implement job runner and scheduler

tests/test_web.py

@@ -5,7 +5,15 @@ from pathlib import Path
 from typing import Any, cast
 
 from repub.datastar import RefreshBroker, render_sse_event, render_stream
-from repub.model import Job, Source, SourceFeed, SourcePangea, create_source
+from repub.model import (
+    Job,
+    JobExecution,
+    JobExecutionStatus,
+    Source,
+    SourceFeed,
+    SourcePangea,
+    create_source,
+)
 from repub.web import (
     create_app,
     get_refresh_broker,
@@ -141,15 +149,20 @@ def test_render_stream_yields_on_connect_and_refresh() -> None:
     asyncio.run(run())
 
 
-def test_render_dashboard_shows_dashboard_information_architecture() -> None:
+def test_render_dashboard_shows_dashboard_information_architecture(
+    monkeypatch, tmp_path: Path
+) -> None:
+    db_path = tmp_path / "dashboard-render.db"
+    monkeypatch.setenv("REPUBLISHER_DB_PATH", str(db_path))
+
     async def run() -> None:
-        body = str(await render_dashboard())
+        app = create_app()
+        body = str(await render_dashboard(app))
 
         assert "Operational snapshot" in body
         assert "Running executions" in body
         assert 'href="/sources"' in body
         assert 'href="/runs"' in body
-        assert "/job/7/execution/104/logs" in body
         assert "Create source" in body
 
     asyncio.run(run())
@@ -569,27 +582,97 @@ def test_create_source_action_validates_duplicate_slug_and_pangea_type(
     asyncio.run(run())
 
 
-def test_render_runs_shows_running_upcoming_and_completed_tables() -> None:
+def test_render_runs_shows_running_upcoming_and_completed_tables(
+    monkeypatch, tmp_path: Path
+) -> None:
+    db_path = tmp_path / "runs-render.db"
+    monkeypatch.setenv("REPUBLISHER_DB_PATH", str(db_path))
+
     async def run() -> None:
-        body = str(await render_runs())
+        app = create_app()
+
+        source = create_source(
+            name="Runs render source",
+            slug="runs-render-source",
+            source_type="feed",
+            notes="",
+            spider_arguments="",
+            enabled=True,
+            cron_minute="*/30",
+            cron_hour="*",
+            cron_day_of_month="*",
+            cron_day_of_week="*",
+            cron_month="*",
+            feed_url="https://example.com/runs.xml",
+        )
+        job = Job.get(Job.source == source)
+        execution = JobExecution.create(
+            job=job,
+            running_status=JobExecutionStatus.SUCCEEDED,
+        )
+
+        body = str(await render_runs(app))
 
         assert "Running job executions" in body
         assert "Upcoming jobs" in body
         assert "Completed job executions" in body
-        assert "Delete confirmation" in body
-        assert "/job/11/execution/101/logs" in body
-        assert "Already running" in body
+        assert "runs-render-source" in body
+        assert f"/job/{job.id}/execution/{execution.get_id()}/logs" in body
+        assert "Already running" not in body
 
     asyncio.run(run())
 
 
-def test_render_execution_logs_uses_app_route() -> None:
-    async def run() -> None:
-        body = str(await render_execution_logs(job_id=7, execution_id=104))
+def test_render_execution_logs_uses_app_route(monkeypatch, tmp_path: Path) -> None:
+    db_path = tmp_path / "logs-render.db"
+    monkeypatch.setenv("REPUBLISHER_DB_PATH", str(db_path))
 
-        assert "Job 7 / execution 104" in body
-        assert "/job/7/execution/104/logs" in body
-        assert "Streaming text log view" in body
+    async def run() -> None:
+        log_dir = tmp_path / "out" / "logs"
+        app = create_app()
+        app.config["REPUB_LOG_DIR"] = log_dir
+
+        source = create_source(
+            name="Log render source",
+            slug="log-render-source",
+            source_type="feed",
+            notes="",
+            spider_arguments="",
+            enabled=False,
+            cron_minute="*/30",
+            cron_hour="*",
+            cron_day_of_month="*",
+            cron_day_of_week="*",
+            cron_month="*",
+            feed_url="https://example.com/logs.xml",
+        )
+        job = Job.get(Job.source == source)
+        execution = JobExecution.create(
+            job=job,
+            running_status=JobExecutionStatus.RUNNING,
+        )
+        log_path = log_dir / f"job-{job.id}-execution-{execution.get_id()}.log"
+        log_path.parent.mkdir(parents=True, exist_ok=True)
+        log_path.write_text(
+            "\n".join(
+                (
+                    "scheduler: run_now requested",
+                    "worker: starting simulated crawl",
+                    "worker: waiting for more log lines ...",
+                )
+            ),
+            encoding="utf-8",
+        )
+
+        body = str(
+            await render_execution_logs(
+                app, job_id=job.id, execution_id=int(execution.get_id())
+            )
+        )
+
+        assert f"Job {job.id} / execution {execution.get_id()}" in body
+        assert f"/job/{job.id}/execution/{execution.get_id()}/logs" in body
+        assert "Route: /job/" in body
         assert "waiting for more log lines" in body
 
     asyncio.run(run())