dls": "http://marklogic.com/xdmp/dls"},
+            ).text
+        )
 
     def get_judgment_version(
         self,

caselawclient/models/documents/__init__.py

@@ -63,6 +63,9 @@ class Document:
     document_noun_plural = "documents"
     """ The noun for a plural of this document type. """
 
+    _default_reparse_document_type: Optional[str] = None
+    """ The default noun to pass to the parser when reparsing given the document type if known. This is used to determine how the document should be parsed and processed."""
+
     type_collection_name: str
 
     attributes_to_validate: list[tuple[str, bool, str]] = [
@@ -465,7 +468,6 @@ class Document:
         now = datetime.datetime.now(datetime.timezone.utc)
         self.api_client.set_property(self.uri, "last_sent_to_parser", now.isoformat())
 
-        parser_type_noun = {"judgment": "judgment", "press summary": "pressSummary"}[self.document_noun]
         checked_date: Optional[str] = (
             self.body.document_date_as_date.isoformat()
             if self.body.document_date_as_date and self.body.document_date_as_date > datetime.date(1001, 1, 1)
@@ -477,16 +479,18 @@ class Document:
         # values are "" from the API, we should pass None instead in this case.
 
         parser_instructions: ParserInstructionsDict = {
-            "documentType": parser_type_noun,
             "metadata": {
                 "name": self.body.name or None,
                 "cite": None,
                 "court": self.body.court or None,
                 "date": checked_date,
                 "uri": self.uri,
-            },
+            }
         }
 
+        if self._default_reparse_document_type:
+            parser_instructions["documentType"] = self._default_reparse_document_type
+
         ## TODO: Remove this hack around the fact that NCNs are assumed to be present for all documents' metadata, but actually different document classes may have different metadata
         if hasattr(self, "neutral_citation"):
             parser_instructions["metadata"]["cite"] = self.neutral_citation

caselawclient/models/judgments.py

@@ -24,6 +24,7 @@ class Judgment(NeutralCitationMixin, Document):
     document_noun = "judgment"
     document_noun_plural = "judgments"
     type_collection_name = "judgment"
+    _default_reparse_document_type = "judgment"
 
     def __init__(self, uri: DocumentURIString, *args: Any, **kwargs: Any) -> None:
         super().__init__(self.document_noun, uri, *args, **kwargs)

caselawclient/models/parser_logs.py

@@ -11,3 +11,4 @@ class ParserLog(Document):
     document_noun = "parser log"
     document_noun_plural = "parser logs"
     type_collection_name = "parser-log"
+    _default_reparse_document_type = "parserLog"

caselawclient/models/press_summaries.py

@@ -25,6 +25,7 @@ class PressSummary(NeutralCitationMixin, Document):
     document_noun = "press summary"
     document_noun_plural = "press summaries"
     type_collection_name = "press-summary"
+    _default_reparse_document_type = "pressSummary"
 
     def __init__(self, uri: DocumentURIString, *args: Any, **kwargs: Any) -> None:
         super().__init__(self.document_noun, uri, *args, **kwargs)

{ds_caselaw_marklogic_api_client-39.0.0.dist-info → ds_caselaw_marklogic_api_client-39.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: ds-caselaw-marklogic-api-client
-Version: 39.0.0
+Version: 39.1.0
 Summary: An API client for interacting with the underlying data in Find Caselaw.
 Keywords: national archives,caselaw
 Author: The National Archives
@@ -13,6 +13,7 @@ Classifier: Programming Language :: Python :: 3.13
 Requires-Dist: boto3 (>=1.26.112,<2.0.0)
 Requires-Dist: certifi (>=2025.7.9,<2025.8.0)
 Requires-Dist: charset-normalizer (>=3.0.0,<4.0.0)
+Requires-Dist: defusedxml (>=0.7.1,<0.8.0)
 Requires-Dist: django-environ (>=0.12.0)
 Requires-Dist: ds-caselaw-utils (>=2.0.0,<3.0.0)
 Requires-Dist: idna (>=3.4,<4.0)

{ds_caselaw_marklogic_api_client-39.0.0.dist-info → ds_caselaw_marklogic_api_client-39.1.0.dist-info}/RECORD

@@ -1,4 +1,4 @@
-caselawclient/Client.py,sha256=AGF7ud4x2hZFm6H--IUXa-n5woOWYmWz2yDzbPPCGdA,45193
+caselawclient/Client.py,sha256=A007KPqb-sn1pUMMVnKQuUfwCSPB8yipVxKhaWECsLY,45282
 caselawclient/__init__.py,sha256=DY-caubLDQWWingSdsBWgovDNXh8KcnkI6kwz08eIFk,612
 caselawclient/client_helpers/__init__.py,sha256=eucyUXwUqI72TPw-C5zLcHlMu4GtFY507a6lQc03lQY,5053
 caselawclient/client_helpers/search_helpers.py,sha256=R99HyRLeYHgsw2L3DOidEqlKLLvs6Tga5rKTuWQViig,1525
@@ -7,7 +7,7 @@ caselawclient/errors.py,sha256=JC16fEGq_MRJX-_KFzfINCV2Cqx8o6OWOt3C16rQd84,3142
 caselawclient/factories.py,sha256=eGj9TiZpmF3todW-08Ps7bHNMvByHqwEbgujRhvU_Yc,7382
 caselawclient/identifier_resolution.py,sha256=B5I1sD7o7YjzsXMECjbKjgiGLDda5bGhejsJ-lYpTIg,2429
 caselawclient/models/__init__.py,sha256=kd23EUpvaC7aLHdgk8farqKAQEx3lf7RvNT2jEatvlg,68
-caselawclient/models/documents/__init__.py,sha256=0MYorde8Ku9qUWLa5G_aQUBKwDeTRZ_1ebsV9RxIBuI,21256
+caselawclient/models/documents/__init__.py,sha256=MTSXGaHqsM2xuwgCuBPOskk0B_MPwd54oBZcVfPIA1A,21471
 caselawclient/models/documents/body.py,sha256=O1ZTV3KHo-YNi7Syd4oCV1CVSuRF7mcLXojwshyY4jg,6601
 caselawclient/models/documents/exceptions.py,sha256=te7PPQTDHjZ9EYVg5pVaiZfF00lMBFy333PHj8_mkC4,443
 caselawclient/models/documents/statuses.py,sha256=Cp4dTQmJOtsU41EJcxy5dV1841pGD2PNWH0VrkDEv4Q,579
@@ -20,10 +20,10 @@ caselawclient/models/identifiers/fclid.py,sha256=hj8z-VhXFrUHKOY6k_ItPvOakIvbhJ5
 caselawclient/models/identifiers/neutral_citation.py,sha256=bYAeXHVm_ls0aDTeYI4uv35iZmJGSKU4-H-iLh2xED0,2912
 caselawclient/models/identifiers/press_summary_ncn.py,sha256=t-x6PsEe2tz1uO1qZKXKK0TugkQYb_49O_xgjd_oiE4,801
 caselawclient/models/identifiers/unpacker.py,sha256=OpFBw1B6pqSuzcyHbnTY3dScHc2Ujt5StGRnh-tKE1Q,2592
-caselawclient/models/judgments.py,sha256=r40irgdEID-NeSNLm3OUdUBznMpRSwjD2SJrGlBgP8o,2208
+caselawclient/models/judgments.py,sha256=esg017Z2vAevk17jCyGhr8n9fa4FEoMz2SDFgeisV1Y,2256
 caselawclient/models/neutral_citation_mixin.py,sha256=jAac3PPuWyPdj9N-n-U_JfwkbgbSIXaqFVQahfu95do,2086
-caselawclient/models/parser_logs.py,sha256=30kF4w0GcowiMIFtymUkl7ZARanNh_PjDpJZezn-cA8,315
-caselawclient/models/press_summaries.py,sha256=PIq9RceZ7n7Z079tESfxhQbfxCmtTc2V2OeFtcn594s,2144
+caselawclient/models/parser_logs.py,sha256=iOhKTAAi87XQvxz1DHjF2lrqScD19g_c8EjSf0vPdfs,364
+caselawclient/models/press_summaries.py,sha256=rtrYs_3BazUXxdA2oYmIJ6YIAiVlKeyc1aSF9uvkJJU,2196
 caselawclient/models/utilities/__init__.py,sha256=LPhyrQwLKc5tIJUO8Bysn9wCiR6Z6jMMTksjOV4JH9U,1041
 caselawclient/models/utilities/aws.py,sha256=NTF2W2aNgbO72e5WklXZC2U2_GPbVeynjTS1Nqu6DcE,8561
 caselawclient/models/utilities/dates.py,sha256=WwORxVjUHM1ZFcBF6Qtwo3Cj0sATsnSECkUZ6ls1N1Q,492
@@ -86,7 +86,7 @@ caselawclient/xquery/xslt_transform.xqy,sha256=cccaFiGkCcvSfDv007UriZ3I4ak2nTLP1
 caselawclient/xquery_type_dicts.py,sha256=zuyDGTkcN6voOXCm3APXItZ-Ey6tZ2hdZummZWzjl50,6489
 caselawclient/xslt/modify_xml_live.xsl,sha256=gNjwBun2-UzOeeuf0wNjFtN3jXm1yrwqv_KT8r1slXw,2370
 caselawclient/xslt/sample.xsl,sha256=IG-v77stjwqiw25pguh391K-5DTKiX651WqILDZixm0,825
-ds_caselaw_marklogic_api_client-39.0.0.dist-info/LICENSE.md,sha256=fGMzyyLuQW-IAXUeDSCrRdsYW536aEWThdbpCjo6ZKg,1108
-ds_caselaw_marklogic_api_client-39.0.0.dist-info/METADATA,sha256=hn5g5HHm3R3WIpBTeSKsv0IY6uVxKY6m7AkCo-pr_y4,4319
-ds_caselaw_marklogic_api_client-39.0.0.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
-ds_caselaw_marklogic_api_client-39.0.0.dist-info/RECORD,,
+ds_caselaw_marklogic_api_client-39.1.0.dist-info/LICENSE.md,sha256=fGMzyyLuQW-IAXUeDSCrRdsYW536aEWThdbpCjo6ZKg,1108
+ds_caselaw_marklogic_api_client-39.1.0.dist-info/METADATA,sha256=y6C7CqUYSY2LHCM5RSuzZP6z2sZjRbkg6B1bqhLLJA8,4362
+ds_caselaw_marklogic_api_client-39.1.0.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+ds_caselaw_marklogic_api_client-39.1.0.dist-info/RECORD,,