cursor-lens/orchestrator.py at main · EbookHun/cursor-lens · GitHub

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
from datetime import datetime

from agents.classifier import ClassifierAgent
from agents.policy import PolicyAgent
from agents.trend import TrendAgent
from agents.lens import LensAgent
from data.mock_posts import MOCK_POSTS
from time_windows import rolling_window_bounds


class Orchestrator:
    def __init__(self, blackboard):
        self.bb = blackboard
        self.classifier = ClassifierAgent(blackboard)
        self.policy = PolicyAgent(blackboard)
        self.trend = TrendAgent(blackboard)
        self.lens = LensAgent(blackboard)

    def process_post(self, post):
        self.bb.write("orchestrator", f"source_post_{post['post_id']}", post)
        classification = self.classifier.run(post)
        policy_result = self.policy.run(post["post_id"])
        return {"post": post, "classification": classification, "policy": policy_result}

    def run_batch(self, posts, source_context=None):
        timestamps = [datetime.fromisoformat(post["created_at"]) for post in posts]
        window_end = max(timestamps)
        window_start, window_end = rolling_window_bounds(window_end)
        window_posts = [
            post
            for post, timestamp in zip(posts, timestamps, strict=True)
            if window_start <= timestamp <= window_end
        ]
        results = []
        verified_posts = []

        for post in window_posts:
            result = self.process_post(post)
            results.append(result)
            if result["policy"].get("passes_filter"):
                verified_posts.append(
                    {
                        "post_id": post["post_id"],
                        "subreddit": post["subreddit"],
                        "title": post["title"],
                        "body": post["body"],
                        "engagement_score": result["classification"]["engagement_score"],
                        "engagement_tier": result["classification"]["engagement_tier"],
                        "created_at": post["created_at"],
                    }
                )

        trend_result = self.trend.run(verified_posts, window_start, window_end)

        total = len(window_posts)
        passed = len(verified_posts)
        signal_quality_score = passed / total if total > 0 else 0
        batch_summary = {
            "total_posts": total,
            "excluded_outside_window": len(posts) - total,
            "verified_posts": passed,
            "pass_rate": signal_quality_score,
            "signal_quality_score": signal_quality_score,
            "classification_cache_hits": 0,
            "estimated_llm_calls_this_run": 2,
            "trend_window": {
                "type": "rolling_30_day",
                "start": window_start.isoformat(),
                "end": window_end.isoformat(),
                "recurrence_rule": "Topic must appear in all four weekly buckets.",
            },
            "source_context": source_context or {"analysis_scope": "unspecified"},
            "trend_analysis": trend_result,
            "individual_results": results,
        }

        lens_context = {
            key: value for key, value in batch_summary.items() if key != "individual_results"
        }
        lens_result = self.lens.run(lens_context)
        return {"batch_summary": batch_summary, "intelligence": lens_result}

    def run_simulation(self):
        batches = [MOCK_POSTS]
        all_results = []
        for i, batch in enumerate(batches):
            result = self.run_batch(
                batch,
                source_context={
                    "analysis_scope": "mock_simulation",
                    "source": "data.mock_posts",
                },
            )
            result["batch_number"] = i + 1
            all_results.append(result)
        return all_results