File size: 9,830 Bytes
0d3f7cc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""Third-party observability integrations: LangSmith, W&B, Arize."""

from __future__ import annotations

import logging
from typing import Any

from hermes.config.settings import get_settings

logger = logging.getLogger(__name__)


class LangSmithIntegration:
    """LangSmith tracing integration for LLM observability."""

    def __init__(self) -> None:
        self.settings = get_settings()
        self._client: Any = None
        self._enabled = False

    async def initialize(self) -> None:
        """Initialize LangSmith client if configured."""
        if not self.settings.observability.langsmith_api_key:
            logger.info("LangSmith not configured (no API key)")
            return

        try:
            from langsmith import Client

            self._client = Client(
                api_key=self.settings.observability.langsmith_api_key,
                api_url=self.settings.observability.langsmith_endpoint,
            )
            self._enabled = True
            logger.info("LangSmith initialized")
        except ImportError:
            logger.warning("langsmith package not installed. Run: pip install langsmith")
        except Exception as e:
            logger.warning(f"Failed to initialize LangSmith: {e}")

    async def create_run(
        self,
        name: str,
        run_type: str = "chain",
        inputs: dict[str, Any] | None = None,
        project_name: str | None = None,
    ) -> str | None:
        """Create a trace run."""
        if not self._enabled or not self._client:
            return None

        try:
            project = project_name or self.settings.observability.langsmith_project
            run = self._client.create_run(
                name=name,
                run_type=run_type,
                inputs=inputs or {},
                project_name=project,
            )
            return str(run.id) if hasattr(run, "id") else None
        except Exception as e:
            logger.debug(f"LangSmith create_run failed: {e}")
            return None

    async def update_run(
        self,
        run_id: str,
        outputs: dict[str, Any] | None = None,
        error: str | None = None,
        end_time: float | None = None,
    ) -> None:
        """Update a trace run with outputs or error."""
        if not self._enabled or not self._client:
            return

        try:
            self._client.update_run(
                run_id=run_id,
                outputs=outputs,
                error=error,
                end_time=end_time,
            )
        except Exception as e:
            logger.debug(f"LangSmith update_run failed: {e}")

    async def trace_llm_call(
        self,
        model: str,
        prompt: str,
        response: str,
        prompt_tokens: int = 0,
        completion_tokens: int = 0,
        duration_ms: float = 0.0,
        tags: list[str] | None = None,
    ) -> None:
        """Trace an LLM call to LangSmith."""
        if not self._enabled:
            return

        run_id = await self.create_run(
            name=f"llm_call_{model}",
            run_type="llm",
            inputs={"prompt": prompt, "model": model},
        )
        if run_id:
            await self.update_run(
                run_id=run_id,
                outputs={
                    "response": response,
                    "token_usage": {
                        "prompt_tokens": prompt_tokens,
                        "completion_tokens": completion_tokens,
                    },
                    "duration_ms": duration_ms,
                },
            )

    @property
    def enabled(self) -> bool:
        """Check if LangSmith is enabled."""
        return self._enabled


class WeightsAndBiasesIntegration:
    """Weights & Biases integration for experiment tracking."""

    def __init__(self) -> None:
        self.settings = get_settings()
        self._run: Any = None
        self._enabled = False

    async def initialize(self, project: str = "hermes-platform") -> None:
        """Initialize W&B run."""
        try:
            import wandb

            self._run = wandb.init(
                project=project,
                config={
                    "observability_enabled": self.settings.observability.enabled,
                    "model_provider": self.settings.model.provider,
                    "model_name": self.settings.model.name,
                },
                reinit=True,
            )
            self._enabled = True
            logger.info("W&B initialized")
        except ImportError:
            logger.warning("wandb package not installed. Run: pip install wandb")
        except Exception as e:
            logger.warning(f"Failed to initialize W&B: {e}")

    async def log_metrics(self, metrics: dict[str, float], step: int | None = None) -> None:
        """Log metrics to W&B."""
        if not self._enabled or not self._run:
            return
        try:
            self._run.log(metrics, step=step)
        except Exception as e:
            logger.debug(f"W&B log_metrics failed: {e}")

    async def log_llm_call(
        self,
        model: str,
        prompt_tokens: int,
        completion_tokens: int,
        duration_ms: float,
        success: bool = True,
    ) -> None:
        """Log LLM call metrics to W&B."""
        await self.log_metrics({
            f"llm/{model}/prompt_tokens": prompt_tokens,
            f"llm/{model}/completion_tokens": completion_tokens,
            f"llm/{model}/total_tokens": prompt_tokens + completion_tokens,
            f"llm/{model}/duration_ms": duration_ms,
            f"llm/{model}/success": 1.0 if success else 0.0,
        })

    async def finish(self) -> None:
        """Finish the W&B run."""
        if self._enabled and self._run:
            try:
                self._run.finish()
            except Exception as e:
                logger.debug(f"W&B finish failed: {e}")

    @property
    def enabled(self) -> bool:
        """Check if W&B is enabled."""
        return self._enabled


class ArizeIntegration:
    """Arize AI integration for LLM observability."""

    def __init__(self) -> None:
        self.settings = get_settings()
        self._client: Any = None
        self._enabled = False

    async def initialize(self, api_key: str = "", space_key: str = "") -> None:
        """Initialize Arize client."""
        if not api_key and not self.settings.observability.langsmith_api_key:
            logger.info("Arize not configured")
            return

        try:
            from arize.api import Client

            self._client = Client(
                api_key=api_key or self.settings.observability.langsmith_api_key,
                space_key=space_key or "hermes-platform",
            )
            self._enabled = True
            logger.info("Arize initialized")
        except ImportError:
            logger.warning("arize package not installed. Run: pip install arize")
        except Exception as e:
            logger.warning(f"Failed to initialize Arize: {e}")

    async def log_llm_event(
        self,
        model: str,
        prompt: str,
        response: str,
        prompt_tokens: int = 0,
        completion_tokens: int = 0,
        latency_ms: float = 0.0,
        tags: dict[str, str] | None = None,
    ) -> None:
        """Log an LLM event to Arize."""
        if not self._enabled or not self._client:
            return
        try:
            self._client.log_llm_record(
                model=model,
                prompt=prompt,
                response=response,
                prompt_tokens=prompt_tokens,
                completion_tokens=completion_tokens,
                latency_ms=latency_ms,
                tags=tags or {},
            )
        except Exception as e:
            logger.debug(f"Arize log_llm_event failed: {e}")

    @property
    def enabled(self) -> bool:
        """Check if Arize is enabled."""
        return self._enabled


class ObservabilityManager:
    """Central manager for all observability integrations."""

    def __init__(self) -> None:
        self.langsmith = LangSmithIntegration()
        self.wandb = WeightsAndBiasesIntegration()
        self.arize = ArizeIntegration()

    async def initialize_all(self) -> None:
        """Initialize all configured integrations."""
        await self.langsmith.initialize()
        await self.wandb.initialize()
        await self.arize.initialize()

    async def trace_llm_call(
        self,
        model: str,
        prompt: str,
        response: str,
        prompt_tokens: int = 0,
        completion_tokens: int = 0,
        duration_ms: float = 0.0,
    ) -> None:
        """Trace an LLM call across all configured providers."""
        await self.langsmith.trace_llm_call(
            model=model,
            prompt=prompt,
            response=response,
            prompt_tokens=prompt_tokens,
            completion_tokens=completion_tokens,
            duration_ms=duration_ms,
        )
        await self.wandb.log_llm_call(
            model=model,
            prompt_tokens=prompt_tokens,
            completion_tokens=completion_tokens,
            duration_ms=duration_ms,
        )
        await self.arize.log_llm_event(
            model=model,
            prompt=prompt,
            response=response,
            prompt_tokens=prompt_tokens,
            completion_tokens=completion_tokens,
            latency_ms=duration_ms,
        )

    async def log_metrics(self, metrics: dict[str, float], step: int | None = None) -> None:
        """Log metrics across all configured providers."""
        await self.wandb.log_metrics(metrics, step=step)

    async def shutdown(self) -> None:
        """Shutdown all integrations gracefully."""
        await self.wandb.finish()


obs_manager = ObservabilityManager()