Route capability entries as evidence pool

2026-06-29 11:55:16 +08:00
parent 178251e7e6
commit 1c08af34fa
9 changed files with 1337 additions and 94 deletions
@@ -0,0 +1,205 @@
+#!/usr/bin/env python3
+"""Evidence-backed topic discovery for ordinary users.
+
+This layer does not calculate astrology. It ranks existing full-reading
+evidence into a small set of next topics a user can tap or ask about.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+
+def _as_dict(value: Any) -> dict[str, Any]:
+    return value if isinstance(value, dict) else {}
+
+
+def _as_list(value: Any) -> list[Any]:
+    return value if isinstance(value, list) else []
+
+
+def _current_dasha(modules: dict[str, Any]) -> tuple[str | None, str | None, str | None, str | None]:
+    dasha = _as_dict(modules.get("dasha"))
+    current = _as_dict(dasha.get("current_dasha"))
+    antar = _as_dict(current.get("antardasha"))
+    return current.get("lord"), antar.get("lord"), current.get("start"), current.get("end")
+
+
+def _convergence_for(modules: dict[str, Any], *tokens: str) -> dict[str, Any]:
+    convergence = _as_dict(modules.get("dasa_convergence"))
+    activations = _as_dict(convergence.get("domain_activations"))
+    for domain, row in activations.items():
+        if any(token in str(domain).lower() for token in tokens):
+            found = dict(_as_dict(row))
+            found.setdefault("domain", domain)
+            return found
+    for row in _as_list(convergence.get("top_convergent_domains")):
+        if isinstance(row, dict) and any(token in str(row.get("domain", "")).lower() for token in tokens):
+            return row
+        if isinstance(row, (list, tuple)) and row and any(token in str(row[0]).lower() for token in tokens):
+            return {"domain": row[0], "convergence_level": row[1] if len(row) > 1 else None}
+    return {}
+
+
+def _vedastro_snapshot(modules: dict[str, Any], domain: str) -> dict[str, Any]:
+    overview = _as_dict(modules.get("vedastro_range_scan_result"))
+    metadata = _as_dict(overview.get("source_metadata"))
+    counts = _as_dict(metadata.get("domain_event_counts"))
+    statuses = _as_dict(metadata.get("domain_statuses"))
+    top = _as_dict(overview.get("top_events_by_domain")).get(domain)
+    status = statuses.get(domain) or overview.get("status")
+    event_count = int(counts.get(domain) or 0)
+    if status == "ok" or event_count:
+        return {
+            "status": "used",
+            "domain": domain,
+            "event_count": event_count,
+            "top_event": top if isinstance(top, dict) else None,
+            "source": "modules.vedastro_range_scan_result",
+        }
+    return {
+        "status": "blocked" if overview else "not_available",
+        "domain": domain,
+        "event_count": event_count,
+        "top_event": None,
+        "source": "modules.vedastro_range_scan_result",
+    }
+
+
+def _evidence_line(label: str, value: Any) -> dict[str, str]:
+    return {"label": label, "value": str(value)}
+
+
+def _topic(
+    *,
+    topic_id: str,
+    title: str,
+    reality_value: str,
+    why: str,
+    evidence: list[dict[str, str]],
+    confidence: str,
+    vedastro: dict[str, Any],
+    questions: list[str],
+    answer_mode: str = "tap_or_ask",
+    priority: int = 50,
+) -> dict[str, Any]:
+    return {
+        "id": topic_id,
+        "title": title,
+        "reality_value": reality_value,
+        "why_worth_exploring": why,
+        "evidence": evidence,
+        "confidence": confidence,
+        "vedastro": vedastro,
+        "suggested_questions": questions,
+        "answer_mode": answer_mode,
+        "priority": priority,
+    }
+
+
+def build_guided_topics(report: dict[str, Any]) -> list[dict[str, Any]]:
+    modules = _as_dict(report.get("modules"))
+    chart = _as_dict(report.get("chart") or modules.get("chart"))
+    planets = _as_dict(chart.get("planets"))
+    md, ad, md_start, md_end = _current_dasha(modules)
+    md_label = f"{md or '-'} / {ad or '-'}"
+    fbm = _as_dict(_as_dict(report.get("ai_prompt_pack")).get("evidence_snapshot")).get("functional_benefic_malefic")
+    fbm = _as_dict(fbm) or _as_dict(modules.get("functional_benefic_malefic"))
+
+    career_conv = _convergence_for(modules, "career", "status", "profession", "work")
+    marriage_conv = _convergence_for(modules, "marriage", "partnership", "relationship")
+    wealth_conv = _convergence_for(modules, "wealth", "finance", "income", "gain")
+    relationship = _as_dict(modules.get("relationship_strict_evidence"))
+    rel_judgement = _as_dict(relationship.get("event_judgement"))
+    rel_present = _as_dict(relationship.get("present_evidence"))
+    d9 = _as_dict(rel_present.get("d9_navamsa"))
+    ul = _as_dict(rel_present.get("upapada_lagna"))
+    dk = _as_dict(rel_present.get("darakaraka"))
+    ketu_house = _as_dict(planets.get("Ketu")).get("house")
+
+    topics = [
+        _topic(
+            topic_id="relationship_partnership",
+            title="婚恋与长期合作为什么是当前强主题",
+            reality_value="帮助用户判断关系、合作、相亲、公开关系或长期承诺是否值得深入推进。",
+            why="婚恋/合作不是靠用户主动问才触发；当前证据里第7宫、D9、UL、DK 与多系统时间层已经可读。",
+            evidence=[
+                _evidence_line("Vimshottari", md_label),
+                _evidence_line("Dasa 收敛", marriage_conv.get("convergence_level") or "not_found"),
+                _evidence_line("D9", f"Asc={_as_dict(d9.get('Ascendant')).get('sign', '-')}; 7th={_as_dict(d9.get('_d9_analysis')).get('navamsa_7th_sign', '-')}"),
+                _evidence_line("UL", f"{ul.get('sign', '-')} H{ul.get('house', '-')}"),
+                _evidence_line("DK", f"{dk.get('dk_planet', '-')} H{dk.get('dk_house', '-')}"),
+                _evidence_line("Strict verdict", rel_judgement.get("verdict") or "not_available"),
+            ],
+            confidence="medium" if marriage_conv or relationship else "low",
+            vedastro=_vedastro_snapshot(modules, "marriage"),
+            questions=[
+                "我现在适合认真发展关系，还是更适合筛选和观察？",
+                "我的伴侣画像、认识场景和相处风险是什么？",
+                "未来哪些时间窗口适合推进关系公开或承诺？",
+            ],
+            priority=90 if marriage_conv else 65,
+        ),
+        _topic(
+            topic_id="career_direction",
+            title="事业定位是否正在重构",
+            reality_value="帮助用户判断是继续深耕、换方向、做产品化，还是先修系统和长期资产。",
+            why="事业主题需要把10宫、A10/D10、多系统 Dasha 与 VedAstro 事业雷达放在一起看。",
+            evidence=[
+                _evidence_line("Vimshottari", md_label),
+                _evidence_line("10宫触发", f"Ketu house={ketu_house}" if ketu_house else "check D10/A10"),
+                _evidence_line("Dasa 收敛", career_conv.get("convergence_level") or "not_found"),
+                _evidence_line("Functional layer", f"benefics={fbm.get('functional_benefics', [])}; malefics={fbm.get('functional_malefics', [])}"),
+            ],
+            confidence="medium" if career_conv or ketu_house == 10 else "low",
+            vedastro=_vedastro_snapshot(modules, "career"),
+            questions=[
+                "我现在适合换方向还是继续深耕？",
+                "2026 年事业吉利在哪里，不利在哪里？",
+                "哪些月份适合推进项目、发布产品或谈合作？",
+            ],
+            priority=82 if career_conv or ketu_house == 10 else 60,
+        ),
+        _topic(
+            topic_id="birth_time_rectification",
+            title="出生时间是否需要微调",
+            reality_value="帮助用户把婚恋、事业、财富应期从泛泛判断推进到可回验时间窗口。",
+            why="D9、D10、UL、A10 对出生时间敏感；如果用户想问具体月份/日期，先校正时间更有价值。",
+            evidence=[
+                _evidence_line("birth time", _as_dict(report.get("birth_info")).get("time", "-")),
+                _evidence_line("sensitive layers", "D9 / D10 / UL / A10"),
+                _evidence_line("current timing", f"{md_label}; {md_start or '-'} to {md_end or '-'}"),
+            ],
+            confidence="medium",
+            vedastro=_vedastro_snapshot(modules, "marriage"),
+            questions=[
+                "我可以用过去事件校正出生时间吗？",
+                "哪些人生事件最适合用来校正出生时间？",
+                "我只知道一个时间区间，系统应该先问我哪些 yes/no 问题？",
+            ],
+            answer_mode="yes_no_or_free_text",
+            priority=80,
+        ),
+        _topic(
+            topic_id="wealth_risk",
+            title="财富、借贷和交易风险怎样用数据拆开",
+            reality_value="帮助用户把收入、现金流、借贷、买卖和投资风险分开判断，而不是只说财运好坏。",
+            why="财富主题必须同时看2宫、11宫、D2/D11、Dasha 与 VedAstro wealth 标签。",
+            evidence=[
+                _evidence_line("Vimshottari", md_label),
+                _evidence_line("Dasa 收敛", wealth_conv.get("convergence_level") or "not_found"),
+                _evidence_line("required vargas", "D2 / D11"),
+            ],
+            confidence="medium" if wealth_conv else "low",
+            vedastro=_vedastro_snapshot(modules, "wealth"),
+            questions=[
+                "2026 年哪些钱可以赚，哪些钱要避险？",
+                "我适合靠项目、投资、合作还是长期积累赚钱？",
+                "哪些时间窗口不适合借贷、买卖或大额投入？",
+            ],
+            priority=70 if wealth_conv else 50,
+        ),
+    ]
+
+    topics.sort(key=lambda item: (-int(item.get("priority", 0)), item["id"]))
+    return topics[:4]