学习路径多阶段化：拓扑层按每阶段≤3技能拆分，阶段命名改为 基础补齐→岗位核心*→项目验证；简历裸姓名/专硕识别修复

app/match/engine.py

@@ -27,27 +27,37 @@ def _coverage(have: set[str], want: list[str]) -> tuple[list[str], list[str], fl
     return matched, missing, cov
 
 
-def topological_stages(missing: list[str], lexicon: dict) -> list[dict]:
-    """把缺失技能按前置关系拓扑排序分阶段（Kahn 分层）。
+def topological_stages(missing: list[str], lexicon: dict,
+                       max_per_stage: int = 3) -> list[dict]:
+    """把缺失技能按前置关系拓扑排序分阶段（Kahn 分层 + 每层按上限再拆分）。
 
     前置关系取自技能词表 prereqs；前置技能若已掌握或不在缺失集中则视为已满足。
+    同一拓扑层技能过多时按 max_per_stage 拆成多个连续阶段，让路线图呈
+    "基础补齐 → 岗位核心 → … → 项目验证"的多阶段结构。
     每个阶段附带结构化验收标准（产出物 / 量化验收项 / 自测问答），按技能类别生成。
     """
     prereq = {s["name"]: [p for p in s.get("prereqs", []) if p in missing]
               for s in lexicon["skills"] if s["name"] in missing}
     cat_of = {s["name"]: s.get("category", "") for s in lexicon["skills"]}
     remaining = set(missing)
-    stages: list[dict] = []
-    stage_no = 1
+    layers: list[list[str]] = []
     while remaining:
         ready = sorted(s for s in remaining if not [p for p in prereq.get(s, []) if p in remaining])
         if not ready:  # 有环兜底：按字典序放出，保证终止
             ready = sorted(remaining)
-        stages.append({"stage": stage_no, "skills": ready,
-                       "goal": f"第 {stage_no} 阶段：掌握 {'、'.join(ready)}",
-                       "acceptance": build_acceptance(ready, cat_of)})
+        layers.append(ready)
         remaining -= set(ready)
-        stage_no += 1
+
+    # 拓扑层 → 按每阶段技能数上限拆分
+    stages: list[dict] = []
+    stage_no = 1
+    for layer in layers:
+        for i in range(0, len(layer), max_per_stage):
+            chunk = layer[i:i + max_per_stage]
+            stages.append({"stage": stage_no, "skills": chunk,
+                           "goal": f"第 {stage_no} 阶段：掌握 {'、'.join(chunk)}",
+                           "acceptance": build_acceptance(chunk, cat_of)})
+            stage_no += 1
     return stages
 
 

tests/test_match.py

@@ -105,3 +105,15 @@ def test_resume_bare_name_and_edu_alias():
     assert p.education == "硕士"                                     # 专硕→硕士
     assert p.address.startswith("辽宁大连")
     assert "Python" in p.skills
+
+
+def test_learning_path_stage_chunking():
+    """同层技能超 3 个时拆成多个连续阶段；拓扑序不变。"""
+    stages = topological_stages(["Linux", "Docker", "Kubernetes", "Python", "Java", "Go", "MySQL"],
+                                load_lexicon())
+    # 每阶段 ≤3 个技能
+    assert all(len(st["skills"]) <= 3 for st in stages)
+    flat = [s for st in stages for s in st["skills"]]
+    # 前置关系仍满足：Linux 在 Docker 前，Docker 在 Kubernetes 前
+    assert flat.index("Linux") < flat.index("Docker") < flat.index("Kubernetes")
+    assert len(stages) >= 3  # 7 个技能至少拆成 3 段

webui/app.js

@@ -1806,9 +1806,10 @@ function renderPath(data) {
     return;
   }
   const weeks = stageWeeks(data.learning_path.length);
-  /* 阶段语义化命名：基础补齐 → 岗位核心 → 项目验证（超出 3 阶段顺延为综合提升） */
-  const STAGE_NAMES = ["基础补齐", "岗位核心", "项目验证"];
-  const stageName = (i) => STAGE_NAMES[i] || "综合提升";
+  const weeks = stageWeeks(data.learning_path.length);
+  /* 阶段语义化命名：基础补齐 → 岗位核心（可多个）→ 项目验证 */
+  const stageName = (i, n) =>
+    i === 0 ? "基础补齐" : i === n - 1 ? "项目验证" : "岗位核心";
   box.innerHTML = (data._demo ? `
     <div class="path-demo-banner">
       <b>示范路线</b>（以零已掌握技能生成）：完成一次人岗诊断后，此处将自动替换为你的个人成长路线图
@@ -1820,7 +1821,7 @@ function renderPath(data) {
       <span class="stage-no">${done ? "✓" : st.stage}</span>
       <div class="stage-body">
         <div class="stage-head">
-          <span class="stage-title">阶段 ${st.stage} · ${stageName(i)}（${st.skills.length} 项技能）</span>
+          <span class="stage-title">阶段 ${st.stage} · ${stageName(i, data.learning_path.length)}（${st.skills.length} 项技能）</span>
           <span class="stage-weeks">建议 ${weeks[i]} 周</span>
         </div>
         ${i > 0 ? `<div class="stage-prereq">前置：建议完成阶段 ${st.stage - 1} 后开始（按技能前置关系拓扑排序）</div>` : ""}