solpbc.org
personal memory agent
1# SPDX-License-Identifier: AGPL-3.0-only
2# Copyright (c) 2026 sol pbc
3
4"""Tests for the sol talent CLI."""
5
6import json
7
8import pytest
9
10from think.talent_cli import (
11 _collect_configs,
12 _format_bytes,
13 _format_cost,
14 _format_tags,
15 _parse_run_stats,
16 _scan_variables,
17 json_output,
18 list_prompts,
19 log_run,
20 logs_runs,
21 show_prompt,
22)
23
24
25def test_collect_configs_returns_prompts():
26 """All configs include known system prompts."""
27 configs = _collect_configs(include_disabled=True)
28 assert "flow" in configs
29 assert "sense" in configs
30 assert "chat" in configs
31
32
33def test_collect_configs_excludes_disabled_by_default():
34 """Disabled prompts are excluded unless include_disabled is set."""
35 without = _collect_configs(include_disabled=False)
36 with_disabled = _collect_configs(include_disabled=True)
37 # include_disabled should return at least as many configs
38 assert len(with_disabled) >= len(without)
39 assert "flow" in without
40 assert "flow" in with_disabled
41
42
43def test_collect_configs_filter_schedule():
44 """Schedule filter returns only matching prompts."""
45 daily = _collect_configs(schedule="daily", include_disabled=True)
46 for key, info in daily.items():
47 assert info.get("schedule") == "daily", f"{key} should be daily"
48
49 segment = _collect_configs(schedule="segment", include_disabled=True)
50 for key, info in segment.items():
51 assert info.get("schedule") == "segment", f"{key} should be segment"
52
53 # No overlap
54 assert not set(daily.keys()) & set(segment.keys())
55
56 activity = _collect_configs(schedule="activity", include_disabled=True)
57 for key, info in activity.items():
58 assert info.get("schedule") == "activity", f"{key} should be activity"
59
60 # decisions is activity-scheduled
61 assert "decisions" in activity
62
63
64def test_collect_configs_filter_source():
65 """Source filter returns only matching prompts."""
66 system = _collect_configs(source="system", include_disabled=True)
67 for key, info in system.items():
68 assert info.get("source") == "system", f"{key} should be system"
69
70 app = _collect_configs(source="app", include_disabled=True)
71 for key, info in app.items():
72 assert info.get("source") == "app", f"{key} should be app"
73
74
75def test_format_tags_hook():
76 """Format tags shows compact output, hook, disabled, and FAIL tags."""
77 # Output format tags
78 assert _format_tags({"output": "md"}) == "md"
79 assert _format_tags({"output": "json"}) == "json"
80 assert _format_tags({}) == ""
81
82 # Hook tags (compact, no =name suffix)
83 assert _format_tags({"hook": {"post": "occurrence"}}) == "post"
84 assert _format_tags({"hook": {"pre": "prep"}}) == "pre"
85 assert _format_tags({"hook": {"pre": "prep", "post": "process"}}) == "pre post"
86
87 # Disabled
88 assert _format_tags({"disabled": True}) == "disabled"
89
90 # FAIL tag
91 assert _format_tags({}, failed=True) == "FAIL"
92 assert _format_tags({"output": "md"}, failed=True) == "md FAIL"
93
94 # Combined: output + hooks + disabled + FAIL
95 tags = _format_tags(
96 {"output": "md", "hook": {"post": "occurrence"}, "disabled": True},
97 failed=True,
98 )
99 assert tags == "md post disabled FAIL"
100
101
102def test_scan_variables():
103 """Variable scanning finds template variables in prompt body."""
104 assert "name" in _scan_variables("Hello $name, welcome")
105 assert "daily_preamble" in _scan_variables("$daily_preamble\n\n# Title")
106 assert _scan_variables("No variables here") == []
107 # Deduplicates
108 result = _scan_variables("$foo and $bar and $foo again")
109 assert result == ["foo", "bar"]
110
111
112def test_list_prompts_output(capsys):
113 """List view outputs expected groups and prompts with column layout."""
114 list_prompts()
115 output = capsys.readouterr().out
116
117 # Column header
118 assert "NAME" in output
119 assert "TITLE" in output
120 assert "LAST RUN" in output
121 assert "TAGS" in output
122 assert "OUTPUT" not in output
123
124 # Group headers
125 assert "segment:" in output
126 assert "daily:" in output
127 assert "activity:" in output
128
129 # Prompt names
130 assert "activity" in output
131 assert "flow" in output
132
133 # Last run column is present
134 assert "LAST RUN" in output
135
136
137def test_list_prompts_schedule_filter(capsys):
138 """Schedule filter shows only matching group."""
139 list_prompts(schedule="segment")
140 output = capsys.readouterr().out
141
142 assert "sense" in output
143 # Should not show daily-only prompts
144 # (but don't assert group headers since they're suppressed with filter)
145
146
147def test_list_prompts_disabled_shown(capsys):
148 """--disabled includes disabled prompts (currently none after cleanup)."""
149 list_prompts(include_disabled=True)
150 output = capsys.readouterr().out
151
152 # all agents should appear in the listing
153 assert "flow" in output
154
155
156def test_show_prompt_known(capsys):
157 """Detail view shows expected fields for a known prompt."""
158 show_prompt("flow")
159 output = capsys.readouterr().out
160
161 assert "talent/flow.md" in output
162 assert "title:" in output
163 assert "schedule:" in output
164 assert "daily" in output
165 assert "hook:" in output
166 assert "occurrence" in output
167 assert "variables:" in output
168 assert "$daily_preamble" in output
169 assert "body:" in output
170 assert "lines" in output
171
172
173def test_show_prompt_not_found(capsys):
174 """Detail view exits with error for unknown prompt."""
175 with pytest.raises(SystemExit):
176 show_prompt("nonexistent_prompt_xyz")
177
178 output = capsys.readouterr().err
179 assert "not found" in output.lower()
180
181
182def test_json_output_format(capsys):
183 """JSON output produces valid JSONL with file field."""
184 json_output()
185 output = capsys.readouterr().out
186
187 lines = [x for x in output.strip().splitlines() if x.strip()]
188 assert len(lines) > 0
189
190 for line in lines:
191 record = json.loads(line)
192 assert "file" in record, f"Missing 'file' key in: {line}"
193 assert record["file"].endswith(".md")
194
195
196def test_json_output_contains_known_prompts(capsys):
197 """JSON output includes known prompts with expected fields."""
198 json_output(include_disabled=True)
199 output = capsys.readouterr().out
200
201 records = [json.loads(x) for x in output.strip().splitlines() if x.strip()]
202 files = {r["file"] for r in records}
203 assert any("flow.md" in f for f in files)
204 assert any("sense.md" in f for f in files)
205
206 # Check a specific record has expected fields
207 flow = next(r for r in records if "flow.md" in r["file"])
208 assert "title" in flow
209 assert "schedule" in flow
210
211
212def test_json_output_schedule_filter(capsys):
213 """JSON output respects schedule filter."""
214 json_output(schedule="segment")
215 output = capsys.readouterr().out
216
217 records = [json.loads(x) for x in output.strip().splitlines() if x.strip()]
218 for r in records:
219 assert r.get("schedule") == "segment", f"Expected segment: {r}"
220
221
222def test_show_prompt_as_json(capsys):
223 """Detail view with --json outputs single JSONL record."""
224 show_prompt("flow", as_json=True)
225 output = capsys.readouterr().out
226
227 lines = [x for x in output.strip().splitlines() if x.strip()]
228 assert len(lines) == 1
229
230 record = json.loads(lines[0])
231 assert record["file"].endswith("flow.md")
232 assert "title" in record
233 assert "schedule" in record
234 # Should not contain expanded instruction text
235 assert "system_instruction" not in record
236
237
238def test_truncate_content():
239 """Content truncation works correctly."""
240 from think.talent_cli import _truncate_content
241
242 # Short content not truncated
243 short = "line1\nline2\nline3"
244 result, omitted = _truncate_content(short, max_lines=10)
245 assert result == short
246 assert omitted == 0
247
248 # Long content truncated
249 long = "\n".join(f"line{i}" for i in range(200))
250 result, omitted = _truncate_content(long, max_lines=100)
251 assert omitted == 100
252 assert "lines omitted" in result
253 assert "line0" in result # First lines kept
254 assert "line199" in result # Last lines kept
255
256
257def test_yesterday():
258 """Yesterday helper returns correct format."""
259 from think.talent_cli import _yesterday
260
261 result = _yesterday()
262 assert len(result) == 8
263 assert result.isdigit()
264
265
266def test_show_prompt_context_segment_validation(capsys):
267 """Segment-scheduled prompts require --segment."""
268 from think.talent_cli import show_prompt_context
269
270 with pytest.raises(SystemExit):
271 show_prompt_context("screen", day="20260101")
272
273 output = capsys.readouterr().err
274 assert "segment-scheduled" in output.lower()
275
276
277def test_show_prompt_context_multi_facet_validation(capsys):
278 """Multi-facet prompts require --facet."""
279 from think.talent_cli import show_prompt_context
280
281 with pytest.raises(SystemExit):
282 show_prompt_context("entities:entities")
283
284 output = capsys.readouterr().err
285 assert "multi-facet" in output.lower()
286
287
288def test_show_prompt_context_day_format_validation(capsys):
289 """Day argument must be YYYYMMDD format."""
290 from think.talent_cli import show_prompt_context
291
292 # Too short
293 with pytest.raises(SystemExit):
294 show_prompt_context("flow", day="2026")
295
296 output = capsys.readouterr().err
297 assert "invalid --day format" in output.lower()
298
299 # Non-numeric
300 with pytest.raises(SystemExit):
301 show_prompt_context("flow", day="abcdefgh")
302
303 output = capsys.readouterr().err
304 assert "invalid --day format" in output.lower()
305
306
307def test_logs_runs_default(capsys):
308 """Logs shows recent runs from fixture day-index files."""
309 logs_runs(count=50)
310 output = capsys.readouterr().out
311
312 # Should have runs from all fixture days (original + R&J)
313 assert "default" in output or "unified" in output
314 assert "flow" in output
315 assert "activity" in output
316 assert "entities" in output
317 assert "meetings" in output
318 assert "knowledge_graph" in output
319 # Error run should show ✗
320 assert "\u2717" in output
321 # Completed runs should show ✓
322 assert "\u2713" in output
323
324
325def test_logs_runs_filter_agent(capsys):
326 """Logs filters to a specific agent."""
327 logs_runs(agent="default")
328 output = capsys.readouterr().out
329
330 lines = [line for line in output.strip().splitlines() if line.strip()]
331 # fixture has 2 "default" runs in 20231114 + 2 from R&J (20260305, 20260310)
332 assert len(lines) == 4
333 for line in lines:
334 assert "default" in line
335 # Should NOT contain other agents
336 assert "flow" not in output
337 assert "activity" not in output
338
339
340def test_logs_runs_count_limit(capsys):
341 """Logs respects count limit."""
342 logs_runs(count=2)
343 output = capsys.readouterr().out
344
345 lines = [line for line in output.strip().splitlines() if line.strip()]
346 assert len(lines) == 2
347
348
349def test_logs_runs_no_results(capsys):
350 """Logs with unknown agent produces empty output."""
351 logs_runs(agent="nonexistent_agent_xyz")
352 output = capsys.readouterr().out
353 assert output.strip() == ""
354
355
356def test_logs_runs_new_columns(capsys):
357 """Logs output includes enriched columns for runs with JSONL files."""
358 logs_runs(count=50)
359 output = capsys.readouterr().out
360 lines = [line for line in output.strip().splitlines() if line.strip()]
361
362 # Find the line for agent_id 1700000000001 (has JSONL file)
363 enriched_line = None
364 for line in lines:
365 if "1700000000001" in line:
366 enriched_line = line
367 break
368 assert enriched_line is not None
369
370 # Should have numeric event/tool counts (not "-")
371 # The fixture has 7 events total, 6 non-request, 1 tool_start
372 assert " 6 " in enriched_line # events
373 assert " 1 " in enriched_line # tools
374
375 # Lines without JSONL files should show "-" for enriched columns
376 # (most lines lack JSONL files)
377 dash_count = sum(1 for line in lines if " - " in line)
378 assert dash_count > 0
379
380
381def test_logs_runs_day_filter(capsys):
382 """--day filters to a specific day."""
383 logs_runs(day="20231114")
384 output = capsys.readouterr().out
385 lines = [line for line in output.strip().splitlines() if line.strip()]
386 # 20231114 has 4 records
387 assert len(lines) == 4
388 # All should be from 20231114
389 for line in lines:
390 assert "1700000" in line # all agent_ids from that day start with 1700000
391
392
393def test_logs_runs_day_filter_no_match(capsys):
394 """--day with nonexistent day produces empty output."""
395 logs_runs(day="20990101")
396 output = capsys.readouterr().out
397 assert output.strip() == ""
398
399
400def test_logs_runs_day_invalid(capsys):
401 """--day with invalid format prints error."""
402 with pytest.raises(SystemExit):
403 logs_runs(day="bad")
404 output = capsys.readouterr().err
405 assert "invalid --day format" in output.lower()
406
407
408def test_logs_runs_errors_filter(capsys):
409 """--errors shows only error runs."""
410 logs_runs(errors=True)
411 output = capsys.readouterr().out
412 lines = [line for line in output.strip().splitlines() if line.strip()]
413 # Only flow on 20231114 has status "error"
414 assert len(lines) == 1
415 assert "flow" in lines[0]
416 assert "✗" in lines[0]
417
418
419def test_logs_runs_daily_filter(capsys):
420 """--daily shows only daily-scheduled runs."""
421 logs_runs(daily=True)
422 output = capsys.readouterr().out
423 lines = [line for line in output.strip().splitlines() if line.strip()]
424 # Daily runs: entities (20231113, schedule=daily), default x2 (20231114,
425 # schedule=daily + legacy fallback)
426 # Should NOT include flow (segment) or activity
427 assert "flow" not in output
428 assert "activity" not in output
429 for line in lines:
430 assert any(
431 name in line
432 for name in ["default", "entities", "meetings", "knowledge_graph"]
433 )
434
435
436def test_logs_runs_daily_bumps_count(capsys):
437 """--daily bumps default count to 50."""
438 # With only 6 total records in fixtures, verify explicit count still applies.
439 logs_runs(daily=True, count=1)
440 output = capsys.readouterr().out
441 lines = [line for line in output.strip().splitlines() if line.strip()]
442 assert len(lines) == 1
443
444
445def test_logs_runs_filter_composition(capsys):
446 """Filters compose with AND logic."""
447 logs_runs(day="20231114", errors=True)
448 output = capsys.readouterr().out
449 lines = [line for line in output.strip().splitlines() if line.strip()]
450 # Only flow on 20231114 is an error
451 assert len(lines) == 1
452 assert "flow" in lines[0]
453
454
455def test_logs_runs_summary(capsys):
456 """--summary shows grouped aggregation."""
457 logs_runs(summary=True, count=50)
458 output = capsys.readouterr().out
459 # Should have agent names (original + R&J)
460 assert "default" in output
461 assert "flow" in output
462 assert "entities" in output
463 assert "activity" in output
464 assert "meetings" in output
465 assert "knowledge_graph" in output
466 # Should have totals line
467 assert "total" in output
468 # Should show pass/fail symbols
469 assert "✓" in output
470 assert "✗" in output
471
472
473def test_logs_runs_daily_summary(capsys):
474 """--daily --summary shows only daily runs in summary."""
475 logs_runs(daily=True, summary=True)
476 output = capsys.readouterr().out
477 # Only daily agents (entities, default, meetings, knowledge_graph)
478 assert "flow" not in output
479 assert "activity" not in output
480 assert "default" in output
481 assert "entities" in output
482 assert "meetings" in output
483 assert "knowledge_graph" in output
484 assert "total" in output
485
486
487def test_parse_run_stats():
488 """Parse run stats extracts correct counts from fixture JSONL."""
489 from pathlib import Path
490
491 jsonl = Path("tests/fixtures/journal/agents/default/1700000000001.jsonl")
492 stats = _parse_run_stats(jsonl)
493 assert stats["event_count"] == 6 # all except request
494 assert stats["tool_count"] == 1 # one tool_start
495 assert stats["model"] == "gpt-4o"
496 assert stats["usage"] == {"input_tokens": 150, "output_tokens": 80}
497 assert stats["request"] is not None
498 assert stats["request"]["prompt"] == "Search for meetings about project updates"
499
500
501def test_parse_run_stats_error():
502 """Parse run stats handles error run JSONL correctly."""
503 from pathlib import Path
504
505 jsonl = Path("tests/fixtures/journal/agents/flow/1700000000002.jsonl")
506 stats = _parse_run_stats(jsonl)
507 assert stats["event_count"] == 2 # start + error (not request)
508 assert stats["tool_count"] == 0
509 assert stats["model"] == "claude-3-haiku"
510 assert stats["usage"] is None
511
512
513def test_format_bytes():
514 """Byte formatting produces human-readable strings."""
515 assert _format_bytes(0) == "0"
516 assert _format_bytes(500) == "500"
517 assert _format_bytes(999) == "999"
518 assert _format_bytes(1000) == "1.0K"
519 assert _format_bytes(1200) == "1.2K"
520 assert _format_bytes(34000) == "34.0K"
521 assert _format_bytes(1500000) == "1.5M"
522
523
524def test_format_cost():
525 """Cost formatting shows rounded cents."""
526 assert _format_cost(None) == "-"
527 assert _format_cost(0.0) == "0¢"
528 assert _format_cost(0.001) == "<1¢"
529 assert _format_cost(0.02) == "2¢"
530 assert _format_cost(0.10) == "10¢"
531 assert _format_cost(1.50) == "150¢"
532
533
534def test_log_run_default(capsys):
535 """Log run shows one-line-per-event output."""
536 log_run("1700000000001")
537 output = capsys.readouterr().out
538 lines = output.strip().splitlines()
539
540 # Fixture has 7 events
541 assert len(lines) == 7
542
543 # Each line should be ≤100 chars
544 for line in lines:
545 assert len(line) <= 100, f"Line too long ({len(line)}): {line}"
546
547 # Check event type labels appear
548 full_output = output
549 assert "request" in full_output
550 assert "start" in full_output
551 assert "think" in full_output
552 assert "tool" in full_output
553 assert "tool_end" in full_output
554 assert "updated" in full_output
555 assert "finish" in full_output
556
557
558def test_log_run_json(capsys):
559 """Log run --json outputs raw JSONL."""
560 log_run("1700000000001", json_mode=True)
561 output = capsys.readouterr().out
562 lines = [line for line in output.strip().splitlines() if line.strip()]
563
564 assert len(lines) == 7
565 # Each line should be valid JSON
566 for line in lines:
567 parsed = json.loads(line)
568 assert "event" in parsed
569
570
571def test_log_run_full(capsys):
572 """Log run --full shows expanded content with escaped newlines."""
573 log_run("1700000000001", full=True)
574 output = capsys.readouterr().out
575
576 # The thinking event in the fixture has actual newlines in "content"
577 # In --full mode, these should appear as literal \n
578 assert "\\n" in output
579
580 # Lines can exceed 100 chars in full mode
581 lines = output.strip().splitlines()
582 assert len(lines) == 7
583
584
585def test_log_run_missing():
586 """Log run with unknown ID exits with error."""
587 with pytest.raises(SystemExit):
588 log_run("nonexistent_id_12345")
589
590
591def test_log_run_error_run(capsys):
592 """Log run displays error events correctly."""
593 log_run("1700000000002")
594 output = capsys.readouterr().out
595 lines = output.strip().splitlines()
596 assert len(lines) == 3 # request, start, error
597 assert "error" in output
598 assert "Rate limit" in output
599
600
601def test_show_prompt_context_activity_requires_facet(capsys):
602 """Activity-scheduled prompts require --facet."""
603 from think.talent_cli import show_prompt_context
604
605 with pytest.raises(SystemExit):
606 show_prompt_context("decisions", day="20260214")
607
608 output = capsys.readouterr().err
609 assert "activity-scheduled" in output.lower()
610 assert "--facet" in output
611
612
613def test_show_prompt_context_activity_requires_activity_id(capsys):
614 """Activity-scheduled prompts require --activity and list available IDs."""
615 from think.talent_cli import show_prompt_context
616
617 with pytest.raises(SystemExit):
618 show_prompt_context("decisions", day="20260214", facet="full-featured")
619
620 output = capsys.readouterr().err
621 assert "--activity" in output
622 assert "coding_093000_300" in output
623 assert "meeting_140000_300" in output
624
625
626def test_show_prompt_context_activity_not_found(capsys):
627 """Activity-scheduled prompt with unknown activity ID errors."""
628 from think.talent_cli import show_prompt_context
629
630 with pytest.raises(SystemExit):
631 show_prompt_context(
632 "decisions",
633 day="20260214",
634 facet="full-featured",
635 activity="nonexistent_999",
636 )
637
638 output = capsys.readouterr().err
639 assert "not found" in output.lower()
640
641
642def test_list_prompts_activity_group(capsys):
643 """List view includes activity group with decisions agent."""
644 list_prompts()
645 output = capsys.readouterr().out
646
647 assert "activity:" in output
648 assert "decisions" in output