merge(app): land run Summary recorded fields
This commit is contained in:
@@ -0,0 +1,89 @@
|
||||
import '../../../core/api_gateway.dart';
|
||||
import '../../pipelines/pipeline_format.dart';
|
||||
import 'run_detail_content.dart';
|
||||
import 'run_detail_visit_facts.dart';
|
||||
|
||||
/// The exact wording for a Summary fact the reply did not record.
|
||||
///
|
||||
/// It is the same word the Summary's usage rows already use for an unreported
|
||||
/// value, so one absent fact never reads as `0`, `false` or an empty field.
|
||||
const String summaryUnknownValue = 'Unavailable';
|
||||
|
||||
/// The run Summary's recorded facts, in the web summary's order: the current
|
||||
/// stage and the run's timing, then the latest recorded visit's effective
|
||||
/// provider/model, target and verdict.
|
||||
///
|
||||
/// The latest visit is the one the Graph tab opens for the current stage, so
|
||||
/// the Summary and the recorded-visit pane describe the same evidence. Tokens
|
||||
/// and cost stay in [detailUsageMetrics], their existing Summary home. A fact
|
||||
/// the server did not record reads [summaryUnknownValue].
|
||||
List<DetailMetric> detailSummaryFields(JsonMap run) {
|
||||
final steps = detailRunSteps(run);
|
||||
final stage = _text(run['current_stage']);
|
||||
final latest = _latestVisit(steps, stage);
|
||||
return [
|
||||
(label: 'Stage', value: stage ?? summaryUnknownValue),
|
||||
(
|
||||
label: 'Started',
|
||||
value:
|
||||
formatLocalTimestamp(_text(run['created_at'])) ?? summaryUnknownValue,
|
||||
),
|
||||
(
|
||||
label: 'Finished',
|
||||
value:
|
||||
formatLocalTimestamp(_text(run['finished_at'])) ??
|
||||
summaryUnknownValue,
|
||||
),
|
||||
(label: 'Duration', value: _duration(run) ?? summaryUnknownValue),
|
||||
(
|
||||
label: 'Provider/model',
|
||||
value: latest == null
|
||||
? summaryUnknownValue
|
||||
: detailVisitProviderModel(latest) ?? summaryUnknownValue,
|
||||
),
|
||||
(
|
||||
label: 'Target',
|
||||
value: latest == null
|
||||
? summaryUnknownValue
|
||||
: _text(latest['target']) ?? summaryUnknownValue,
|
||||
),
|
||||
(label: 'Verdict', value: _verdict(latest) ?? summaryUnknownValue),
|
||||
];
|
||||
}
|
||||
|
||||
/// The latest recorded visit of [stage], or the latest visit when the reply
|
||||
/// records no current stage. It is the same visit the Graph tab opens for that
|
||||
/// stage, so a fact and its Stage row never come from different visits.
|
||||
JsonMap? _latestVisit(List<JsonMap> steps, String? stage) {
|
||||
JsonMap? latest;
|
||||
for (final step in steps) {
|
||||
if (stage == null || _text(step['stage']) == stage) latest = step;
|
||||
}
|
||||
return latest;
|
||||
}
|
||||
|
||||
/// The run's stored duration from its own start and finish, never a live
|
||||
/// estimate and never a zero for an unfinished run.
|
||||
String? _duration(JsonMap run) {
|
||||
final seconds = durationSeconds(
|
||||
_text(run['created_at']),
|
||||
_text(run['finished_at']),
|
||||
);
|
||||
return seconds == null ? null : formatDuration(seconds);
|
||||
}
|
||||
|
||||
/// The recorded verdict as the web pane spells it: the option above its
|
||||
/// reason, one part when only one was recorded, nothing when neither was.
|
||||
String? _verdict(JsonMap? step) {
|
||||
if (step == null) return null;
|
||||
final verdict = detailVisitVerdict(step);
|
||||
if (verdict == null) return null;
|
||||
final parts = [verdict.option, verdict.reason].whereType<String>();
|
||||
return parts.isEmpty ? null : parts.join('\n');
|
||||
}
|
||||
|
||||
String? _text(dynamic value) {
|
||||
if (value == null) return null;
|
||||
final text = '$value'.trim();
|
||||
return text.isEmpty ? null : text;
|
||||
}
|
||||
@@ -3,10 +3,11 @@ import 'package:flutter/material.dart';
|
||||
|
||||
import '../../../core/api_gateway.dart';
|
||||
import 'run_detail_content.dart';
|
||||
import 'run_detail_summary_fields.dart';
|
||||
|
||||
/// The Summary tab: the stored original prompt with its copy control and the
|
||||
/// reported usage metrics. Neither the prompt nor a metric is replaced by a
|
||||
/// fallback value.
|
||||
/// The Summary tab: the stored original prompt with its copy control, the
|
||||
/// run's recorded facts and the reported usage metrics. Neither the prompt nor
|
||||
/// a fact is replaced by a fallback value.
|
||||
Widget runDetailSummaryTab({
|
||||
required BuildContext context,
|
||||
required JsonMap run,
|
||||
@@ -16,6 +17,7 @@ Widget runDetailSummaryTab({
|
||||
required VoidCallback onCopy,
|
||||
}) {
|
||||
final prompt = detailPrompt(run);
|
||||
final facts = detailSummaryFields(run);
|
||||
final metrics = detailUsageMetrics(run['usage']);
|
||||
if (cupertino) {
|
||||
final theme = CupertinoTheme.of(context);
|
||||
@@ -55,6 +57,7 @@ Widget runDetailSummaryTab({
|
||||
),
|
||||
),
|
||||
_cupertinoMetrics(metrics),
|
||||
_cupertinoFacts(facts),
|
||||
],
|
||||
);
|
||||
}
|
||||
@@ -89,6 +92,7 @@ Widget runDetailSummaryTab({
|
||||
const SizedBox(height: 12),
|
||||
],
|
||||
_materialMetrics(metrics),
|
||||
_materialFacts(facts),
|
||||
],
|
||||
);
|
||||
}
|
||||
@@ -123,6 +127,24 @@ Widget runDetailFieldTab({
|
||||
);
|
||||
}
|
||||
|
||||
Widget _materialFacts(List<DetailMetric> facts) => Column(
|
||||
children: [
|
||||
for (final fact in facts)
|
||||
ListTile(title: Text(fact.label), subtitle: Text(fact.value)),
|
||||
],
|
||||
);
|
||||
|
||||
Widget _cupertinoFacts(List<DetailMetric> facts) =>
|
||||
CupertinoListSection.insetGrouped(
|
||||
children: [
|
||||
for (final fact in facts)
|
||||
CupertinoListTile(
|
||||
title: Text(fact.label),
|
||||
subtitle: Text(fact.value),
|
||||
),
|
||||
],
|
||||
);
|
||||
|
||||
Widget _materialMetrics(List<DetailMetric> metrics) => ExpansionTile(
|
||||
title: const Text('Metrics'),
|
||||
children: [
|
||||
|
||||
@@ -0,0 +1,180 @@
|
||||
import 'package:flutter/cupertino.dart';
|
||||
import 'package:flutter/material.dart';
|
||||
import 'package:flutter_test/flutter_test.dart';
|
||||
import 'package:pipelines_app/core/api_gateway.dart';
|
||||
import 'package:pipelines_app/features/pipelines/pipeline_format.dart';
|
||||
import 'package:pipelines_app/features/runs/ui/run_detail_summary_fields.dart';
|
||||
|
||||
import 'run_detail_graph_support.dart';
|
||||
|
||||
/// A run whose latest visit recorded every summary fact, plus reported usage.
|
||||
JsonMap recordedRunFixture() => <String, dynamic>{
|
||||
'id': 51,
|
||||
'title': 'Summary run',
|
||||
'name': 'release',
|
||||
'status': 'succeeded',
|
||||
'current_stage': 'review',
|
||||
'created_at': '2026-10-19T09:30:00Z',
|
||||
'finished_at': '2026-10-19T09:32:05Z',
|
||||
'input': <String, dynamic>{'prompt': 'ship it'},
|
||||
'steps': <dynamic>[
|
||||
<String, dynamic>{
|
||||
'stage': 'build',
|
||||
'sequence': 1,
|
||||
'status': 'completed',
|
||||
'outcome': 'success',
|
||||
'provider': 'openai',
|
||||
'model': 'gpt-luna',
|
||||
'target': 'review',
|
||||
},
|
||||
<String, dynamic>{
|
||||
'stage': 'review',
|
||||
'sequence': 2,
|
||||
'status': 'completed',
|
||||
'outcome': 'success',
|
||||
'provider': 'anthropic',
|
||||
'model': 'claude-x',
|
||||
'target': '#success',
|
||||
'verdict': <String, dynamic>{
|
||||
'option': 'approve',
|
||||
'reason': 'Looks good.',
|
||||
},
|
||||
},
|
||||
],
|
||||
'usage': <String, dynamic>{
|
||||
'stages': 1,
|
||||
'tokens_total': 123,
|
||||
'tokens_known_stages': 1,
|
||||
'tokens_in': 100,
|
||||
'tokens_out': 23,
|
||||
'split_known': true,
|
||||
'cost_known_stages': 1,
|
||||
'cost_usd': 0.5,
|
||||
},
|
||||
};
|
||||
|
||||
/// A run that recorded none of the summary facts and reported no usage.
|
||||
JsonMap absentRunFixture() => <String, dynamic>{
|
||||
'id': 51,
|
||||
'title': 'Sparse run',
|
||||
'name': 'release',
|
||||
'status': 'succeeded',
|
||||
'current_stage': ' ',
|
||||
'created_at': '',
|
||||
'finished_at': '',
|
||||
'input': <String, dynamic>{'prompt': 'ship it'},
|
||||
'steps': <dynamic>[
|
||||
<String, dynamic>{
|
||||
'stage': 'build',
|
||||
'sequence': 1,
|
||||
'status': 'completed',
|
||||
'provider': ' ',
|
||||
'model': '',
|
||||
'target': ' ',
|
||||
'verdict': null,
|
||||
},
|
||||
],
|
||||
};
|
||||
|
||||
const summaryFacts = [
|
||||
'Stage',
|
||||
'Started',
|
||||
'Finished',
|
||||
'Duration',
|
||||
'Provider/model',
|
||||
'Target',
|
||||
'Verdict',
|
||||
];
|
||||
|
||||
void main() {
|
||||
for (final platform in RunDetailTestPlatform.values) {
|
||||
testWidgets('${platform.name} summary renders each recorded field', (
|
||||
tester,
|
||||
) async {
|
||||
await pumpRunDetail(
|
||||
tester,
|
||||
platform,
|
||||
run: recordedRunFixture(),
|
||||
size: const Size(900, 1600),
|
||||
);
|
||||
|
||||
for (final label in summaryFacts) {
|
||||
expect(find.text(label), findsOneWidget, reason: label);
|
||||
}
|
||||
expect(find.text('review'), findsOneWidget);
|
||||
final started = formatLocalTimestamp('2026-10-19T09:30:00Z');
|
||||
final finished = formatLocalTimestamp('2026-10-19T09:32:05Z');
|
||||
expect(find.text(started!), findsOneWidget);
|
||||
expect(find.text(finished!), findsOneWidget);
|
||||
expect(find.text('2m 5s'), findsOneWidget);
|
||||
expect(find.text('anthropic / claude-x'), findsOneWidget);
|
||||
expect(find.text('#success'), findsOneWidget);
|
||||
expect(find.text('approve\nLooks good.'), findsOneWidget);
|
||||
|
||||
// Tokens and cost keep their existing Summary home and exact values.
|
||||
await tester.ensureVisible(find.text('Metrics'));
|
||||
await tester.tap(find.text('Metrics'));
|
||||
await tester.pumpAndSettle();
|
||||
expect(find.text('123'), findsOneWidget);
|
||||
expect(find.text('100'), findsOneWidget);
|
||||
expect(find.text('23'), findsOneWidget);
|
||||
expect(find.text('\$0.500000'), findsOneWidget);
|
||||
});
|
||||
|
||||
testWidgets('${platform.name} summary states absent facts as unknown', (
|
||||
tester,
|
||||
) async {
|
||||
await pumpRunDetail(
|
||||
tester,
|
||||
platform,
|
||||
run: absentRunFixture(),
|
||||
size: const Size(900, 1600),
|
||||
);
|
||||
|
||||
for (final label in summaryFacts) {
|
||||
expect(find.text(label), findsOneWidget, reason: label);
|
||||
}
|
||||
expect(
|
||||
find.text(summaryUnknownValue),
|
||||
findsNWidgets(summaryFacts.length),
|
||||
);
|
||||
expect(find.text('0'), findsNothing);
|
||||
expect(find.text('false'), findsNothing);
|
||||
|
||||
await tester.ensureVisible(find.text('Metrics'));
|
||||
await tester.tap(find.text('Metrics'));
|
||||
await tester.pumpAndSettle();
|
||||
expect(find.text('No reported usage.'), findsOneWidget);
|
||||
expect(find.text('0'), findsNothing);
|
||||
});
|
||||
}
|
||||
|
||||
test('visit facts come from the current stage latest visit', () {
|
||||
final run = recordedRunFixture()..['current_stage'] = 'build';
|
||||
final fields = detailSummaryFields(run);
|
||||
String value(String label) =>
|
||||
fields.firstWhere((field) => field.label == label).value;
|
||||
expect(value('Provider/model'), 'openai / gpt-luna');
|
||||
expect(value('Target'), 'review');
|
||||
expect(value('Verdict'), 'Unavailable');
|
||||
});
|
||||
|
||||
test('a current stage without a visit records no visit facts', () {
|
||||
final run = recordedRunFixture()..['current_stage'] = 'audit';
|
||||
final fields = detailSummaryFields(run);
|
||||
for (final label in ['Provider/model', 'Target', 'Verdict']) {
|
||||
expect(
|
||||
fields.firstWhere((field) => field.label == label).value,
|
||||
'Unavailable',
|
||||
reason: label,
|
||||
);
|
||||
}
|
||||
});
|
||||
|
||||
test('unknown wording is the web usage word and never zero', () {
|
||||
expect(summaryUnknownValue, 'Unavailable');
|
||||
final fields = detailSummaryFields(absentRunFixture());
|
||||
expect(fields.map((field) => field.label), summaryFacts);
|
||||
expect(fields.every((field) => field.value == 'Unavailable'), isTrue);
|
||||
});
|
||||
}
|
||||
Reference in New Issue
Block a user