fix(report): project annual sections and thicken dasha reading

完整数据版第二轮。年运章写出资料包里已有的节，没有数据的不新算。大运解释补回可能段落。瑜伽作用不再断言寿命。术语替换不进标识符。罗睺和计都不写守护第未列出宫。未合入 staging。

scripts/tajika_mudda.py

@@ -18,8 +18,10 @@ of the same rule. What it adds on top of the upstream function:
 - `periods` are cropped to the annual window [annual_start, annual_end) and
   carry `months` (share of 12) for the existing month-window consumers; the
   upstream ten-group surface (the last group runs past the year end) stays
-  in `periods_full`. Upstream's sub-periods are not carried (no product
-  surface shows them).
+  in `periods_full`. Sub-periods already computed by
+  `calc_mudda_with_dasha_balance` are copied as `antardasha_periods` and
+  cropped to the same annual window. Day rows and a second balance profile
+  are not added here.
 
 The legacy `tajika.calc_mudda_dasha` (first lord = year lord) is no longer
 called from any product path.
@@ -90,6 +92,23 @@ def annual_mudda_dasha(
             "start": p_start.isoformat(sep=" "),
             "end": p_end.isoformat(sep=" "),
         })
+    antardasha_periods = []
+    for row in raw.get("antardasha_periods") or []:
+        if not isinstance(row, dict) or not row.get("start") or not row.get("end"):
+            continue
+        p_start = max(_parse(row["start"]), start)
+        p_end = min(_parse(row["end"]), end)
+        if p_start >= p_end:
+            continue
+        antardasha_periods.append({
+            "main_order": row.get("main_order"),
+            "sub_order": row.get("sub_order"),
+            "main_lord": row.get("main_lord"),
+            "sub_lord": row.get("sub_lord"),
+            "duration_days": (p_end - p_start).total_seconds() / 86400.0,
+            "start": p_start.isoformat(sep=" "),
+            "end": p_end.isoformat(sep=" "),
+        })
     return {
         # The first-lord rule is the astrologer's round-5 ruling; dates are
         # not externally replayed (upstream keeps its own parameter_sensitive).
@@ -110,6 +129,7 @@ def annual_mudda_dasha(
         "balance_profile": raw.get("balance_profile"),
         "balance_fraction_remaining": raw.get("balance_fraction_remaining"),
         "periods": periods,
+        "antardasha_periods": antardasha_periods,
         "periods_full": raw.get("periods"),
         "independent_of_year_lord": True,
         "source": SOURCE,