ds-caselaw-marklogic-api-client 39.0.0__tar.gz → 39.2.0__tar.gz

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/PKG-INFO

@@ -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 → ds_caselaw_marklogic_api_client-39.2.0}/README.md

@@ -4,7 +4,9 @@ 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 → ds_caselaw_marklogic_api_client-39.2.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "ds-caselaw-marklogic-api-client"
-version = "39.0.0"
+version = "39.2.0"
 description = "An API client for interacting with the underlying data in Find Caselaw."
 authors = ["The National Archives"]
 homepage = "https://github.com/nationalarchives/ds-caselaw-custom-api-client"
@@ -12,14 +12,14 @@ packages = [
 
 [tool.poetry.dependencies]
 python = "^3.10.0"
-certifi = ">=2025.7.9,<2025.8.0"
+certifi = ">=2025.8.3,<2025.9.0"
 charset-normalizer = "^3.0.0"
 django-environ = ">=0.12.0"
 idna = "^3.4"
 requests = "^2.28.2"
 requests-toolbelt = ">=0.10.1,<1.1.0"
 memoization = "^0.4.0"
-lxml = "^5.0.0"
+lxml = "^6.0.0"
 ds-caselaw-utils = "^2.0.0"
 boto3 = "^1.26.112"
 typing-extensions = "^4.7.1"
@@ -29,15 +29,17 @@ pytz = ">2024"
 python-dateutil = "^2.9.0-post.0"
 saxonche = "^12.5.0"
 sqids = "^0.5.0"
+defusedxml = "^0.7.1"
 
 [tool.poetry.group.dev.dependencies]
-coverage = "7.9.2"
+coverage = "7.10.3"
 pytest = "8.4.1"
+pytest-cov = "6.2.1"
 beautifulsoup4 = "4.13.4"
-responses = "0.25.7"
+responses = "0.25.8"
 python-dotenv = "1.1.1"
-time-machine = "2.16.0"
-moto = {version = "5.1.8", extras = ["all"]}
+time-machine = "2.17.0"
+moto = {version = "5.1.9", extras = ["all"]}
 
 [tool.poetry.group.docs]
 optional = true
@@ -69,14 +71,15 @@ line-length = 120
 
 [tool.ruff.lint]
 ignore = ["E501", "G004", "PLR2004", "RUF005", "RUF012", "UP040"] # longlines, fstrings in logs, magic values, consider not concat, mutable classbits, type instead of TypeAlias
-extend-select = ["W", "I", "SLF", "SIM"]
-# extend-select = [ "B", "Q", "C90", "I", "UP", "YTT", "ASYNC", "S", "BLE", "A", "COM", "C4", "DTZ", "T10", "DJ", "EM", "EXE", "FA",
+extend-select = ["W", "I", "SLF", "SIM", "C90", "S"]
+# extend-select = [ "B", "Q", "I", "UP", "YTT", "ASYNC", "BLE", "A", "COM", "C4", "DTZ", "T10", "DJ", "EM", "EXE", "FA",
 #                  "ISC", "ICN", "G", "INP", "PIE", "T20", "PYI", "PT", "Q", "RSE", "RET", "SLOT", "TID", "TCH", "INT", "PTH",
 #                  "FIX", "PGH", "PL", "TRY", "FLY", "PERF", "RUF"]
 unfixable = ["ERA"]
 
 [tool.ruff.lint.extend-per-file-ignores]
-"tests/*" = ["S101"]  # `assert` is fine in tests
+"tests/*" = ["S101"] # `assert` is fine in tests
+"smoketest/*" = ["S101"] # `assert` is fine in tests
 "tests/client/test_client.py" = ["SLF001"]  # TODO: This really shouldn't be the case, but it's not important to fix right now.
 
 # things skipped:

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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
 

ds_caselaw_marklogic_api_client-39.2.0/src/caselawclient/__init__.py

@@ -0,0 +1,63 @@
+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
+
+Include the API client in your project using Pip or Poetry:
+
+```bash
+pip install ds-caselaw-marklogic-api-client
+```
+
+```bash
+poetry add ds-caselaw-marklogic-api-client
+```
+
+or in your projects `requirements.txt` with:
+
+```text
+ds-caselaw-marklogic-api-client{__pip_version_string__}
+```
+
+or `pyproject.toml` for Poetry with:
+
+```text
+ds-caselaw-marklogic-api-client{__poetry_version_string__}
+```
+
+# Usage
+
+## Initialising the client
+
+Import the `MarklogicApiClient` class and instantiate with appropriate credentials:
+
+```python
+
+from caselawclient.Client import MarklogicApiClient
+
+client = MarklogicApiClient(
+    host="hostname",
+    username="username",
+    password="password",
+    use_https=True,
+)
+
+```
+
+"""

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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)

ds_caselaw_marklogic_api_client-39.2.0/src/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())

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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(

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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)

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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"

{ds_caselaw_marklogic_api_client-39.0.0 → ds_caselaw_marklogic_api_client-39.2.0}/src/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 → ds_caselaw_marklogic_api_client-39.2.0}/src/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/src/caselawclient/__init__.py

@@ -1,40 +0,0 @@
-"""
-
-# Installation
-
-Include the API client in your project using Pip or Poetry:
-
-```bash
-pip install ds-caselaw-marklogic-api-client
-```
-
-```bash
-poetry add ds-caselaw-marklogic-api-client
-```
-
-or in your projects `requirements.txt` with:
-
-```text
-ds-caselaw-marklogic-api-client~=12.0.0
-```
-
-# Usage
-
-## Initialising the client
-
-Import the `MarklogicApiClient` class and instantiate with appropriate credentials:
-
-```python
-
-from caselawclient.Client import MarklogicApiClient
-
-client = MarklogicApiClient(
-    host="hostname",
-    username="username",
-    password="password",
-    use_https=True,
-)
-
-```
-
-"""