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
# ingester_agent.py
"""
Ingester Agent - Receives and validates incoming events.

Responsibilities:
- Accept events from external sources
- Validate schema and format
- Normalize data structure
- Rate limiting
- Send to processor
"""

import asyncio
from typing import Dict, Any, List, Optional
from datetime import datetime
from pydantic import BaseModel, ValidationError

from agentweave import SecureAgent, capability
from agentweave.types import TaskResult, Message, DataPart
from agentweave.exceptions import AgentCallError


class Event(BaseModel):
    """Event schema."""
    event_id: str
    event_type: str
    timestamp: datetime
    user_id: str
    properties: Dict[str, Any]
    source: str


class IngesterAgent(SecureAgent):
    """Ingests and validates events from external sources."""

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

        self.processor_id = "spiffe://agentweave.io/agent/processor"

        # Metrics
        self._events_received = 0
        self._events_validated = 0
        self._events_rejected = 0

    @capability("ingest_event")
    async def ingest_event(self, event_data: Dict[str, Any]) -> TaskResult:
        """
        Ingest a single event.

        Process:
        1. Validate event schema
        2. Normalize format
        3. Check for duplicates
        4. Send to processor
        """
        self._events_received += 1

        try:
            # Validate event schema
            event = Event(**event_data)

            # Check for duplicates (simplified)
            if await self._is_duplicate(event.event_id):
                self.logger.warning(
                    "Duplicate event detected",
                    extra={"event_id": event.event_id}
                )
                return TaskResult(
                    status="completed",
                    messages=[Message(
                        role="assistant",
                        parts=[DataPart(data={
                            "status": "duplicate",
                            "event_id": event.event_id
                        })]
                    )]
                )

            # Normalize event
            normalized = await self._normalize_event(event)

            # Send to processor
            result = await self.call_agent(
                target=self.processor_id,
                task_type="process_event",
                payload={
                    "event": normalized,
                    "ingested_at": datetime.utcnow().isoformat(),
                    "ingester_id": str(self.spiffe_id)
                },
                timeout=30.0
            )

            if result.status == "completed":
                self._events_validated += 1
                self.logger.info(
                    "Event ingested successfully",
                    extra={
                        "event_id": event.event_id,
                        "event_type": event.event_type
                    }
                )
            else:
                raise AgentCallError(f"Processor failed: {result.error}")

            return TaskResult(
                status="completed",
                messages=[Message(
                    role="assistant",
                    parts=[DataPart(data={
                        "status": "ingested",
                        "event_id": event.event_id,
                        "processed": True
                    })]
                )]
            )

        except ValidationError as e:
            self._events_rejected += 1
            self.logger.error(
                "Event validation failed",
                extra={"errors": str(e), "event_data": event_data}
            )

            # Send to dead letter queue
            await self._send_to_dlq(event_data, str(e))

            return TaskResult(
                status="failed",
                error=f"Invalid event schema: {e}"
            )

        except AgentCallError as e:
            self.logger.error(f"Failed to send to processor: {e}")
            # Could implement retry logic here
            return TaskResult(
                status="failed",
                error=f"Processing failed: {e}"
            )

    @capability("ingest_batch")
    async def ingest_batch(
        self,
        events: List[Dict[str, Any]]
    ) -> TaskResult:
        """
        Ingest a batch of events.

        Processes events in parallel with backpressure control.
        """
        self.logger.info(f"Ingesting batch of {len(events)} events")

        # Process events with concurrency limit
        semaphore = asyncio.Semaphore(10)  # Max 10 concurrent

        async def process_one(event_data):
            async with semaphore:
                return await self.ingest_event(event_data)

        results = await asyncio.gather(
            *[process_one(e) for e in events],
            return_exceptions=True
        )

        # Aggregate results
        succeeded = sum(1 for r in results if isinstance(r, TaskResult) and r.status == "completed")
        failed = len(results) - succeeded

        return TaskResult(
            status="completed",
            messages=[Message(
                role="assistant",
                parts=[DataPart(data={
                    "total": len(events),
                    "succeeded": succeeded,
                    "failed": failed
                })]
            )]
        )

    @capability("health")
    async def health(self) -> TaskResult:
        """Get ingester health metrics."""
        return TaskResult(
            status="completed",
            messages=[Message(
                role="assistant",
                parts=[DataPart(data={
                    "agent": "ingester",
                    "status": "healthy",
                    "metrics": {
                        "events_received": self._events_received,
                        "events_validated": self._events_validated,
                        "events_rejected": self._events_rejected,
                        "rejection_rate": self._events_rejected / max(self._events_received, 1)
                    }
                })]
            )]
        )

    async def _is_duplicate(self, event_id: str) -> bool:
        """Check if event was already processed."""
        # In production, check Redis or database
        return False

    async def _normalize_event(self, event: Event) -> Dict[str, Any]:
        """Normalize event format."""
        return {
            "event_id": event.event_id,
            "event_type": event.event_type,
            "timestamp": event.timestamp.isoformat(),
            "user_id": event.user_id,
            "properties": event.properties,
            "source": event.source,
            "version": "1.0"
        }

    async def _send_to_dlq(self, event_data: Dict[str, Any], error: str):
        """Send invalid event to dead letter queue."""
        self.logger.warning(
            "Sending to DLQ",
            extra={"event": event_data, "error": error}
        )
        # In production, send to SQS, Kafka, etc.


async def main():
    agent = IngesterAgent.from_config("config/ingester.yaml")
    await agent.run()


if __name__ == "__main__":
    asyncio.run(main())

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
# processor_agent.py
"""
Processor Agent - Enriches and transforms events.

Responsibilities:
- Enrich event data (lookup user info, etc.)
- Apply business rules
- Filter and transform
- Send to storage
"""

import asyncio
from typing import Dict, Any
from datetime import datetime

from agentweave import SecureAgent, capability, requires_peer
from agentweave.types import TaskResult, Message, DataPart
from agentweave.exceptions import AgentCallError


class ProcessorAgent(SecureAgent):
    """Processes and enriches events."""

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

        self.storage_id = "spiffe://agentweave.io/agent/storage"

        # Metrics
        self._events_processed = 0
        self._events_filtered = 0

    @capability("process_event")
    @requires_peer("spiffe://agentweave.io/agent/ingester")
    async def process_event(
        self,
        event: Dict[str, Any],
        ingested_at: str,
        ingester_id: str
    ) -> TaskResult:
        """
        Process an event.

        Processing pipeline:
        1. Enrich with additional data
        2. Apply business rules
        3. Transform format
        4. Send to storage
        """
        self.logger.info(
            "Processing event",
            extra={
                "event_id": event["event_id"],
                "event_type": event["event_type"]
            }
        )

        # Step 1: Enrich event
        enriched = await self._enrich_event(event)

        # Step 2: Apply business rules
        if not await self._should_process(enriched):
            self._events_filtered += 1
            self.logger.debug(
                "Event filtered by business rules",
                extra={"event_id": event["event_id"]}
            )
            return TaskResult(
                status="completed",
                messages=[Message(
                    role="assistant",
                    parts=[DataPart(data={"status": "filtered"})]
                )]
            )

        # Step 3: Transform
        transformed = await self._transform_event(enriched)

        # Step 4: Send to storage
        try:
            result = await self.call_agent(
                target=self.storage_id,
                task_type="store_event",
                payload={
                    "event": transformed,
                    "processed_at": datetime.utcnow().isoformat(),
                    "processor_id": str(self.spiffe_id)
                },
                timeout=30.0
            )

            if result.status != "completed":
                raise AgentCallError(f"Storage failed: {result.error}")

            self._events_processed += 1

            return TaskResult(
                status="completed",
                messages=[Message(
                    role="assistant",
                    parts=[DataPart(data={
                        "status": "processed",
                        "event_id": event["event_id"]
                    })]
                )]
            )

        except AgentCallError as e:
            self.logger.error(f"Failed to store event: {e}")
            return TaskResult(
                status="failed",
                error=f"Storage failed: {e}"
            )

    async def _enrich_event(self, event: Dict[str, Any]) -> Dict[str, Any]:
        """
        Enrich event with additional data.

        In production:
        - Lookup user profile
        - Geo-location from IP
        - Product catalog data
        """
        enriched = event.copy()

        # Simulate user lookup
        user_data = await self._lookup_user(event["user_id"])
        enriched["user"] = user_data

        # Add computed fields
        enriched["computed"] = {
            "day_of_week": datetime.fromisoformat(event["timestamp"]).strftime("%A"),
            "hour": datetime.fromisoformat(event["timestamp"]).hour
        }

        return enriched

    async def _should_process(self, event: Dict[str, Any]) -> bool:
        """
        Apply business rules to determine if event should be processed.

        Examples:
        - Filter test users
        - Skip certain event types
        - Apply sampling
        """
        # Skip test users
        if event.get("user", {}).get("is_test", False):
            return False

        # Skip certain event types
        skip_types = ["heartbeat", "debug"]
        if event["event_type"] in skip_types:
            return False

        return True

    async def _transform_event(self, event: Dict[str, Any]) -> Dict[str, Any]:
        """
        Transform event to storage format.

        - Flatten nested structures
        - Convert types
        - Add metadata
        """
        return {
            "id": event["event_id"],
            "type": event["event_type"],
            "timestamp": event["timestamp"],
            "user_id": event["user_id"],
            "user_name": event.get("user", {}).get("name"),
            "user_tier": event.get("user", {}).get("tier", "free"),
            "properties": event["properties"],
            "source": event["source"],
            "day_of_week": event["computed"]["day_of_week"],
            "hour": event["computed"]["hour"],
            "version": event["version"]
        }

    async def _lookup_user(self, user_id: str) -> Dict[str, Any]:
        """Lookup user data (mock)."""
        # In production, query user service or database
        return {
            "id": user_id,
            "name": f"User {user_id}",
            "tier": "premium",
            "is_test": False
        }


async def main():
    agent = ProcessorAgent.from_config("config/processor.yaml")
    await agent.run()


if __name__ == "__main__":
    asyncio.run(main())

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
# storage_agent.py
"""
Storage Agent - Persists events to data stores.

Responsibilities:
- Save to database
- Send to data warehouse
- Update search indexes
- Trigger downstream analytics
"""

import asyncio
from typing import Dict, Any
from datetime import datetime

from agentweave import SecureAgent, capability, requires_peer
from agentweave.types import TaskResult, Message, DataPart


class StorageAgent(SecureAgent):
    """Stores events in multiple data stores."""

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

        # Metrics
        self._events_stored = 0
        self._storage_errors = 0

    @capability("store_event")
    @requires_peer("spiffe://agentweave.io/agent/processor")
    async def store_event(
        self,
        event: Dict[str, Any],
        processed_at: str,
        processor_id: str
    ) -> TaskResult:
        """
        Store event in multiple destinations.

        Destinations:
        1. Primary database (PostgreSQL)
        2. Data warehouse (S3/Parquet)
        3. Search index (Elasticsearch)
        """
        self.logger.info(
            "Storing event",
            extra={"event_id": event["id"], "event_type": event["type"]}
        )

        results = {}

        try:
            # Store in parallel
            db_task = self._store_in_database(event)
            warehouse_task = self._store_in_warehouse(event)
            search_task = self._update_search_index(event)

            db_result, warehouse_result, search_result = await asyncio.gather(
                db_task, warehouse_task, search_task,
                return_exceptions=True
            )

            results["database"] = "success" if not isinstance(db_result, Exception) else str(db_result)
            results["warehouse"] = "success" if not isinstance(warehouse_result, Exception) else str(warehouse_result)
            results["search"] = "success" if not isinstance(search_result, Exception) else str(search_result)

            # Check if critical storage (database) succeeded
            if isinstance(db_result, Exception):
                self._storage_errors += 1
                raise db_result

            self._events_stored += 1

            # Trigger downstream (async, don't wait)
            asyncio.create_task(self._trigger_analytics(event))

            return TaskResult(
                status="completed",
                messages=[Message(
                    role="assistant",
                    parts=[DataPart(data={
                        "status": "stored",
                        "event_id": event["id"],
                        "destinations": results
                    })]
                )]
            )

        except Exception as e:
            self.logger.error(
                f"Storage failed: {e}",
                extra={"event_id": event["id"]}
            )
            return TaskResult(
                status="failed",
                error=f"Storage failed: {e}"
            )

    async def _store_in_database(self, event: Dict[str, Any]):
        """Store in primary database."""
        self.logger.debug(f"Storing event {event['id']} in database")

        # In production, use async database client
        # await db.events.insert_one(event)

        # Simulate database write
        await asyncio.sleep(0.01)

    async def _store_in_warehouse(self, event: Dict[str, Any]):
        """Store in data warehouse."""
        self.logger.debug(f"Storing event {event['id']} in warehouse")

        # In production:
        # - Buffer events
        # - Write to S3 as Parquet
        # - Partition by date
        # await s3_client.write_parquet(event)

        await asyncio.sleep(0.005)

    async def _update_search_index(self, event: Dict[str, Any]):
        """Update search index."""
        self.logger.debug(f"Indexing event {event['id']} in search")

        # In production, use Elasticsearch client
        # await es_client.index(index="events", document=event)

        await asyncio.sleep(0.005)

    async def _trigger_analytics(self, event: Dict[str, Any]):
        """Trigger downstream analytics (fire and forget)."""
        self.logger.debug(f"Triggering analytics for event {event['id']}")

        # In production:
        # - Send to Kafka topic
        # - Trigger Lambda function
        # - Update real-time dashboards

        await asyncio.sleep(0.01)


async def main():
    agent = StorageAgent.from_config("config/storage.yaml")
    await agent.run()


if __name__ == "__main__":
    asyncio.run(main())

1
2
3
4
5
6
# Ingester uses semaphore for concurrency control
semaphore = asyncio.Semaphore(10)

async def process_one(event):
    async with semaphore:
        return await self.ingest_event(event)