Pi cannot be read as another linear JSONL stream. Its history is a tree with a durable leaf, orphan roots, branch summaries, and two compaction forms, so the active context is something the format states rather than something line order implies. The adapter keeps those semantics inside itself and projects the result into the existing canonical tables. Sessions are keyed by (normalized header cwd, header id) rather than by path, because Pi's --session-id lookup is project-local: two projects may reuse an id, while a move or an identical copy is still one session. Discovery covers both layouts Pi writes and fingerprints each file by mtime, ctime, size and inode, so a rewrite that preserves mtime is not read as unchanged. Abandoned branches are preserved rather than dropped. Visibility becomes three-state -- visible, inactive, hidden -- and helpers return only visible rows until includeInactive asks for the superseded path, labeling every row so a caller knows which it holds. Usage counts all three, because an abandoned call still spent tokens; message_count reports only the visible transcript. A committed MIT-licensed oracle transcribed from Pi 0.83.0 pins the context algorithms, and a fixed-seed differential runs 512 generated sessions against it on every test run. Schema changes are additive.
175 lines
6.6 KiB
JavaScript
175 lines
6.6 KiB
JavaScript
import { test } from 'node:test';
|
|
import assert from 'node:assert/strict';
|
|
import { assembleSessionDetail } from '../app/src/shared/session-detail-assembly.mjs';
|
|
|
|
test('session assembly preserves thinking and attaches tool result and subagent evidence', () => {
|
|
const messages = [
|
|
{
|
|
uuid: 'thinking-1', timestamp: '2026-06-10T10:00:00Z',
|
|
type: 'assistant', content_type: 'thinking', text: 'reasoning',
|
|
},
|
|
{
|
|
uuid: 'answer-1', timestamp: '2026-06-10T10:00:01Z',
|
|
type: 'assistant', content_type: 'text', text: 'answer',
|
|
},
|
|
{
|
|
uuid: 'tool-1', timestamp: '2026-06-10T10:00:02Z',
|
|
type: 'assistant', content_type: 'tool_use', text: '',
|
|
},
|
|
{
|
|
uuid: 'result-1', timestamp: '2026-06-10T10:00:03Z',
|
|
type: 'user', content_type: 'tool_result', text: '',
|
|
},
|
|
];
|
|
const assembled = assembleSessionDetail({
|
|
messages,
|
|
toolCalls: [{ id: 'call-1', message_uuid: 'tool-1', name: 'Agent', input_json: '{"description":"inspect"}' }],
|
|
toolResults: [{ tool_use_id: 'call-1', message_uuid: 'result-1', content: 'done', is_error: 0 }],
|
|
subagents: [{ agent_id: 'agent-1', parent_tool_use_id: 'call-1', agent_type: 'reviewer', description: 'inspect' }],
|
|
workflows: [],
|
|
}).messages;
|
|
|
|
assert.equal(assembled.length, 1);
|
|
assert.equal(assembled[0].uuid, 'answer-1');
|
|
assert.equal(assembled[0]._thinking, 'reasoning');
|
|
assert.deepEqual(assembled[0].tool_calls[0].result.content, 'done');
|
|
assert.equal(assembled[0].tool_calls[0].subagent.agent_id, 'agent-1');
|
|
});
|
|
|
|
test('session assembly keeps Skill evidence standalone and embeds matching workflow agents', () => {
|
|
const assembled = assembleSessionDetail({
|
|
messages: [
|
|
{ uuid: 'skill-1', type: 'assistant', content_type: 'tool_use', text: '' },
|
|
{ uuid: 'skill-md', type: 'user', content_type: 'skill_instructions', is_meta: 1, text: '# Skill instructions' },
|
|
{ uuid: 'workflow-1', type: 'assistant', content_type: 'tool_use', text: '' },
|
|
],
|
|
toolCalls: [
|
|
{ id: 'call-skill', message_uuid: 'skill-1', name: 'Skill', presentation: 'skill', input_json: '{"skill":"obelisk"}' },
|
|
{ id: 'call-workflow', message_uuid: 'workflow-1', name: 'Workflow', presentation: 'default', input_json: '{}' },
|
|
],
|
|
toolResults: [{ tool_use_id: 'call-workflow', content: 'complete', is_error: 0 }],
|
|
subagents: [],
|
|
workflows: [{
|
|
run_id: 'run-1',
|
|
parent_tool_use_id: 'call-workflow',
|
|
workflow_name: 'review',
|
|
status: 'complete',
|
|
agents: [{ agent_id: 'agent-1', phase: 'review', label: 'Reviewer', state: 'complete' }],
|
|
}],
|
|
}).messages;
|
|
|
|
assert.equal(assembled[0]._skillMd, '# Skill instructions');
|
|
assert.equal(assembled[1].tool_calls[0].workflow.run_id, 'run-1');
|
|
assert.deepEqual(assembled[1].tool_calls[0].workflow.agents, [{
|
|
agent_id: 'agent-1',
|
|
phase: 'review',
|
|
label: 'Reviewer',
|
|
state: 'complete',
|
|
tokens: null,
|
|
duration_ms: null,
|
|
}]);
|
|
});
|
|
|
|
test('session assembly trusts canonical classification instead of parsing provider text', () => {
|
|
const detail = assembleSessionDetail({
|
|
messages: [{
|
|
uuid: 'provider-owned-classification',
|
|
type: 'user',
|
|
content_type: 'text',
|
|
is_meta: 0,
|
|
text: '<system-reminder>text alone does not define presentation semantics</system-reminder>',
|
|
}],
|
|
});
|
|
|
|
assert.equal(detail.messages[0].is_meta, 0);
|
|
});
|
|
|
|
test('canonical ordering is stable across provider and SQLite iteration order', () => {
|
|
const detail = assembleSessionDetail([
|
|
{
|
|
kind: 'message', uuid: 'b', session_id: 'session', type: 'user', parent_uuid: null,
|
|
timestamp: '2026-06-10T10:00:00Z', role: 'user', text: 'second', content_type: 'text',
|
|
is_meta: 0, visibility: 'visible', model: null, is_sidechain: 0, agent_id: null,
|
|
input_tokens: null, output_tokens: null, cwd: null, skill: null, source: 'test',
|
|
},
|
|
{
|
|
kind: 'message', uuid: 'a', session_id: 'session', type: 'user', parent_uuid: null,
|
|
timestamp: '2026-06-10T10:00:00Z', role: 'user', text: 'first', content_type: 'text',
|
|
is_meta: 0, visibility: 'visible', model: null, is_sidechain: 0, agent_id: null,
|
|
input_tokens: null, output_tokens: null, cwd: null, skill: null, source: 'test',
|
|
},
|
|
]);
|
|
|
|
assert.deepEqual(detail.messages.map(message => message.uuid), ['a', 'b']);
|
|
});
|
|
|
|
test('session detail remains active-only across direct and persisted visibility values', () => {
|
|
const message = (uuid, visibility) => ({
|
|
kind: 'message',
|
|
uuid,
|
|
session_id: 'session',
|
|
type: 'user',
|
|
parent_uuid: null,
|
|
timestamp: `2026-06-10T10:00:0${uuid.length}Z`,
|
|
role: 'user',
|
|
text: uuid,
|
|
content_type: 'text',
|
|
is_meta: 0,
|
|
visibility,
|
|
model: null,
|
|
is_sidechain: 0,
|
|
agent_id: null,
|
|
input_tokens: null,
|
|
output_tokens: null,
|
|
cwd: null,
|
|
skill: null,
|
|
source: 'pi',
|
|
});
|
|
const summary = (id, visibility) => ({
|
|
kind: 'summary',
|
|
id,
|
|
session_id: 'session',
|
|
timestamp: null,
|
|
source: 'pi:branch_summary',
|
|
content: id,
|
|
visibility,
|
|
input_tokens: null,
|
|
output_tokens: null,
|
|
});
|
|
const direct = assembleSessionDetail([
|
|
message('visible', 'visible'),
|
|
message('inactive', 'inactive'),
|
|
message('hidden', 'hidden'),
|
|
summary('visible-summary', 'visible'),
|
|
summary('inactive-summary', 'inactive'),
|
|
summary('hidden-summary', 'hidden'),
|
|
]);
|
|
assert.deepEqual(direct.messages.map(row => row.text), ['visible']);
|
|
assert.deepEqual(direct.summaries.map(row => row.content), ['visible-summary']);
|
|
|
|
const persisted = assembleSessionDetail({
|
|
messages: [
|
|
{ ...message('visible', 'visible'), kind: undefined },
|
|
{ ...message('inactive', 'inactive'), kind: undefined },
|
|
{ ...message('hidden', 'hidden'), kind: undefined },
|
|
{ ...message('unknown', 'future-state'), kind: undefined },
|
|
],
|
|
summaries: [
|
|
{ ...summary('visible-summary', 'visible'), kind: undefined },
|
|
{ ...summary('inactive-summary', 'inactive'), kind: undefined },
|
|
{ ...summary('hidden-summary', 'hidden'), kind: undefined },
|
|
{ ...summary('unknown-summary', 'future-state'), kind: undefined },
|
|
],
|
|
});
|
|
assert.deepEqual(persisted.messages.map(row => row.text), ['visible']);
|
|
assert.deepEqual(persisted.summaries.map(row => row.content), ['visible-summary']);
|
|
});
|
|
|
|
test('direct session assembly rejects an incomplete provider delta', () => {
|
|
assert.throws(() => assembleSessionDetail([{
|
|
kind: 'session', id: 'session', title: null, project: null,
|
|
started_at: null, ended_at: null, git_branch: null, version: null,
|
|
message_count: 1, countMode: 'delta', jsonl_path: '/session.jsonl', source: 'test',
|
|
}]), /fresh full parse/);
|
|
});
|