-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtest_openinference_cpp.cpp
More file actions
365 lines (307 loc) · 13 KB
/
test_openinference_cpp.cpp
File metadata and controls
365 lines (307 loc) · 13 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
// Issue #9 — parity coverage for the C++ OpenInference layer.
//
// Mirrors the assertions in `bindings/python/tests/test_openinference.py`:
// - openinference_tracer opens a CHAIN root + per-node CHAIN children
// - OpenInferenceProvider opens an LLM-kind child span carrying the
// full LLM attribute set (model, params, in/out messages, usage)
// - Streaming overloads append per-token events on the LLM span
// - Provider exceptions surface ERROR status on the span and re-raise
//
// The Tracer adapter under test is an in-memory recorder (no
// opentelemetry-cpp dependency).
#include <gtest/gtest.h>
#include <neograph/observability/openinference.h>
#include <neograph/observability/tracer.h>
#include <neograph/provider.h>
#include <neograph/graph/types.h>
#include <neograph/json.h>
#include <atomic>
#include <memory>
#include <mutex>
#include <stdexcept>
#include <string>
#include <unordered_map>
#include <vector>
using neograph::ChatCompletion;
using neograph::ChatMessage;
using neograph::CompletionParams;
using neograph::Provider;
using neograph::StreamCallback;
using neograph::graph::GraphEvent;
namespace obs = neograph::observability;
// ---------------------------------------------------------------------------
// In-memory tracer adapter.
// ---------------------------------------------------------------------------
namespace {
struct RecordedEvent {
std::string name;
std::string payload;
};
struct RecordedSpan {
std::string name;
obs::Span* parent = nullptr;
std::unordered_map<std::string, std::string> attrs_str;
std::unordered_map<std::string, int64_t> attrs_int;
std::unordered_map<std::string, bool> attrs_bool;
std::vector<RecordedEvent> events;
std::string status = "unset"; // "unset" | "ok" | "error"
std::string status_message;
bool ended = false;
};
class InMemorySpan : public obs::Span {
public:
explicit InMemorySpan(RecordedSpan* rec) : rec_(rec) {}
void set_attribute(std::string_view k, std::string_view v) override {
rec_->attrs_str[std::string(k)] = std::string(v);
}
void set_attribute(std::string_view k, int64_t v) override {
rec_->attrs_int[std::string(k)] = v;
}
void set_attribute(std::string_view k, double v) override {
rec_->attrs_str[std::string(k)] = std::to_string(v);
}
void set_attribute_bool(std::string_view k, bool v) override {
rec_->attrs_bool[std::string(k)] = v;
}
void add_event(std::string_view name, std::string_view payload) override {
rec_->events.push_back({std::string(name), std::string(payload)});
}
void set_status_ok() override { rec_->status = "ok"; }
void set_status_error(std::string_view msg) override {
rec_->status = "error";
rec_->status_message = std::string(msg);
}
void end() override { rec_->ended = true; }
private:
RecordedSpan* rec_;
};
class InMemoryTracer : public obs::Tracer {
public:
std::unique_ptr<obs::Span> start_span(std::string_view name,
obs::Span* parent) override {
std::lock_guard<std::mutex> lock(mu_);
spans_.push_back(std::make_unique<RecordedSpan>());
spans_.back()->name = std::string(name);
spans_.back()->parent = parent;
// Map the underlying record to its handle so the wrapper Span
// returned to the caller stays in sync with the recorded one.
return std::make_unique<InMemorySpan>(spans_.back().get());
}
std::vector<RecordedSpan*> snapshot() const {
std::lock_guard<std::mutex> lock(mu_);
std::vector<RecordedSpan*> out;
out.reserve(spans_.size());
for (const auto& s : spans_) out.push_back(s.get());
return out;
}
private:
mutable std::mutex mu_;
std::vector<std::unique_ptr<RecordedSpan>> spans_;
};
// ---------------------------------------------------------------------------
// Fake provider — records calls, emits configurable response.
// ---------------------------------------------------------------------------
class FakeProvider : public Provider {
public:
std::string reply = "ok";
int prompt_tokens = 7;
int completion_tokens = 3;
std::vector<std::string> stream_chunks;
std::atomic<int> calls{0};
ChatCompletion complete(const CompletionParams& /*p*/) override {
++calls;
return make_completion();
}
ChatCompletion complete_stream(const CompletionParams& /*p*/,
const StreamCallback& on_chunk) override {
++calls;
std::string acc;
for (const auto& c : stream_chunks) {
acc += c;
if (on_chunk) on_chunk(c);
}
if (acc.empty()) acc = reply;
auto comp = make_completion();
comp.message.content = acc;
return comp;
}
std::string get_name() const override { return "fake"; }
private:
ChatCompletion make_completion() const {
ChatCompletion c;
c.message.role = "assistant";
c.message.content = reply;
c.usage.prompt_tokens = prompt_tokens;
c.usage.completion_tokens = completion_tokens;
c.usage.total_tokens = prompt_tokens + completion_tokens;
return c;
}
};
class ThrowingProvider : public Provider {
public:
ChatCompletion complete(const CompletionParams&) override {
throw std::runtime_error("boom");
}
ChatCompletion complete_stream(const CompletionParams&,
const StreamCallback&) override {
return {};
}
std::string get_name() const override { return "throw"; }
};
} // namespace
// ---------------------------------------------------------------------------
// openinference_tracer — root + per-node spans
// ---------------------------------------------------------------------------
TEST(OpenInferenceCpp, TracerOpensChainRootAndPerNodeChild) {
InMemoryTracer tracer;
auto session = obs::openinference_tracer(tracer);
GraphEvent start{GraphEvent::Type::NODE_START, "researcher", neograph::json::object()};
GraphEvent end{GraphEvent::Type::NODE_END, "researcher",
neograph::json{{"summary", "found 3 hits"}}};
session.cb(start);
session.cb(end);
session.close();
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 2u);
// Root: graph.run, CHAIN, ended
EXPECT_EQ(spans[0]->name, "graph.run");
EXPECT_EQ(spans[0]->attrs_str["openinference.span.kind"], "CHAIN");
EXPECT_TRUE(spans[0]->ended);
// Node child: name "node.researcher", parent = root, CHAIN
EXPECT_EQ(spans[1]->name, "node.researcher");
EXPECT_NE(spans[1]->parent, nullptr);
EXPECT_EQ(spans[1]->attrs_str["openinference.span.kind"], "CHAIN");
EXPECT_EQ(spans[1]->attrs_str["neograph.node"], "researcher");
EXPECT_EQ(spans[1]->attrs_str["input.mime_type"], "application/json");
EXPECT_EQ(spans[1]->attrs_str["output.mime_type"], "application/json");
EXPECT_EQ(spans[1]->status, "ok");
EXPECT_TRUE(spans[1]->ended);
// input.value JSON contains node name; output.value JSON contains the data.
auto in_json = neograph::json::parse(spans[1]->attrs_str["input.value"]);
EXPECT_EQ(in_json["node"].get<std::string>(), "researcher");
auto out_json = neograph::json::parse(spans[1]->attrs_str["output.value"]);
EXPECT_EQ(out_json["summary"].get<std::string>(), "found 3 hits");
}
TEST(OpenInferenceCpp, TracerSurfacesErrorAndInterrupt) {
InMemoryTracer tracer;
auto session = obs::openinference_tracer(tracer);
session.cb({GraphEvent::Type::NODE_START, "a", neograph::json::object()});
session.cb({GraphEvent::Type::ERROR, "a", neograph::json("kaboom")});
session.cb({GraphEvent::Type::NODE_START, "b", neograph::json::object()});
session.cb({GraphEvent::Type::INTERRUPT, "b", neograph::json::object()});
session.close();
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 3u); // root + a + b
// a: ERROR
EXPECT_EQ(spans[1]->name, "node.a");
EXPECT_EQ(spans[1]->status, "error");
EXPECT_EQ(spans[1]->status_message, "kaboom");
EXPECT_EQ(spans[1]->attrs_str["neograph.error"], "kaboom");
EXPECT_TRUE(spans[1]->ended);
// b: INTERRUPT
EXPECT_EQ(spans[2]->name, "node.b");
EXPECT_TRUE(spans[2]->attrs_bool["neograph.interrupted"]);
EXPECT_TRUE(spans[2]->ended);
}
TEST(OpenInferenceCpp, TracerRecordsLLMTokenAsSpanEvent) {
InMemoryTracer tracer;
auto session = obs::openinference_tracer(tracer);
session.cb({GraphEvent::Type::NODE_START, "chat", neograph::json::object()});
session.cb({GraphEvent::Type::LLM_TOKEN, "chat", neograph::json("Hello")});
session.cb({GraphEvent::Type::LLM_TOKEN, "chat", neograph::json(" world")});
session.cb({GraphEvent::Type::NODE_END, "chat", neograph::json::object()});
session.close();
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 2u);
const auto& node_span = *spans[1];
ASSERT_EQ(node_span.events.size(), 2u);
EXPECT_EQ(node_span.events[0].name, "llm.token");
EXPECT_EQ(node_span.events[0].payload, "Hello");
EXPECT_EQ(node_span.events[1].payload, " world");
}
TEST(OpenInferenceCpp, TracerCloseAlsoEndsStragglerNodeSpan) {
InMemoryTracer tracer;
auto session = obs::openinference_tracer(tracer);
session.cb({GraphEvent::Type::NODE_START, "a", neograph::json::object()});
// No NODE_END before close — close() must still end the open span.
session.close();
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 2u);
EXPECT_TRUE(spans[0]->ended); // root
EXPECT_TRUE(spans[1]->ended); // straggler node span
}
// ---------------------------------------------------------------------------
// OpenInferenceProvider — LLM span attributes
// ---------------------------------------------------------------------------
TEST(OpenInferenceCpp, ProviderEmitsLLMSpanWithFullAttributes) {
InMemoryTracer tracer;
auto inner = std::make_shared<FakeProvider>();
inner->reply = "hello world";
inner->prompt_tokens = 7;
inner->completion_tokens = 3;
obs::OpenInferenceProvider wrapped(inner, tracer);
CompletionParams params;
params.model = "fake-model-1";
params.temperature = 0.5f;
params.max_tokens = 100;
params.messages.push_back({"system", "be helpful"});
params.messages.push_back({"user", "hi"});
auto result = wrapped.complete(params);
EXPECT_EQ(result.message.content, "hello world");
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 1u);
const auto& s = *spans[0];
EXPECT_EQ(s.name, "llm.complete");
EXPECT_EQ(s.attrs_str.at("openinference.span.kind"), "LLM");
EXPECT_EQ(s.attrs_str.at("llm.model_name"), "fake-model-1");
EXPECT_EQ(s.attrs_str.at("llm.input_messages.0.message.role"), "system");
EXPECT_EQ(s.attrs_str.at("llm.input_messages.0.message.content"), "be helpful");
EXPECT_EQ(s.attrs_str.at("llm.input_messages.1.message.role"), "user");
EXPECT_EQ(s.attrs_str.at("llm.input_messages.1.message.content"), "hi");
EXPECT_EQ(s.attrs_str.at("llm.output_messages.0.message.role"), "assistant");
EXPECT_EQ(s.attrs_str.at("llm.output_messages.0.message.content"), "hello world");
EXPECT_EQ(s.attrs_int.at("llm.token_count.prompt"), 7);
EXPECT_EQ(s.attrs_int.at("llm.token_count.completion"), 3);
EXPECT_EQ(s.attrs_int.at("llm.token_count.total"), 10);
EXPECT_EQ(s.attrs_str.at("input.mime_type"), "application/json");
EXPECT_EQ(s.attrs_str.at("output.mime_type"), "text/plain");
EXPECT_EQ(s.status, "ok");
EXPECT_TRUE(s.ended);
}
TEST(OpenInferenceCpp, ProviderStreamAppendsPerTokenEvents) {
InMemoryTracer tracer;
auto inner = std::make_shared<FakeProvider>();
inner->stream_chunks = {"foo", "bar", "baz"};
obs::OpenInferenceProvider wrapped(inner, tracer);
CompletionParams params;
params.model = "fake-stream";
params.messages.push_back({"user", "hi"});
std::vector<std::string> user_received;
auto result = wrapped.complete_stream(
params, [&](const std::string& c) { user_received.push_back(c); });
EXPECT_EQ(result.message.content, "foobarbaz");
EXPECT_EQ(user_received, (std::vector<std::string>{"foo", "bar", "baz"}));
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 1u);
const auto& s = *spans[0];
EXPECT_EQ(s.events.size(), 3u);
EXPECT_EQ(s.events[0].name, "llm.token");
EXPECT_EQ(s.events[0].payload, "foo");
EXPECT_EQ(s.events[2].payload, "baz");
EXPECT_EQ(s.attrs_str.at("output.value"), "foobarbaz");
EXPECT_TRUE(s.ended);
}
TEST(OpenInferenceCpp, ProviderPropagatesExceptionAndMarksSpanError) {
InMemoryTracer tracer;
auto inner = std::make_shared<ThrowingProvider>();
obs::OpenInferenceProvider wrapped(inner, tracer);
CompletionParams params;
params.model = "x";
params.messages.push_back({"user", "hi"});
EXPECT_THROW(wrapped.complete(params), std::runtime_error);
auto spans = tracer.snapshot();
ASSERT_EQ(spans.size(), 1u);
EXPECT_EQ(spans[0]->status, "error");
EXPECT_NE(spans[0]->status_message.find("boom"), std::string::npos);
EXPECT_TRUE(spans[0]->ended);
}