ds-caselaw-marklogic-api-client 39.0.0__py3-none-any.whl → 39.2.0__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
@@ -55,6 +56,11 @@ from .errors import (
 )
 
 env = environ.Env()
+
+# Requests timeouts: https://requests.readthedocs.io/en/latest/user/advanced/
+CONNECT_TIMEOUT = float(os.environ.get("CONNECT_TIMEOUT", "3.05"))
+READ_TIMEOUT = float(os.environ.get("READ_TIMEOUT", "10.0"))
+
 ROOT_DIR = os.path.dirname(os.path.realpath(__file__))
 DEFAULT_XSL_TRANSFORM = "accessible-html.xsl"
 
@@ -169,6 +175,7 @@ class MarklogicApiClient:
     error_code_classes: dict[str, Type[MarklogicAPIError]] = {
         "XDMP-DOCNOTFOUND": MarklogicResourceNotFoundError,
         "XDMP-LOCKCONFLICT": MarklogicResourceLockedError,
+        "XDMP-LOCKED": MarklogicResourceLockedError,
         "DLS-UNMANAGED": MarklogicResourceUnmanagedError,
         "DLS-NOTCHECKEDOUT": MarklogicResourceNotCheckedOutError,
         "DLS-CHECKOUTCONFLICT": MarklogicCheckoutConflictError,
@@ -258,10 +265,12 @@ 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"
 
@@ -314,11 +323,13 @@ class MarklogicApiClient:
         self,
         vars: query_dicts.MarkLogicAPIDict,
         xquery_file_name: str,
+        timeout: tuple[float, float] = (CONNECT_TIMEOUT, READ_TIMEOUT),
     ) -> requests.Response:
         return self.eval(
             self._xquery_path(xquery_file_name),
             vars=json.dumps(vars),
             accept_header="application/xml",
+            timeout=timeout,
         )
 
     def _eval_and_decode(
@@ -682,10 +693,12 @@ 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,
@@ -719,6 +732,7 @@ class MarklogicApiClient:
         xquery_path: str,
         vars: str,
         accept_header: str = "multipart/mixed",
+        timeout: tuple[float, float] = (CONNECT_TIMEOUT, READ_TIMEOUT),
     ) -> requests.Response:
         headers = {
             "Content-type": "application/x-www-form-urlencoded",
@@ -738,6 +752,7 @@ class MarklogicApiClient:
             url=self._path_to_request_url(path),
             headers=headers,
             data=data,
+            timeout=timeout,
         )
         # Raise relevant exception for an erroneous response
         self._raise_for_status(response)
@@ -1047,10 +1062,6 @@ class MarklogicApiClient:
         if show_unpublished and not self.user_can_view_unpublished_judgments(
             self.username,
         ):
-            # The user cannot view unpublished judgments but is requesting to see them
-            logging.warning(
-                f"User {self.username} is attempting to view unpublished judgments but does not have that privilege.",
-            )
             return False
         return show_unpublished
 

caselawclient/__init__.py

@@ -1,4 +1,21 @@
-"""
+import os
+
+if os.getenv("PDOC_DYNAMIC_VERSION") == "1":
+    from pathlib import Path
+
+    import tomllib
+
+    pyproject_path = Path(__file__).parent.parent.parent / "pyproject.toml"
+    with pyproject_path.open("rb") as f:
+        __version__ = tomllib.load(f)["tool"]["poetry"]["version"]
+        __pip_version_string__ = f"~={__version__}"
+        __poetry_version_string__ = f' = "^{__version__}"'
+
+else:
+    __pip_version_string__ = ""
+    __poetry_version_string__ = ""
+
+__doc__ = f"""
 
 # Installation
 
@@ -15,7 +32,13 @@ poetry add ds-caselaw-marklogic-api-client
 or in your projects `requirements.txt` with:
 
 ```text
-ds-caselaw-marklogic-api-client~=12.0.0
+ds-caselaw-marklogic-api-client{__pip_version_string__}
+```
+
+or `pyproject.toml` for Poetry with:
+
+```text
+ds-caselaw-marklogic-api-client{__poetry_version_string__}
 ```
 
 # Usage

caselawclient/models/documents/__init__.py

@@ -9,6 +9,7 @@ from ds_caselaw_utils.courts import CourtNotFoundException
 from ds_caselaw_utils.types import NeutralCitationString
 from requests_toolbelt.multipart import decoder
 
+import caselawclient.models.documents.comparison as comparison
 from caselawclient.errors import (
     DocumentNotFoundError,
     NotSupportedOnVersion,
@@ -63,6 +64,9 @@ 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 +469,6 @@ 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 +480,18 @@ 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
@@ -579,3 +584,6 @@ class Document:
         return self.body.apply_xslt(
             "modify_xml_live.xsl", work_uri=work_uri, expression_uri=expression_uri, manifestation_uri=manifestation_uri
         )
+
+    def compare_to(self, that_doc: "Document") -> comparison.Comparison:
+        return comparison.Comparison(self, that_doc)

caselawclient/models/documents/comparison.py

@@ -0,0 +1,42 @@
+from typing import TypedDict
+
+import caselawclient.models.documents
+
+
+class AttributeComparison(TypedDict):
+    """Results from a comparison of one attribute across two documents"""
+
+    label: str
+    this_value: str
+    that_value: str
+    match: bool
+
+
+class Comparison(dict[str, AttributeComparison]):
+    def __init__(
+        self, this_doc: "caselawclient.models.documents.Document", that_doc: "caselawclient.models.documents.Document"
+    ):
+        # court, date, title
+        self["court"] = AttributeComparison(
+            label="court",
+            this_value=this_doc.court,
+            that_value=that_doc.court,
+            match=this_doc.court == that_doc.court,
+        )
+
+        self["date"] = AttributeComparison(
+            label="date",
+            this_value=this_doc.body.document_date_as_string,
+            that_value=that_doc.body.document_date_as_string,
+            match=this_doc.body.document_date_as_string == that_doc.body.document_date_as_string,
+        )
+        self["name"] = AttributeComparison(
+            label="name",
+            this_value=this_doc.name,
+            that_value=that_doc.name,
+            match=this_doc.name == that_doc.name,
+        )
+
+    def match(self) -> bool:
+        """Is this comparison an exact match across all attributes?"""
+        return all(x["match"] for x in self.values())

caselawclient/models/identifiers/__init__.py

@@ -166,8 +166,10 @@ class Identifier(ABC):
         """
         resolutions = [
             resolution
-            for resolution in api_client.resolve_from_identifier_value(identifier_value=self.value)
-            if resolution.identifier_namespace == self.schema.namespace
+            for resolution in api_client.resolve_from_identifier_value(
+                identifier_value=self.value, published_only=False
+            )
+            if resolution.identifier_namespace == self.schema.namespace and resolution.identifier_uuid != self.uuid
         ]
         if len(resolutions) > 0:
             return SuccessFailureMessageTuple(

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)

caselawclient/models/utilities/aws.py

@@ -96,6 +96,8 @@ def check_docx_exists(uri: DocumentURIString) -> bool:
     except botocore.exceptions.ClientError as e:
         if e.response["Error"]["Code"] == "404":
             return False
+        if e.response["Error"]["Code"] == "403":
+            e.add_note("403 on reading {s3_key}")
         raise
 
 

{ds_caselaw_marklogic_api_client-39.0.0.dist-info → ds_caselaw_marklogic_api_client-39.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: ds-caselaw-marklogic-api-client
-Version: 39.0.0
+Version: 39.2.0
 Summary: An API client for interacting with the underlying data in Find Caselaw.
 Keywords: national archives,caselaw
 Author: The National Archives
@@ -11,12 +11,13 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 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: certifi (>=2025.8.3,<2025.9.0)
 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)
-Requires-Dist: lxml (>=5.0.0,<6.0.0)
+Requires-Dist: lxml (>=6.0.0,<7.0.0)
 Requires-Dist: memoization (>=0.4.0,<0.5.0)
 Requires-Dist: mypy-boto3-s3 (>=1.26.104,<2.0.0)
 Requires-Dist: mypy-boto3-sns (>=1.26.69,<2.0.0)
@@ -36,7 +37,9 @@ This repository is part of the [Find Case Law](https://caselaw.nationalarchives.
 
 # MarkLogic API Client
 
-[![PyPI](https://img.shields.io/pypi/v/ds-caselaw-marklogic-api-client)](https://pypi.org/project/ds-caselaw-marklogic-api-client/) ![Libraries.io dependency status for latest release](https://img.shields.io/librariesio/release/pypi/ds-caselaw-marklogic-api-client)
+[![PyPI](https://img.shields.io/pypi/v/ds-caselaw-marklogic-api-client)](https://pypi.org/project/ds-caselaw-marklogic-api-client/)
+![Libraries.io dependency status for latest release](https://img.shields.io/librariesio/release/pypi/ds-caselaw-marklogic-api-client)
+![Code Coverage](https://img.shields.io/codecov/c/github/nationalarchives/ds-caselaw-custom-api-client)
 
 This is an API Client for connecting to Marklogic for The National Archive's Caselaw site.
 

{ds_caselaw_marklogic_api_client-39.0.0.dist-info → ds_caselaw_marklogic_api_client-39.2.0.dist-info}/RECORD

@@ -1,5 +1,5 @@
-caselawclient/Client.py,sha256=AGF7ud4x2hZFm6H--IUXa-n5woOWYmWz2yDzbPPCGdA,45193
-caselawclient/__init__.py,sha256=DY-caubLDQWWingSdsBWgovDNXh8KcnkI6kwz08eIFk,612
+caselawclient/Client.py,sha256=rcdgfEJ5N_Ajryf8CFHBlEWqvRB7J5IxqE1tCcudTqU,45496
+caselawclient/__init__.py,sha256=QZtsOB_GR5XfFnWMJ6E9_fBany-JXFIrQmzs1mD_KVg,1225
 caselawclient/client_helpers/__init__.py,sha256=eucyUXwUqI72TPw-C5zLcHlMu4GtFY507a6lQc03lQY,5053
 caselawclient/client_helpers/search_helpers.py,sha256=R99HyRLeYHgsw2L3DOidEqlKLLvs6Tga5rKTuWQViig,1525
 caselawclient/content_hash.py,sha256=0cPC4OoABq0SC2wYFX9-24DodNigeOqksDxgxQH_hUA,2221
@@ -7,25 +7,26 @@ 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=jD6lIPeGKOGhs3q0AOsQBhl92c7AqxpkDcf3I3EpOJk,21661
 caselawclient/models/documents/body.py,sha256=O1ZTV3KHo-YNi7Syd4oCV1CVSuRF7mcLXojwshyY4jg,6601
+caselawclient/models/documents/comparison.py,sha256=KwFZQByOcYcZKe8csjAntttACKq4BZb28n2VeV5rK54,1355
 caselawclient/models/documents/exceptions.py,sha256=te7PPQTDHjZ9EYVg5pVaiZfF00lMBFy333PHj8_mkC4,443
 caselawclient/models/documents/statuses.py,sha256=Cp4dTQmJOtsU41EJcxy5dV1841pGD2PNWH0VrkDEv4Q,579
 caselawclient/models/documents/transforms/html.xsl,sha256=XyUQLFcJ7_GwthWQ6ShU0bmzrgpl7xDFU-U8VLgOvEs,38258
 caselawclient/models/documents/xml.py,sha256=BVra2VL_0JyImM8GC3wdouu1tApy79C-e2dHvQyrXPE,2195
-caselawclient/models/identifiers/__init__.py,sha256=N2xcIOUlnG5w7yCzCAvGnlkN_rm6ceMHwA-CS1ZNoGk,7631
+caselawclient/models/identifiers/__init__.py,sha256=bcXiree1FKIlJklWggvS_IKMHMppAjDbadOpxCJx3yw,7727
 caselawclient/models/identifiers/collection.py,sha256=kGlziJiLAqoyd6LaaZ5tsgUf2fD6Y-7fv1It9S4-Otw,7448
 caselawclient/models/identifiers/exceptions.py,sha256=6LVjvx-UOwqkrpxU19ydmrphKNw0rcG5GXwjTFyf8Dk,130
 caselawclient/models/identifiers/fclid.py,sha256=hj8z-VhXFrUHKOY6k_ItPvOakIvbhJ5xEbZ04E2j7t8,1521
 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/aws.py,sha256=_JUoJatfC1m_etT5MDwbCrOHpdqqTRqGCyzzNiW1WRA,8660
 caselawclient/models/utilities/dates.py,sha256=WwORxVjUHM1ZFcBF6Qtwo3Cj0sATsnSECkUZ6ls1N1Q,492
 caselawclient/models/utilities/move.py,sha256=MXdUqkSiyqRb8YKs_66B6ICWn8EWM6DiJV95fuJO1Us,3610
 caselawclient/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -86,7 +87,7 @@ 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-39.2.0.dist-info/LICENSE.md,sha256=fGMzyyLuQW-IAXUeDSCrRdsYW536aEWThdbpCjo6ZKg,1108
+ds_caselaw_marklogic_api_client-39.2.0.dist-info/METADATA,sha256=FVy1r1XjQwAKBA4kTqk4TEVzKWjAMr7ooMHMHjz84BY,4466
+ds_caselaw_marklogic_api_client-39.2.0.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+ds_caselaw_marklogic_api_client-39.2.0.dist-info/RECORD,,