File size: 7,062 Bytes
a9e46a4 | 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 | from __future__ import annotations
from pathlib import Path
from typing import Any
import json
from .schemas import ExperimentReport, Hypothesis, Insight, PipelineConfiguration, ValidationReport
class SharedKnowledgeSpace:
"""File-based bridge between V1 and T1."""
def __init__(self, root: str | Path):
self.root = Path(root)
self.reports_dir = self.root / "experiment_reports"
self.insights_dir = self.root / "insights"
self.configs_dir = self.root / "pipeline_configs"
self.hypotheses_dir = self.root / "hypotheses"
self.validation_reports_dir = self.root / "validation_reports"
self.summary_stats_file = self.root / "summary_statistics.json"
for d in (
self.reports_dir,
self.insights_dir,
self.configs_dir,
self.hypotheses_dir,
self.validation_reports_dir,
):
d.mkdir(parents=True, exist_ok=True)
@staticmethod
def _write_json(path: Path, payload: dict[str, Any]) -> None:
path.write_text(json.dumps(payload, indent=2, ensure_ascii=True), encoding="utf-8")
@staticmethod
def _read_json(path: Path) -> dict[str, Any]:
return json.loads(path.read_text(encoding="utf-8"))
def save_report(self, report: ExperimentReport) -> Path:
fp = self.reports_dir / f"{report.run_id}.json"
self._write_json(fp, report.to_dict())
return fp
def save_insight(self, insight: Insight) -> Path:
fp = self.insights_dir / f"{insight.insight_id}.json"
self._write_json(fp, insight.to_dict())
return fp
def save_pipeline_config(self, config: PipelineConfiguration) -> Path:
fp = self.configs_dir / f"{config.config_id}.json"
self._write_json(fp, config.to_dict())
return fp
def save_hypothesis(self, hypothesis: Hypothesis) -> Path:
fp = self.hypotheses_dir / f"{hypothesis.hypothesis_id}.json"
self._write_json(fp, hypothesis.to_dict())
return fp
def save_validation_report(self, report: ValidationReport) -> Path:
fp = self.validation_reports_dir / f"{report.validation_id}.json"
self._write_json(fp, report.to_dict())
return fp
def list_reports(self) -> list[dict[str, Any]]:
items: list[dict[str, Any]] = []
for fp in sorted(self.reports_dir.glob("*.json")):
items.append(self._read_json(fp))
return items
def list_insights(self) -> list[dict[str, Any]]:
items: list[dict[str, Any]] = []
for fp in sorted(self.insights_dir.glob("*.json")):
items.append(self._read_json(fp))
return items
def list_configs(self) -> list[dict[str, Any]]:
items: list[dict[str, Any]] = []
for fp in sorted(self.configs_dir.glob("*.json")):
items.append(self._read_json(fp))
return items
def list_hypotheses(self) -> list[dict[str, Any]]:
items: list[dict[str, Any]] = []
for fp in sorted(self.hypotheses_dir.glob("*.json")):
items.append(self._read_json(fp))
return items
def list_validation_reports(self) -> list[dict[str, Any]]:
items: list[dict[str, Any]] = []
for fp in sorted(self.validation_reports_dir.glob("*.json")):
items.append(self._read_json(fp))
return items
def latest_config_for_task(self, task_scope: str) -> dict[str, Any] | None:
matched = [cfg for cfg in self.list_configs() if cfg.get("task_scope") == task_scope]
return matched[-1] if matched else None
def get_hypothesis_context(self, domain: str, top_k: int = 5) -> list[dict[str, Any]]:
domain_l = domain.strip().lower()
hypotheses = self.list_hypotheses()
validations = self.list_validation_reports()
by_hid = {}
for vr in validations:
hid = vr.get("hypothesis_id", "")
if not hid:
continue
by_hid.setdefault(hid, []).append(vr)
scored: list[dict[str, Any]] = []
for h in hypotheses:
h_domain = str(h.get("domain", "")).lower()
if domain_l and domain_l not in h_domain and domain_l not in " ".join(h.get("tags", [])).lower():
continue
reports = by_hid.get(h.get("hypothesis_id", ""), [])
if reports:
avg_score = sum(float(r.get("score", 0.0)) for r in reports) / len(reports)
avg_conf = sum(float(r.get("confidence", 0.0)) for r in reports) / len(reports)
else:
avg_score = 0.0
avg_conf = 0.0
scored.append(
{
"hypothesis": h,
"validation_reports": reports,
"avg_score": avg_score,
"avg_confidence": avg_conf,
}
)
scored.sort(key=lambda x: (x["avg_score"], x["avg_confidence"]), reverse=True)
return scored[: max(0, top_k)]
def update_summary_statistics(self) -> dict[str, Any]:
validations = self.list_validation_reports()
tag_stats: dict[str, dict[str, float]] = {}
failure_reason_counter: dict[str, int] = {}
success_scores: list[float] = []
hypotheses = {h.get("hypothesis_id", ""): h for h in self.list_hypotheses()}
for r in validations:
status = r.get("status", "inconclusive")
score = float(r.get("score", 0.0))
if status == "success":
success_scores.append(score)
reason = str(r.get("failure_reason", "")).strip()
if status == "failed" and reason:
failure_reason_counter[reason] = failure_reason_counter.get(reason, 0) + 1
hid = r.get("hypothesis_id", "")
h = hypotheses.get(hid, {})
for tag in h.get("tags", []):
s = tag_stats.setdefault(tag, {"count": 0.0, "success": 0.0, "score_sum": 0.0})
s["count"] += 1
if status == "success":
s["success"] += 1
s["score_sum"] += score
tag_summary = {}
for tag, s in tag_stats.items():
count = max(1.0, s["count"])
tag_summary[tag] = {
"count": int(s["count"]),
"success_rate": s["success"] / count,
"avg_score": s["score_sum"] / count,
}
payload = {
"total_validation_reports": len(validations),
"overall_avg_success_score": (sum(success_scores) / len(success_scores)) if success_scores else 0.0,
"tag_summary": tag_summary,
"failure_reason_counter": failure_reason_counter,
}
self._write_json(self.summary_stats_file, payload)
return payload
def get_summary_statistics(self) -> dict[str, Any]:
if not self.summary_stats_file.exists():
return self.update_summary_statistics()
return self._read_json(self.summary_stats_file)
|