Merge pull request #44 from OCHA-DAP/add_tests

HAPI-154 Add functional tests

.github/workflows/run-python-tests.yaml

@@ -0,0 +1,48 @@
+# This workflow will install Python dependencies, lint and run tests
+# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions
+
+name: Run tests
+
+on:
+  workflow_dispatch: # add run button in github
+  push:
+    branches-ignore:
+      - gh-pages
+      - 'dependabot/**'
+  pull_request:
+    branches-ignore:
+      - gh-pages
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.11'
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install --upgrade hatch
+    - name: Test with hatch/pytest
+      run: |
+        hatch run test:test
+    - name: Check styling
+      if: always()
+      run: |
+        hatch run lint:style
+    - name: Publish Unit Test Results
+      uses: EnricoMi/publish-unit-test-result-action@v2
+      if: always()
+      with:
+        github_token: ${{ secrets.GITHUB_TOKEN }}
+        junit_files: test-results.xml
+    - name: Publish in Coveralls
+      uses: coverallsapp/github-action@v2
+      with:
+        github-token: ${{ secrets.GITHUB_TOKEN }}
+        flag-name: tests
+        format: lcov

requirements.txt

@@ -59,7 +59,7 @@ filelock==3.12.4
     # via virtualenv
 frictionless==5.16.0
     # via hdx-python-utilities
-google-auth==2.23.2
+google-auth==2.23.3
     # via
     #   google-auth-oauthlib
     #   gspread
@@ -69,7 +69,7 @@ greenlet==3.0.0
     # via sqlalchemy
 gspread==5.11.3
     # via hdx-python-scraper
-hapi-schema==0.0.5
+hapi-schema==0.0.6
     # via hapi-pipelines (pyproject.toml)
 hdx-python-api==6.1.3
     # via

src/hapi/pipelines/app/__init__.py

@@ -0,0 +1,15 @@
+from os.path import join
+from typing import Dict
+
+from hdx.utilities.loader import load_and_merge_yaml
+from hdx.utilities.path import script_dir_plus_file
+from hdx.utilities.typehint import ListTuple
+
+
+def load_yamls(config_files: ListTuple[str]) -> Dict:
+    input_files = [
+        script_dir_plus_file(join("..", "configs", file), load_yamls)
+        for file in config_files
+    ]
+    project_config = load_and_merge_yaml(input_files)
+    return project_config

src/hapi/pipelines/app/__main__.py

@@ -5,7 +5,6 @@
 from os import getenv
 from typing import List, Optional
 
-import yaml
 from hdx.api.configuration import Configuration
 from hdx.database import Database
 from hdx.database.dburi import get_params_from_connection_uri
@@ -15,9 +14,10 @@
 from hdx.utilities.dictandlist import args_to_dict
 from hdx.utilities.easy_logging import setup_logging
 from hdx.utilities.errors_onexit import ErrorsOnExit
-from hdx.utilities.path import script_dir_plus_file, temp_dir
+from hdx.utilities.path import temp_dir
 
 from hapi.pipelines._version import __version__
+from hapi.pipelines.app import load_yamls
 from hapi.pipelines.app.pipelines import Pipelines
 
 setup_logging()
@@ -64,19 +64,6 @@ def parse_args():
     return parser.parse_args()
 
 
-def compile_YAMLs(config_files):
-    project_config = {}
-    for config_file in config_files:
-        file_address = script_dir_plus_file(f"../configs/{config_file}", main)
-        with open(file_address) as file:
-            config = yaml.safe_load(file)
-        project_config = project_config | config
-
-    output_address = script_dir_plus_file("project_configuration.yaml", main)
-    with open(output_address, "w") as file:
-        yaml.dump(project_config, file)
-
-
 def main(
     db_uri: Optional[str] = None,
     db_params: Optional[str] = None,
@@ -170,17 +157,14 @@ def main(
         "population.yaml",
         "operational_presence.yaml",
     ]
-    compile_YAMLs(project_configs)
-    project_config_yaml = script_dir_plus_file(
-        "project_configuration.yaml", main
-    )
+    project_config_dict = load_yamls(project_configs)
     facade(
         main,
         hdx_key=hdx_key,
         user_agent=user_agent,
         preprefix=preprefix,
         hdx_site=hdx_site,
-        project_config_yaml=project_config_yaml,
+        project_config_dict=project_config_dict,
         db_uri=db_uri,
         db_params=args.db_params,
         scrapers_to_run=scrapers_to_run,

src/hapi/pipelines/database/population.py

@@ -119,7 +119,7 @@ def _validate_hxl_tag(hxl_tag: str) -> bool:
         #population+f+age_5_12
         #population+f+age_80_plus
     """
-    # TODO: add tests for this (HAPI-154)
+    # TODO: add tests for this (HAPI-159)
     return bool(_HXL_PATTERN.match(hxl_tag))
 
 

tests/fixtures/input/afghanistan-who-does-what-where-january-to-march-2023.json

@@ -0,0 +1 @@
+{"archived": false, "creator_user_id": "391f0864-b6e4-425f-9d46-df87aa456c2b", "data_update_frequency": "180", "dataseries_name": "OCHA Afghanistan - Who does What Where", "dataset_date": "[2023-01-01T00:00:00 TO 2023-03-31T23:59:59]", "dataset_preview": "resource_id", "dataset_source": "Humanitarian partners", "due_date": "2023-11-05T07:35:47", "has_geodata": false, "has_quickcharts": true, "has_showcases": true, "id": "f50bc211-d42f-4fcd-8291-3e4a57a0dfcf", "is_requestdata_type": false, "isopen": true, "last_modified": "2023-05-09T07:35:47.682796", "license_id": "cc-by", "license_title": "Creative Commons Attribution International", "license_url": "http://www.opendefinition.org/licenses/cc-by", "maintainer": "391f0864-b6e4-425f-9d46-df87aa456c2b", "maintainer_email": "[email protected]", "metadata_created": "2023-05-04T06:04:43.041555", "metadata_modified": "2023-09-07T08:36:12.493166", "methodology": "Direct Observational Data/Anecdotal Data", "name": "afghanistan-who-does-what-where-january-to-march-2023", "notes": "The Who does What Where (3W) is a core humanitarian coordination dataset. It is critical to know where humanitarian organizations are working, what they are doing and their capability in order to identify gaps, avoid duplication of efforts, and plan for future humanitarian response (if needed). The data includes a list of humanitarian organizations by district and cluster, as well as a unique count of organizations. An interactive map of the 3W data can be accessed [here](https://www.humanitarianresponse.info/en/operations/afghanistan/3w). ", "num_resources": 3, "num_tags": 4, "organization": {"id": "10e168ce-5b51-49ac-8616-a142d48618e5", "name": "ocha-afghanistan", "title": "OCHA Afghanistan", "type": "organization", "description": "OCHA resumed its operation in Afghanistan in 2009, providing humanitarian assistance in a complex environment where separate \u2013 and not always complementary \u2013 military, political and security objectives pose challenges to the implementation of humanitarian principles, the ability of responders to reach people in need and the safety and security of aid workers", "image_url": "", "created": "2014-08-06T14:26:07.642577", "is_organization": true, "approval_status": "approved", "state": "active"}, "overdue_date": "2023-12-05T07:35:47", "owner_org": "10e168ce-5b51-49ac-8616-a142d48618e5", "package_creator": "murtaza", "pageviews_last_14_days": 1, "private": false, "qa_completed": false, "solr_additions": "{\"countries\": [\"Afghanistan\"]}", "state": "active", "subnational": "1", "title": "Afghanistan - Who does What Where (January to March 2023)", "total_res_downloads": 129, "type": "dataset", "url": null, "version": null, "groups": [{"description": "", "display_name": "Afghanistan", "id": "afg", "image_display_url": "", "name": "afg", "title": "Afghanistan"}], "tags": [{"display_name": "hxl", "id": "a0fbb23a-6aad-4ccc-8062-e9ef9f20e5d2", "name": "hxl", "state": "active", "vocabulary_id": "b891512e-9516-4bf5-962a-7a289772a2a1"}, {"display_name": "operational capacity", "id": "d8a59526-9f9a-4c71-b38f-5d9f2eb1615a", "name": "operational capacity", "state": "active", "vocabulary_id": "b891512e-9516-4bf5-962a-7a289772a2a1"}, {"display_name": "operational presence", "id": "a25059f9-7e1f-49be-b629-ccccd97a95f8", "name": "operational presence", "state": "active", "vocabulary_id": "b891512e-9516-4bf5-962a-7a289772a2a1"}, {"display_name": "who is doing what and where-3w-4w-5w", "id": "ec53893c-6dba-4656-978b-4a32289ea2eb", "name": "who is doing what and where-3w-4w-5w", "state": "active", "vocabulary_id": "b891512e-9516-4bf5-962a-7a289772a2a1"}], "relationships_as_subject": [], "relationships_as_object": [], "is_fresh": true, "update_status": "fresh", "x_resource_grouping": [], "resources": [{"alt_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/129a828a-9735-418e-ad41-691a6dac4144/download/", "cache_last_updated": null, "cache_url": null, "created": "2023-05-09T07:35:36.019511", "dataset_preview_enabled": false, "datastore_active": false, "description": "", "download_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/129a828a-9735-418e-ad41-691a6dac4144/download/afghanistan-3w-january-march-2023.xlsx", "format": "XLSX", "fs_check_info": "[{\"state\": \"processing\", \"message\": \"The processing of the file structure check has started\", \"timestamp\": \"2023-05-09T07:35:35.550193\"}, {\"state\": \"success\", \"message\": \"Hxl Proxy data received successfully\", \"timestamp\": \"2023-05-09T07:35:37.820231\", \"hxl_proxy_response\": {\"url_or_filename\": null, \"format\": \"XLSX\", \"sheets\": [{\"name\": \"MetaData\", \"is_hidden\": false, \"nrows\": 11, \"ncols\": 1, \"has_merged_cells\": false, \"is_hxlated\": false, \"header_hash\": \"8a9c0818e0622aadfcb2e59cfe5299b4\", \"hxl_header_hash\": null}, {\"name\": \"List_District\", \"is_hidden\": false, \"nrows\": 403, \"ncols\": 21, \"has_merged_cells\": false, \"is_hxlated\": true, \"header_hash\": \"30d419bf10484e12f5a01f983141414f\", \"hxl_header_hash\": \"13a225f52ee1aef7a638c73adb893a32\"}, {\"name\": \"Count_District\", \"is_hidden\": false, \"nrows\": 403, \"ncols\": 21, \"has_merged_cells\": false, \"is_hxlated\": true, \"header_hash\": \"30d419bf10484e12f5a01f983141414f\", \"hxl_header_hash\": \"13a225f52ee1aef7a638c73adb893a32\"}]}}]", "hash": "", "hdx_rel_url": "/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/129a828a-9735-418e-ad41-691a6dac4144/download/afghanistan-3w-january-march-2023.xlsx", "id": "129a828a-9735-418e-ad41-691a6dac4144", "last_modified": "2023-05-09T07:35:35.641553", "metadata_modified": "2023-05-09T14:11:46.390076", "microdata": false, "mimetype": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet", "mimetype_inner": null, "name": "afghanistan-3w-january-march-2023.xlsx", "originalHash": "-988266717", "package_id": "f50bc211-d42f-4fcd-8291-3e4a57a0dfcf", "pii": "false", "pii_predict_score": 0.8562966605953982, "pii_report_flag": "FINDINGS", "pii_report_id": "/resources/129a828a-9735-418e-ad41-691a6dac4144/pii.2023-05-09T07-35-40.main.json", "pii_timestamp": "2023-05-09T07:36:15.845000", "position": 0, "resource_type": "file.upload", "size": 142558, "state": "active", "url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/129a828a-9735-418e-ad41-691a6dac4144/download/afghanistan-3w-january-march-2023.xlsx", "url_type": "upload"}, {"alt_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/1d28eb8a-5585-4694-ba2f-766212a2f5ae/download/", "cache_last_updated": null, "cache_url": null, "created": "2023-05-04T06:04:44.805532", "dataset_preview_enabled": true, "datastore_active": false, "description": "", "download_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/1d28eb8a-5585-4694-ba2f-766212a2f5ae/download/afghanistan-3w-operational-presence-january-march-2023.csv", "format": "CSV", "hash": "", "hdx_rel_url": "/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/1d28eb8a-5585-4694-ba2f-766212a2f5ae/download/afghanistan-3w-operational-presence-january-march-2023.csv", "id": "1d28eb8a-5585-4694-ba2f-766212a2f5ae", "last_modified": "2023-05-09T07:35:47.682796", "metadata_modified": "2023-06-05T10:39:33.579461", "microdata": false, "mimetype": "text/csv", "mimetype_inner": null, "name": "afghanistan-3w-operational-presence-january-march-2023.csv", "originalHash": -1590159593, "package_id": "f50bc211-d42f-4fcd-8291-3e4a57a0dfcf", "pii": "false", "pii_predict_score": 0.7961682088135019, "pii_report_flag": "FINDINGS", "pii_report_id": "/resources/1d28eb8a-5585-4694-ba2f-766212a2f5ae/pii.2023-05-09T07-35-51.main.json", "pii_timestamp": "2023-05-09T07:36:40.941000", "position": 1, "resource_type": "file.upload", "size": 1477495, "state": "active", "url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/1d28eb8a-5585-4694-ba2f-766212a2f5ae/download/afghanistan-3w-operational-presence-january-march-2023.csv", "url_type": "upload"}, {"alt_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/f4b8a3d4-b952-4375-b9b8-926c78047c59/download/", "cache_last_updated": null, "cache_url": null, "created": "2023-05-04T06:04:54.297605", "dataset_preview_enabled": false, "datastore_active": false, "description": "", "download_url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/f4b8a3d4-b952-4375-b9b8-926c78047c59/download/afghanistan-3w-operational-capacity-january-march-2023.csv", "format": "CSV", "hash": "", "hdx_rel_url": "/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/f4b8a3d4-b952-4375-b9b8-926c78047c59/download/afghanistan-3w-operational-capacity-january-march-2023.csv", "id": "f4b8a3d4-b952-4375-b9b8-926c78047c59", "last_modified": "2023-05-04T06:04:54.071665", "metadata_modified": "2023-05-04T06:05:52.859718", "microdata": false, "mimetype": "text/csv", "mimetype_inner": null, "name": "afghanistan-3w-operational-capacity-january-march-2023.csv", "originalHash": "-988266717", "package_id": "f50bc211-d42f-4fcd-8291-3e4a57a0dfcf", "pii": "false", "pii_predict_score": 0.7829160250103557, "pii_report_flag": "FINDINGS", "pii_report_id": "/resources/f4b8a3d4-b952-4375-b9b8-926c78047c59/pii.2023-05-04T06-04-57.main.json", "pii_timestamp": "2023-05-04T06:05:52.383000", "position": 2, "resource_type": "file.upload", "size": 1302201, "state": "active", "url": "https://data.humdata.org/dataset/f50bc211-d42f-4fcd-8291-3e4a57a0dfcf/resource/f4b8a3d4-b952-4375-b9b8-926c78047c59/download/afghanistan-3w-operational-capacity-january-march-2023.csv", "url_type": "upload"}]}