❗ [Includes historical git commits] Merge latest version of ministryofjustice-data-platform-catalogue

.gitignore

@@ -14,7 +14,6 @@ dist/
 downloads/
 eggs/
 .eggs/
-lib/
 lib64/
 parts/
 sdist/
@@ -25,6 +24,8 @@ share/python-wheels/
 .installed.cfg
 *.egg
 MANIFEST
+.vscode/**
+!.vscode/launch.json.default
 
 # PyInstaller
 #  Usually these files are written by a python script from a template

.vscode/launch.json → .vscode/launch.json.default

@@ -1,7 +1,4 @@
 {
-  // Use IntelliSense to learn about possible attributes.
-  // Hover to view descriptions of existing attributes.
-  // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
   "version": "0.2.0",
   "configurations": [
     {

Dockerfile

@@ -24,13 +24,16 @@ ENV POETRY_NO_INTERACTION=1 \
   POETRY_CACHE_DIR=/tmp/poetry_cache
 
 COPY pyproject.toml poetry.lock ./
+COPY lib ./lib
 
 RUN poetry install --without dev --no-root && rm -rf $POETRY_CACHE_DIR
 RUN poetry run python -m nltk.downloader punkt
 
 # The runtime image, used to just run the code provided its virtual environment
 FROM python:3.11-slim-buster as runtime
 
+WORKDIR /app
+
 ENV VIRTUAL_ENV=/app/.venv \
   PATH="/app/.venv/bin:$PATH"
 

home/service/base.py

@@ -1,11 +1,10 @@
-from data_platform_catalogue.client import BaseCatalogueClient
-from data_platform_catalogue.client.datahub import DataHubCatalogueClient
+from data_platform_catalogue.client.datahub_client import DataHubCatalogueClient
 from django.conf import settings
 
 
 class GenericService:
     @staticmethod
-    def _get_catalogue_client() -> BaseCatalogueClient:
+    def _get_catalogue_client() -> DataHubCatalogueClient:
         return DataHubCatalogueClient(
             jwt_token=settings.CATALOGUE_TOKEN, api_url=settings.CATALOGUE_URL
         )

home/service/details.py

@@ -5,54 +5,6 @@
 from .base import GenericService
 
 
-class DataProductDetailsService(GenericService):
-    def __init__(self, urn: str):
-        self.urn = urn
-        self.client = self._get_catalogue_client()
-
-        filter_value = [MultiSelectFilter("urn", [urn])]
-        search_results = self.client.search(query="", page=None, filters=filter_value)
-
-        if not search_results.page_results:
-            raise ObjectDoesNotExist(urn)
-
-        self.result = search_results.page_results[0]
-        self.assets_in_data_product = self._get_data_product_entities()
-        self.context = self._get_context()
-
-    def _get_data_product_entities(self):
-        # we might want to implement pagination for data product children
-        # details at some point
-        data_product_search = self.client.list_data_product_assets(
-            urn=self.urn, count=500
-        ).page_results
-
-        assets_in_data_product = []
-        for result in data_product_search:
-            assets_in_data_product.append(
-                {
-                    "name": result.name,
-                    "urn": result.id,
-                    "description": result.description,
-                    "type": "TABLE",
-                }
-            )
-
-        assets_in_data_product = sorted(assets_in_data_product, key=lambda d: d["name"])
-
-        return assets_in_data_product
-
-    def _get_context(self):
-        context = {
-            "result": self.result,
-            "result_type": "Data product",
-            "tables": self.assets_in_data_product,
-            "h1_value": "Details",
-        }
-
-        return context
-
-
 class DatabaseDetailsService(GenericService):
     def __init__(self, urn: str):
         self.urn = urn
@@ -69,7 +21,7 @@ def __init__(self, urn: str):
         self.context = self._get_context()
 
     def _get_database_entities(self):
-        # we might want to implement pagination for data product children
+        # we might want to implement pagination for database children
         # details at some point
         database_search = self.client.list_database_tables(
             urn=self.urn, count=500
@@ -80,7 +32,7 @@ def _get_database_entities(self):
             entities_in_database.append(
                 {
                     "name": result.name,
-                    "urn": result.id,
+                    "urn": result.urn,
                     "description": result.description,
                     "type": "TABLE",
                 }
@@ -122,11 +74,7 @@ def __init__(self, urn: str):
             # v0.12, assigning to multiple data products is not possible and we don't
             # have datasets with multiple parent containers.
             self.parent_entity = parents[0]
-            self.dataset_parent_type = (
-                ResultType.DATABASE.name.lower()
-                if "container" in self.parent_entity.id.split(":")
-                else ResultType.DATA_PRODUCT.name.lower()
-            )
+            self.dataset_parent_type = ResultType.DATABASE.name.lower()
         else:
             self.parent_entity = None
             self.dataset_parent_type = None

home/service/search.py

@@ -5,6 +5,7 @@
 from data_platform_catalogue.search_types import (
     MultiSelectFilter,
     ResultType,
+    SearchResponse,
     SortOption,
 )
 from django.core.paginator import Paginator
@@ -62,7 +63,7 @@ def _build_entity_types(_, entity_types: list[str]) -> tuple[ResultType]:
         )
         return chosen_entities if chosen_entities else default_entities
 
-    def _get_search_results(self, page: str, items_per_page: int):
+    def _get_search_results(self, page: str, items_per_page: int) -> SearchResponse:
         if self.form.is_bound:
             form_data = self.form.cleaned_data
         else:

home/urls.py

@@ -9,7 +9,7 @@
     path("search", views.search_view, name="search"),
     path("glossary", views.glossary_view, name="glossary"),
     path(
-        "details/<str:result_type>/<str:id>",
+        "details/<str:result_type>/<str:urn>",
         views.details_view,
         name="details",
     ),

home/views.py

@@ -1,12 +1,11 @@
-from django.core.exceptions import ObjectDoesNotExist
+from data_platform_catalogue.client.exceptions import EntityDoesNotExist
 from django.http import Http404, HttpResponseBadRequest
 from django.shortcuts import render
 
 from home.forms.search import SearchForm
 from home.service.details import (
     ChartDetailsService,
     DatabaseDetailsService,
-    DataProductDetailsService,
     DatasetDetailsService,
 )
 from home.service.glossary import GlossaryService
@@ -19,58 +18,44 @@ def home_view(request):
     return render(request, "home.html", context)
 
 
-def details_view(request, result_type, id):
-    if result_type == "data_product":
-        context = data_product_details(id)
-        return render(request, "details_data_product.html", context)
+def details_view(request, result_type, urn):
     if result_type == "table":
-        context = dataset_details(id)
+        context = dataset_details(urn)
         return render(request, "details_table.html", context)
     if result_type == "database":
-        context = database_details(id)
+        context = database_details(urn)
         return render(request, "details_database.html", context)
     if result_type == "chart":
-        context = chart_details(id)
+        context = chart_details(urn)
         return render(request, "details_chart.html", context)
 
 
-def data_product_details(id):
+def database_details(urn):
     try:
-        service = DataProductDetailsService(id)
-    except ObjectDoesNotExist:
+        service = DatabaseDetailsService(urn)
+    except EntityDoesNotExist:
         raise Http404("Asset does not exist")
 
     context = service.context
 
     return context
 
 
-def database_details(id):
+def dataset_details(urn):
     try:
-        service = DatabaseDetailsService(id)
-    except ObjectDoesNotExist:
+        service = DatasetDetailsService(urn)
+    except EntityDoesNotExist:
         raise Http404("Asset does not exist")
 
     context = service.context
 
     return context
 
 
-def dataset_details(id):
+def chart_details(urn):
     try:
-        service = DatasetDetailsService(id)
-    except ObjectDoesNotExist:
-        raise Http404("Asset does not exist")
-
-    context = service.context
-
-    return context
-
-
-def chart_details(id):
-    try:
-        service = ChartDetailsService(id)
-    except ObjectDoesNotExist:
+        service = ChartDetailsService(urn)
+    except EntityDoesNotExist:
         raise Http404("Asset does not exist")
 
     context = service.context

lib/datahub-client/.gitignore

@@ -0,0 +1,10 @@
+.env
+coverage/
+venv/
+env/
+.DS_STORE
+.vscode
+*.code-workspace
+dist/
+__pycache__
+.idea/