Repository navigation
Expand file tree
/
Copy pathtest_tools.py
More file actions
692 lines (572 loc) · 23.6 KB
/
Copy pathtest_tools.py
File metadata and controls
692 lines (572 loc) · 23.6 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
# Copyright (c) Microsoft Corporation. All rights reserved.
"""Unit tests for define_tool"""
import json
import pytest
from pydantic import BaseModel, ConfigDict, Field, RootModel, field_validator
from copilot import define_tool
from copilot.generated.rpc import ExternalToolTextResultForLlm
from copilot.tools import (
ToolBinaryResult,
ToolInvocation,
ToolResult,
_normalize_result,
convert_mcp_call_tool_result,
tool_result_to_external_tool_text_result_for_llm,
)
class TestDefineTool:
@pytest.mark.parametrize("patch", ["", "*** Begin Patch\n*** End Patch"])
async def test_string_schema_handler_preserves_scalar_arguments(self, patch: str):
calls = []
@define_tool("apply_patch", description="Apply a patch", overrides_built_in_tool=True)
def apply_patch(params: RootModel[str], invocation: ToolInvocation) -> str:
calls.append((params.root, invocation.arguments))
return "accepted"
result = await apply_patch.handler(
ToolInvocation(
session_id="session-1",
tool_call_id="call-1",
tool_name="apply_patch",
arguments=patch,
)
)
assert result.result_type == "success", result.error
assert calls == [(patch, patch)]
def test_creates_tool_with_correct_name_and_description(self):
class Params(BaseModel):
query: str
@define_tool("search", description="Search for something")
def search(params: Params, invocation: ToolInvocation) -> str:
return "result"
assert search.name == "search"
assert search.description == "Search for something"
assert search.handler is not None
assert search.parameters is not None
def test_infers_name_from_function(self):
class Params(BaseModel):
query: str
@define_tool(description="Search for something")
def my_search_tool(params: Params) -> str:
return "result"
assert my_search_tool.name == "my_search_tool"
def test_generates_schema_from_pydantic_model(self):
class Params(BaseModel):
city: str = Field(description="City name")
unit: str = Field(description="Temperature unit")
@define_tool("get_weather", description="Get weather")
def get_weather(params: Params, invocation: ToolInvocation) -> str:
return "sunny"
schema = get_weather.parameters
assert schema is not None
assert schema["type"] == "object"
assert "city" in schema["properties"]
assert "unit" in schema["properties"]
assert schema["properties"]["city"]["description"] == "City name"
async def test_handler_receives_typed_arguments(self):
class Params(BaseModel):
name: str
count: int
received_params = None
@define_tool("test", description="Test tool")
def test_tool(params: Params, invocation: ToolInvocation) -> str:
nonlocal received_params
received_params = params
return "ok"
invocation = ToolInvocation(
session_id="session-1",
tool_call_id="call-1",
tool_name="test",
arguments={"name": "Alice", "count": 42},
)
await test_tool.handler(invocation)
assert received_params is not None
assert received_params.name == "Alice"
assert received_params.count == 42
async def test_handler_receives_invocation(self):
class Params(BaseModel):
pass
received_inv = None
@define_tool("test", description="Test tool")
def test_tool(params: Params, invocation: ToolInvocation) -> str:
nonlocal received_inv
received_inv = invocation
return "ok"
invocation = ToolInvocation(
session_id="session-123",
tool_call_id="call-456",
tool_name="test",
arguments={},
)
await test_tool.handler(invocation)
assert received_inv.session_id == "session-123"
assert received_inv.tool_call_id == "call-456"
async def test_zero_param_handler(self):
"""Handler with no parameters: def handler() -> str"""
called = False
@define_tool("test", description="Test tool")
def test_tool() -> str:
nonlocal called
called = True
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="test",
arguments={},
)
result = await test_tool.handler(invocation)
assert called
assert result.text_result_for_llm == "ok"
async def test_invocation_only_handler(self):
"""Handler with only invocation: def handler(invocation) -> str"""
received_inv = None
@define_tool("test", description="Test tool")
def test_tool(invocation: ToolInvocation) -> str:
nonlocal received_inv
received_inv = invocation
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="test",
arguments={},
)
await test_tool.handler(invocation)
assert received_inv is not None
assert received_inv.session_id == "s1"
async def test_params_only_handler(self):
"""Handler with only params: def handler(params) -> str"""
class Params(BaseModel):
value: str
received_params = None
@define_tool("test", description="Test tool")
def test_tool(params: Params) -> str:
nonlocal received_params
received_params = params
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="test",
arguments={"value": "hello"},
)
await test_tool.handler(invocation)
assert received_params is not None
assert received_params.value == "hello"
async def test_handler_error_is_hidden_from_llm(self):
class Params(BaseModel):
pass
@define_tool("failing", description="A failing tool")
def failing_tool(params: Params, invocation: ToolInvocation) -> str:
raise ValueError("secret error message")
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="failing",
arguments={},
)
result = await failing_tool.handler(invocation)
assert result.result_type == "failure"
assert "secret error message" not in result.text_result_for_llm
assert "error" in result.text_result_for_llm.lower()
# But the actual error is stored internally
assert result.error == "secret error message"
async def test_validation_error_is_surfaced_to_llm(self):
class Params(BaseModel):
username: str
@field_validator("username")
@classmethod
def check_username(cls, v: str) -> str:
if v == "admin":
raise ValueError("username 'admin' is reserved")
return v
@define_tool("validate", description="A validating tool")
def validating_tool(params: Params) -> str:
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="validate",
arguments={"username": "admin"},
)
result = await validating_tool.handler(invocation)
assert result.result_type == "failure"
assert result.text_result_for_llm.startswith("Invalid tool arguments:")
assert "username 'admin' is reserved" in result.text_result_for_llm
# Full detail is retained in the debug field.
assert result.error is not None
async def test_validation_error_extra_forbid_includes_field_name(self):
class Params(BaseModel):
model_config = ConfigDict(extra="forbid")
request: str
@define_tool("strict", description="A strict tool")
def strict_tool(params: Params) -> str:
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="strict",
arguments={"request": "ok", "extra_field": "unexpected"},
)
result = await strict_tool.handler(invocation)
assert result.result_type == "failure"
assert result.text_result_for_llm.startswith("Invalid tool arguments:")
# The offending key name is carried in `loc` even though the generic
# message is "Extra inputs are not permitted".
assert "extra_field" in result.text_result_for_llm
assert result.error is not None
async def test_validation_error_from_handler_body_is_redacted(self):
class Params(BaseModel):
pass
class Internal(BaseModel):
count: int
@define_tool("body", description="A tool that validates internally")
def body_tool(params: Params) -> str:
Internal.model_validate({"count": "secret-not-an-int"})
return "ok"
invocation = ToolInvocation(
session_id="s1",
tool_call_id="c1",
tool_name="body",
arguments={},
)
result = await body_tool.handler(invocation)
assert result.result_type == "failure"
# A ValidationError from the handler body must not be surfaced as an
# argument-validation error; it stays redacted like any other exception.
assert not result.text_result_for_llm.startswith("Invalid tool arguments:")
assert "secret-not-an-int" not in result.text_result_for_llm
assert "error" in result.text_result_for_llm.lower()
assert result.error is not None
async def test_function_style_api(self):
class Params(BaseModel):
value: str
tool = define_tool(
"my_tool",
description="My tool",
handler=lambda params, inv: params.value.upper(),
params_type=Params,
)
assert tool.name == "my_tool"
assert tool.description == "My tool"
result = await tool.handler(
ToolInvocation(
session_id="s",
tool_call_id="c",
tool_name="my_tool",
arguments={"value": "hello"},
)
)
assert result.text_result_for_llm == "HELLO"
def test_function_style_can_create_declaration_only_tool(self):
class Params(BaseModel):
value: str = Field(description="Value to look up")
tool = define_tool(
"my_tool",
description="My tool",
params_type=Params,
)
assert tool.name == "my_tool"
assert tool.description == "My tool"
assert tool.handler is None
assert tool.parameters is not None
assert tool.parameters["properties"]["value"]["description"] == "Value to look up"
def test_function_style_requires_name(self):
class Params(BaseModel):
value: str
with pytest.raises(ValueError, match="name is required"):
define_tool(
description="My tool",
handler=lambda params, inv: params.value.upper(),
params_type=Params,
)
class TestNormalizeResult:
def test_none_returns_empty_success(self):
result = _normalize_result(None)
assert result.text_result_for_llm == ""
assert result.result_type == "success"
def test_string_passes_through(self):
result = _normalize_result("hello world")
assert result.text_result_for_llm == "hello world"
assert result.result_type == "success"
def test_tool_result_passes_through(self):
input_result = ToolResult(
text_result_for_llm="custom",
result_type="failure",
error="some error",
)
result = _normalize_result(input_result)
assert result.text_result_for_llm == "custom"
assert result.result_type == "failure"
def test_dict_is_json_serialized(self):
result = _normalize_result({"key": "value", "num": 42})
parsed = json.loads(result.text_result_for_llm)
assert parsed == {"key": "value", "num": 42}
assert result.result_type == "success"
def test_list_is_json_serialized(self):
result = _normalize_result(["a", "b", "c"])
assert result.text_result_for_llm == '["a", "b", "c"]'
assert result.result_type == "success"
def test_pydantic_model_is_serialized(self):
class Response(BaseModel):
status: str
count: int
result = _normalize_result(Response(status="ok", count=5))
parsed = json.loads(result.text_result_for_llm)
assert parsed == {"status": "ok", "count": 5}
def test_list_of_pydantic_models_is_serialized(self):
class Item(BaseModel):
name: str
value: int
items = [Item(name="a", value=1), Item(name="b", value=2)]
result = _normalize_result(items)
parsed = json.loads(result.text_result_for_llm)
assert parsed == [{"name": "a", "value": 1}, {"name": "b", "value": 2}]
assert result.result_type == "success"
def test_pydantic_model_with_non_primitive_fields_is_serialized(self):
from datetime import date, datetime
from decimal import Decimal
from enum import Enum
from uuid import UUID
class Status(Enum):
ACTIVE = "active"
class Record(BaseModel):
id: UUID
created: datetime
day: date
score: Decimal
status: Status
tags: set[str]
record = Record(
id=UUID("12345678-1234-5678-1234-567812345678"),
created=datetime(2026, 1, 15, 10, 30, 0),
day=date(2026, 1, 15),
score=Decimal("99.5"),
status=Status.ACTIVE,
tags={"python", "sdk"},
)
result = _normalize_result(record)
parsed = json.loads(result.text_result_for_llm)
assert parsed == {
"id": "12345678-1234-5678-1234-567812345678",
"created": "2026-01-15T10:30:00",
"day": "2026-01-15",
"score": "99.5",
"status": "active",
"tags": parsed["tags"],
}
assert set(parsed["tags"]) == {"python", "sdk"}
assert result.result_type == "success"
def test_plain_dict_with_non_primitive_fields_is_serialized(self):
from datetime import date, datetime, time
from decimal import Decimal
from enum import Enum
from uuid import UUID
class Status(Enum):
ACTIVE = "active"
result = _normalize_result(
{
"id": UUID("12345678-1234-5678-1234-567812345678"),
"created": datetime(2026, 1, 15, 10, 30, 0),
"day": date(2026, 1, 15),
"at": time(10, 30, 0),
"score": Decimal("99.5"),
"status": Status.ACTIVE,
"tags": {"python", "sdk"},
}
)
parsed = json.loads(result.text_result_for_llm)
assert parsed == {
"id": "12345678-1234-5678-1234-567812345678",
"created": "2026-01-15T10:30:00",
"day": "2026-01-15",
"at": "10:30:00",
"score": "99.5",
"status": "active",
"tags": parsed["tags"],
}
assert set(parsed["tags"]) == {"python", "sdk"}
assert result.result_type == "success"
def test_raises_for_unserializable_value(self):
# Functions cannot be JSON serialized
with pytest.raises(TypeError, match="Failed to serialize"):
_normalize_result(lambda x: x)
class TestConvertMcpCallToolResult:
def test_text_only_call_tool_result(self):
result = convert_mcp_call_tool_result(
{
"content": [{"type": "text", "text": "hello"}],
}
)
assert result.text_result_for_llm == "hello"
assert result.result_type == "success"
def test_multiple_text_blocks(self):
result = convert_mcp_call_tool_result(
{
"content": [
{"type": "text", "text": "line 1"},
{"type": "text", "text": "line 2"},
],
}
)
assert result.text_result_for_llm == "line 1\nline 2"
def test_is_error_maps_to_failure(self):
result = convert_mcp_call_tool_result(
{
"content": [{"type": "text", "text": "oops"}],
"isError": True,
}
)
assert result.result_type == "failure"
def test_is_error_false_maps_to_success(self):
result = convert_mcp_call_tool_result(
{
"content": [{"type": "text", "text": "ok"}],
"isError": False,
}
)
assert result.result_type == "success"
def test_image_content_to_binary(self):
result = convert_mcp_call_tool_result(
{
"content": [{"type": "image", "data": "base64data", "mimeType": "image/png"}],
}
)
assert result.binary_results_for_llm is not None
assert len(result.binary_results_for_llm) == 1
assert result.binary_results_for_llm[0].data == "base64data"
assert result.binary_results_for_llm[0].mime_type == "image/png"
assert result.binary_results_for_llm[0].type == "image"
def test_resource_text_to_text_result(self):
result = convert_mcp_call_tool_result(
{
"content": [
{
"type": "resource",
"resource": {"uri": "file:///data.txt", "text": "file contents"},
},
],
}
)
assert result.text_result_for_llm == "file contents"
def test_resource_blob_to_binary(self):
result = convert_mcp_call_tool_result(
{
"content": [
{
"type": "resource",
"resource": {
"uri": "file:///img.png",
"blob": "blobdata",
"mimeType": "image/png",
},
},
],
}
)
assert result.binary_results_for_llm is not None
assert len(result.binary_results_for_llm) == 1
assert result.binary_results_for_llm[0].data == "blobdata"
assert result.binary_results_for_llm[0].mime_type == "image/png"
assert result.binary_results_for_llm[0].description == "file:///img.png"
def test_resource_blob_defaults_missing_or_empty_mime_type(self):
result = convert_mcp_call_tool_result(
{
"content": [
{
"type": "resource",
"resource": {"uri": "file:///data.bin", "blob": "binarydata"},
},
{
"type": "resource",
"resource": {
"uri": "file:///empty-mime.bin",
"blob": "binarydata2",
"mimeType": "",
},
},
],
}
)
assert result.binary_results_for_llm is not None
assert len(result.binary_results_for_llm) == 2
assert result.binary_results_for_llm[0].mime_type == "application/octet-stream"
assert result.binary_results_for_llm[1].mime_type == "application/octet-stream"
def test_empty_content_array(self):
result = convert_mcp_call_tool_result({"content": []})
assert result.text_result_for_llm == ""
assert result.result_type == "success"
def test_call_tool_result_dict_is_json_serialized_by_normalize(self):
"""_normalize_result does NOT auto-detect MCP results; it JSON-serializes them."""
result = _normalize_result({"content": [{"type": "text", "text": "hello"}]})
parsed = json.loads(result.text_result_for_llm)
assert parsed == {"content": [{"type": "text", "text": "hello"}]}
class TestToolReferences:
def test_tool_references_pass_through_normalize(self):
input_result = ToolResult(
text_result_for_llm="found 2 tools",
result_type="success",
tool_references=["get_weather", "check_status"],
)
result = _normalize_result(input_result)
assert result.tool_references == ["get_weather", "check_status"]
def test_tool_references_serialized_to_wire(self):
wire = ExternalToolTextResultForLlm(
text_result_for_llm="found 2 tools",
result_type="success",
tool_references=["get_weather", "check_status"],
)
data = wire.to_dict()
assert data["toolReferences"] == ["get_weather", "check_status"]
def test_tool_references_omitted_when_none(self):
wire = ExternalToolTextResultForLlm(
text_result_for_llm="ok",
result_type="success",
)
assert "toolReferences" not in wire.to_dict()
def test_tool_references_round_trip_from_wire(self):
wire = ExternalToolTextResultForLlm.from_dict(
{
"textResultForLlm": "found tools",
"resultType": "success",
"toolReferences": ["alpha", "beta"],
}
)
assert wire.tool_references == ["alpha", "beta"]
class TestToolResultToExternalToolTextResultForLlm:
def test_forwards_binary_results_and_session_log(self):
tool_result = ToolResult(
text_result_for_llm="screenshot captured",
binary_results_for_llm=[
ToolBinaryResult(
data="base64data",
mime_type="image/png",
type="image",
description="screenshot.png",
)
],
session_log="tool execution details",
tool_telemetry={"duration_ms": 42},
)
rpc_result = tool_result_to_external_tool_text_result_for_llm(tool_result)
assert rpc_result.text_result_for_llm == "screenshot captured"
assert rpc_result.session_log == "tool execution details"
assert rpc_result.tool_telemetry == {"duration_ms": 42}
assert rpc_result.binary_results_for_llm is not None
assert len(rpc_result.binary_results_for_llm) == 1
assert rpc_result.binary_results_for_llm[0].data == "base64data"
assert rpc_result.binary_results_for_llm[0].mime_type == "image/png"
assert rpc_result.binary_results_for_llm[0].type.value == "image"
assert rpc_result.binary_results_for_llm[0].description == "screenshot.png"
def test_omits_binary_results_when_none(self):
tool_result = ToolResult(text_result_for_llm="done")
rpc_result = tool_result_to_external_tool_text_result_for_llm(tool_result)
assert rpc_result.binary_results_for_llm is None
assert rpc_result.session_log is None
def test_forwards_tool_references(self):
tool_result = ToolResult(
text_result_for_llm="found tools",
result_type="success",
tool_references=["get_weather", "check_status"],
)
rpc_result = tool_result_to_external_tool_text_result_for_llm(tool_result)
assert rpc_result.tool_references == ["get_weather", "check_status"]