feat(feasibility): P3可行性研究——opp_feasibility引擎(fp_estimate规则引擎FP/draft_feasibility八节报告/score_candidate评分/promote_to_project立项门禁)+report_type列(models/DDL/CRUD重生成)+详情页与PPT按类型选四节或八节模板(sections_for_type单一规则源)+列表类型徽标+prompt链路C+params/demand_mining字典种子+i18n词条

- fp_calc.calc()返回键实锤核对: total_fp/counts/errors(原total_ufp/by_type兜底猜错,已修)
- promote门禁修正: resolve_approval会把status置approved,confirmed只是中间态→门禁接受confirmed/approval_initiated/approved(原只认confirmed会导致审批通过后反而不能立项)
- _cluster_ctx补项目隔离(对齐_check_batch_visible口径)+全personal类别如实拒绝可行性样本不足
- 八节模板兜底节=需求规格;四节存量解析口径零变化
- opp_approvals CRUD手改(项目过滤)被xls2ui重生成吃掉→已checkout还原,只保留opp_reports新列部分

README.md

@@ -151,6 +151,45 @@ CRUD 页（`opp_reports/` `opp_approvals/`）保留供后台运维直接编辑
 - 验证脚本：scripts/p0_probe_vdb_embed*.py（VDB/embed 协议探测）、
   p1_cluster_analyze.py（多τ对比+缓存，调参用）、p1_create_tables.py（幂等建表）。
 
+## 可行性研究（P3，2026-09-14）
+
+挖掘类别 → 可行性研究 → 评分 → 立项（report_type=feasibility 八节报告）：
+
+- **opp_feasibility.py**（引擎）：
+  - `fp_estimate`：功能点估算。LLM 只提供结构化功能清单
+    （name/type=EI|EO|EQ|ILF|EIF/det/ret/ftr/evidence/confidence），
+    复杂度判定与 FP 加权**全部走规则引擎 fp_calc.py**
+    （pipeline-core `skills_library/all/function-point-counting/`，部署位
+    `skills/global/`，进程内 importlib 加载；加载失败如实报错，绝不 LLM 心算 FP）。
+    返回键对齐 fp_calc.calc() 实际输出：total_fp/counts(→by_type)/errors/pending。
+  - `draft_feasibility(cluster_id)`：类别须已完成共性提取（normalize done）且有
+    core/ext 标准需求（全 personal 时如实拒绝"可行性样本不足"）。FP 输入=每条
+    核心/扩展标准需求映射 1 ILF+1 EI+1 EQ（DET 按证据原子数粗估）；成本=FP×
+    opp_fp_pm_per_fp（默认0.05人月/FP）×opp_cost_wan_per_pm（默认3.0万元/人月），
+    params 可配。LLM 只写三节叙述（架构与可行性/商业价值/风险），其余五节用
+    挖掘真实数据拼装。落库 opp_reports（report_type=feasibility）。
+  - `score_candidate(cluster_id)`：规则评分=需求热度40%（log 规模防大簇垄断）+
+    共性集中度35%（core+ext 覆盖文档占比）+规模适配25%（标准需求 8-25 个最佳）。
+  - `promote_to_project(report_id)`：立项门禁——仅 report_type=feasibility 且
+    过人工确认（status∈confirmed/approval_initiated/approved）且审批单 approved
+    才可建 sd_projects（pipeline=opportunity_general，project_type=demand_mining，
+    字典种子在 init/data.json）。agent 不能自批自审。
+  - 项目隔离：`_cluster_ctx` 对齐 `_check_batch_visible` 口径（批次挂项目的仅
+    同项目会话可见，平台级批次本机构可见）。
+- **八节模板**（report_type=feasibility）：`## 市场需求 / ## 共性需求 / ## 需求规格 /
+  ## 架构与可行性 / ## FP与成本 / ## 商业价值 / ## 风险 / ## 结论建议`。
+  `sections_for_type(report_type)` 是单一规则源：详情页（opp_report_detail.dspy）
+  与 PPT（opp_ppt.py FEASIBILITY_ORDER）按类型选四节/八节渲染；
+  research 报告四节模板不变，存量报告解析口径零变化。
+- **opp_reports.report_type 列**（VARCHAR(16) DEFAULT 'research'）：
+  测试机/生产走 pipeline-app `deploy/migrations/m0030_opp_reports_report_type.json`
+  （dmig apply，幂等 ADD COLUMN）；models/opp_reports.json 已同步（新建表天然带列）。
+- **报告列表标签**：opp_reports_list.dspy 行内加类型徽标（调研=蓝/可行性研究=紫）。
+- PPT 文件名区分：feasibility_report_<software>_<rid8>.pptx。
+- 4 个新工具注册进 OPP_HANDLERS（opp_fp_estimate/opp_draft_feasibility/
+  opp_score_candidate/opp_promote_to_project），OPP_PROMPT 加「可行性研究与立项」
+  流程段；opp-workflow 技能同步（链路C）。
+
 ## 技能
 
 产线技能在 pipeline-core 技能库：`skills_library/pipelines/opportunity_general/`

i18n/en/msg.txt

@@ -37,3 +37,14 @@
 下载=Download
 无权查看=No Access
 暂无成果文件=No output files yet
+可行性研究=Feasibility Study
+调研=Research
+需求挖掘立项=Demand-Mining Initiative
+市场需求=Market Demand
+共性需求=Common Requirements
+需求规格=Requirement Spec
+架构与可行性=Architecture & Feasibility
+FP与成本=FP & Cost
+商业价值=Business Value
+风险=Risks
+结论建议=Conclusion

i18n/zh/msg.txt

@@ -37,3 +37,14 @@
 下载=下载
 无权查看=无权查看
 暂无成果文件=暂无成果文件
+可行性研究=可行性研究
+调研=调研
+需求挖掘立项=需求挖掘立项
+市场需求=市场需求
+共性需求=共性需求
+需求规格=需求规格
+架构与可行性=架构与可行性
+FP与成本=FP与成本
+商业价值=商业价值
+风险=风险
+结论建议=结论建议

init/data.json

@@ -29,6 +29,21 @@
                 {"k": "approved", "v": "审批通过"},
                 {"k": "rejected", "v": "审批驳回"}
             ]
+        },
+        {
+            "parentid": "sd_project_type",
+            "parentname": "项目类型",
+            "items": [
+                {"k": "demand_mining", "v": "需求挖掘立项"}
+            ]
+        },
+        {
+            "parentid": "opp_report_type",
+            "parentname": "研发报告类型",
+            "items": [
+                {"k": "research", "v": "调研报告"},
+                {"k": "feasibility", "v": "可行性研究"}
+            ]
         }
     ],
     "params": [
@@ -43,6 +58,8 @@
         {"name": "opp_norm_tau2", "value": "0.85", "desc": "归一化相似度触发阈值(只当触发器,LLM裁决)"},
         {"name": "opp_norm_topk", "value": "10", "desc": "原子kNN邻居数"},
         {"name": "opp_cov_core", "value": "0.6", "desc": "核心共性覆盖率阈值"},
-        {"name": "opp_cov_ext", "value": "0.4", "desc": "扩展共性覆盖率阈值"}
+        {"name": "opp_cov_ext", "value": "0.4", "desc": "扩展共性覆盖率阈值"},
+        {"name": "opp_fp_pm_per_fp", "value": "0.05", "desc": "FP→人月经验系数(可行性研究成本估算,P3)"},
+        {"name": "opp_cost_wan_per_pm", "value": "3.0", "desc": "人月单价万元(可行性研究成本估算,P3)"}
     ]
 }

models/opp_reports.json

@@ -50,6 +50,13 @@
             "nullable": "no",
             "default": "draft"
         },
+        {
+            "name": "report_type",
+            "title": "报告类型(research调研/feasibility可行性研究)",
+            "type": "str",
+            "length": 16,
+            "default": "research"
+        },
         {
             "name": "confirm_task_id",
             "title": "确认任务ID",

mysql.ddl.sql

@@ -17,6 +17,7 @@ CREATE TABLE opp_reports
   `title` VARCHAR(255) NOT NULL  comment '报告标题',
   `content` longtext CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci   comment '报告正文',
   `status` VARCHAR(24) NOT NULL  DEFAULT 'draft' comment '报告状态',
+  `report_type` VARCHAR(16)   DEFAULT 'research' comment '报告类型(research调研/feasibility可行性研究)',
   `confirm_task_id` VARCHAR(32)   comment '确认任务ID',
   `ppt_path` VARCHAR(500)   comment 'PPT文件路径',
   `confirmed_by` VARCHAR(64)   comment '确认人',

pipeline_opportunity/opp_ability.py

@@ -93,6 +93,14 @@ OPP_TOOLS = [
                    parameters={"cluster_id": "类别ID"}, category="mining"),
     ToolDefinition(name="opp_coverage_report", description="查看某类别的共性需求覆盖率报告：核心共性(覆盖率≥60%)/扩展共性(≥40%)/个性需求三层 + 能力域分组 + 每条标准需求的覆盖文档数与证据",
                    parameters={"cluster_id": "类别ID"}, category="mining"),
+    ToolDefinition(name="opp_fp_estimate", description="软件功能点估算（IFPUG）：传入结构化功能清单(name/type=EI|EO|EQ|ILF|EIF/det/ret/ftr/evidence/confidence)，复杂度与FP由规则引擎fp_calc计算（禁LLM心算），返回总FP/分类型计数/待确认项/四件套markdown",
+                   parameters={"functions_json": "功能清单JSON数组字符串", "project_name": "项目名(可选)"}, category="mining"),
+    ToolDefinition(name="opp_draft_feasibility", description="起草八节可行性研究报告（report_type=feasibility）：市场需求/共性需求/需求规格/架构与可行性/FP与成本/商业价值/风险/结论建议。FP走规则引擎，数据全部来自挖掘结果。返回 report_id",
+                   parameters={"cluster_id": "类别ID(须已完成共性提取)"}, category="mining"),
+    ToolDefinition(name="opp_score_candidate", description="研发项目候选评分（规则模型）：需求热度40%+共性集中度35%+规模适配25%→总分与★优先级",
+                   parameters={"cluster_id": "类别ID"}, category="mining"),
+    ToolDefinition(name="opp_promote_to_project", description="可行性报告立项：仅 report_type=feasibility 且过人工确认+审批approved 的报告可建 sd_projects 研发项目（门禁：agent不能自批自审）",
+                   parameters={"report_id": "报告ID"}, category="mining"),
 ]
 
 # ══════════════════ 产线 prompt 片段 ══════════════════
@@ -149,10 +157,28 @@ opp_create_report / opp_update_report 的正文必须用以下四个二级标题
 2. opp_mining_status 轮询进度（每轮回复用户当前状态；failed 时如实转述原因，禁隐瞒）。
 3. done 后 opp_list_clusters 看类别排名（热度/需求数/占比），向用户呈现 TopX 表格。
 4. 用户选定类别后 opp_cluster_detail 取类内样例（每条带来源URL），结合数据写分析报告；
-   需要立项走可行性研究时，用 opp_create_report 建报告（数据基础引用挖掘批次与类别证据）。
+   需要立项走可行性研究时用 **opp_draft_feasibility**（见下节），不要用
+   opp_create_report 手写可行性报告。
 硬规则：聚类结果只能来自 opp_* mining 工具返回，禁止自己编类别/编排名/编数量；
 类别命名证据（naming_evidence 样例）要能支撑命名，命名可疑时如实说明。
 
+## 可行性研究与立项（P3，report_type=feasibility）
+用户要「可行性研究/评估立项/能不能做这个项目」时走这条链：
+1. 类别先做共性提取：opp_normalize_cluster（未提取过时报错引导），
+   opp_coverage_report 看核心/扩展/个性三层共性。
+2. opp_fp_estimate——功能点估算：LLM 只识别功能（name/type=EI|EO|EQ|ILF|EIF/det/ftr），
+   复杂度与 FP 加权由规则引擎 fp_calc 计算，**禁止自己心算/编造 FP 数值**。
+3. opp_draft_feasibility(cluster_id)——自动生成八节可行性研究报告草稿
+   （市场需求/共性需求/需求规格/架构与可行性/FP与成本/商业价值/风险/结论建议），
+   数据全部来自挖掘与覆盖率结果，返回 report_id。可用 opp_update_report 补写完善。
+4. opp_score_candidate(cluster_id)——规则评分（需求热度40%+共性集中度35%+规模适配25%）
+   给出★优先级，向用户如实呈现评分因子。
+5. 立项门禁（不可绕过，agent 不能自批自审）：
+   opp_submit_report 提交人工确认 → 人工确认通过（confirmed）→ 发起研发审批 →
+   审批回流 approved 后，才可用 opp_promote_to_project(report_id) 立项
+   （sd_projects，pipeline=opportunity_general，project_type=demand_mining）。
+   报告未过确认/审批时 promote 会被门禁拒绝——如实转述拒绝原因，禁止绕道。
+
 ## 硬规则
 - **统计优先用聚合工具，明细只取样例（2026-09-11 实测教训）**：
   数量/排名/预算合计/平均价这类**统计**一律用聚合工具——opp_hot_software（招标侧）、
@@ -532,6 +558,55 @@ async def _h_coverage_report(sor, p, ctx):
         "标准需求": rows}, ensure_ascii=False, default=str)
 
 
+async def _h_fp_estimate(sor, p, ctx):
+    from .opp_feasibility import fp_estimate
+    fj = p.get("functions_json") or ""
+    if not fj:
+        return _fmt(False, "functions_json 必填")
+    ok, res = await fp_estimate(sor, fj, project_name=p.get("project_name") or "")
+    if not ok:
+        return _fmt(False, res)
+    return json.dumps({"总FP": res["total_fp"], "分类型": res["by_type"],
+                       "待确认": res["pending"],
+                       "输入校验错误": res["errors"],
+                       "四件套": res["markdown"][:2000]}, ensure_ascii=False, default=str)
+
+
+async def _h_draft_feasibility(sor, p, ctx):
+    from .opp_feasibility import draft_feasibility
+    cid = (p.get("cluster_id") or "").strip()
+    if not cid:
+        return _fmt(False, "cluster_id 必填")
+    ok, rid = await draft_feasibility(sor, cid, ctx)
+    if not ok:
+        return _fmt(False, rid)
+    return _fmt(True, "可行性报告草稿已建 report_id=%s（八节模板，FP 由规则引擎计算）。"
+                      "补写后可 opp_submit_report 提交人工确认" % rid)
+
+
+async def _h_score_candidate(sor, p, ctx):
+    from .opp_feasibility import score_candidate
+    cid = (p.get("cluster_id") or "").strip()
+    if not cid:
+        return _fmt(False, "cluster_id 必填")
+    ok, res = await score_candidate(sor, cid, ctx)
+    if not ok:
+        return _fmt(False, res)
+    return json.dumps(res, ensure_ascii=False, default=str)
+
+
+async def _h_promote_to_project(sor, p, ctx):
+    from .opp_feasibility import promote_to_project
+    rid = (p.get("report_id") or "").strip()
+    if not rid:
+        return _fmt(False, "report_id 必填")
+    ok, pid = await promote_to_project(sor, rid, ctx)
+    if not ok:
+        return _fmt(False, pid)
+    return _fmt(True, "已立项 project_id=%s（pipeline=opportunity_general，"
+                      "project_type=demand_mining）" % pid)
+
+
 OPP_HANDLERS = {
     "opp_hot_software": _h_hot_software,
     "opp_daily_ai_tenders": _h_daily_ai,
@@ -559,6 +634,11 @@ OPP_HANDLERS = {
     # 类内共性提取（P2）
     "opp_normalize_cluster": _h_normalize_cluster,
     "opp_coverage_report": _h_coverage_report,
+    # 可行性研究（P3）
+    "opp_fp_estimate": _h_fp_estimate,
+    "opp_draft_feasibility": _h_draft_feasibility,
+    "opp_score_candidate": _h_score_candidate,
+    "opp_promote_to_project": _h_promote_to_project,
 }
 
 

pipeline_opportunity/opp_feasibility.py

@@ -0,0 +1,296 @@
+# -*- coding:utf-8 -*-
+"""P3 可行性研究引擎：FP 估算（规则引擎禁 LLM 报数）→ 八节可行性报告 → 评分 → 立项。
+
+分工铁律（function-point-counting 技能）：LLM 只做语义（功能识别/DET/FTR 提取），
+复杂度判定与 FP 加权必须走规则引擎 fp_calc.py（skills/global 部署位，进程内 importlib 加载，
+加载失败如实报错，绝不 LLM 心算 FP）。
+
+八节可行性模板（report_type=feasibility）：
+  ## 市场需求 / ## 共性需求 / ## 需求规格 / ## 架构与可行性 /
+  ## FP与成本 / ## 商业价值 / ## 风险 / ## 结论建议
+"""
+import importlib.util
+import json
+import logging
+import os
+
+from .opp_common import get_db, new_id, rows_to_dicts, get_param
+from .opp_normalize import coverage_report, _llm
+
+logger = logging.getLogger("pipeline.opp_feasibility")
+
+# fp_calc.py 部署位（build.sh 6b 段复制 skills_library/all → skills/global）
+FP_CALC_CANDIDATES = [
+    "/d/pipeline/pipeline-app/skills/global/function-point-counting/scripts/fp_calc.py",
+    os.path.join(os.path.dirname(os.path.abspath(__file__)),
+                 "../../skills/global/function-point-counting/scripts/fp_calc.py"),
+]
+
+# 工作量/成本规则（人月/单价，params 可配；IFPUG FP→人月经验系数）
+P = {
+    "opp_fp_pm_per_fp": "0.05",     # 人月/FP（经验系数，可调）
+    "opp_cost_wan_per_pm": "3.0",   # 万元/人月
+}
+
+
+def _load_fp_calc():
+    for path in FP_CALC_CANDIDATES:
+        p = os.path.abspath(path)
+        if os.path.isfile(p):
+            spec = importlib.util.spec_from_file_location("fp_calc_mod", p)
+            if not spec or not spec.loader:
+                continue
+            mod = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(mod)
+            return mod
+    return None
+
+
+async def get_fp_params(sor):
+    cfg = {}
+    for k, dv in P.items():
+        cfg[k] = await get_param(sor, k, dv)
+    for k in P:
+        try:
+            cfg[k] = float(cfg[k])
+        except (TypeError, ValueError):
+            cfg[k] = float(P[k])
+    return cfg
+
+
+async def fp_estimate(sor, functions_json, project_name=""):
+    """functions_json: [{name,type(EI/EO/EQ/ILF/EIF),desc,evidence,det,ret,ftr,confidence}]
+    返回 (ok, dict|err)。复杂度与 FP 全部由 fp_calc 规则引擎算，LLM 只提供结构化输入。"""
+    try:
+        funcs = json.loads(functions_json) if isinstance(functions_json, str) else functions_json
+    except Exception:
+        return False, "functions_json 解析失败"
+    if not isinstance(funcs, list) or not funcs:
+        return False, "functions 必须是非空数组"
+    data = {"project": project_name or "可行性估算", "boundary": {"internal": [], "external": []},
+            "functions": [], "pending": [], "assumptions": []}
+    for f in funcs:
+        if not isinstance(f, dict):
+            continue
+        ft = str(f.get("type") or "").upper()
+        if ft not in ("EI", "EO", "EQ", "ILF", "EIF"):
+            data["pending"].append({"item": str(f.get("name") or "?"),
+                                    "reason": "类型非法: %s" % ft, "ask": "确认功能类型"})
+            continue
+        data["functions"].append({
+            "name": str(f.get("name") or "")[:80], "type": ft,
+            "desc": str(f.get("desc") or "")[:200],
+            "evidence": str(f.get("evidence") or "")[:200],
+            "det": int(f.get("det") or 0), "ret": int(f.get("ret") or (1 if ft in ("ILF", "EIF") else 0)),
+            "ftr": int(f.get("ftr") or 0),
+            "confidence": str(f.get("confidence") or "medium"),
+            "assumption": str(f.get("assumption") or "")[:200]})
+    mod = _load_fp_calc()
+    if not mod:
+        return False, "fp_calc.py 规则引擎未找到（skills/global 未部署 function-point-counting）"
+    try:
+        res = mod.calc(data)
+    except Exception as e:
+        return False, "fp_calc 计算失败: %s" % str(e)[:200]
+    # fp_calc.calc() 实际返回键（2026-09-14 核对源码）：
+    # functions / counts / total_fp / errors / pending / assumptions / boundary
+    return True, {
+        "total_fp": res.get("total_fp") or 0,
+        "by_type": res.get("counts") or {},
+        "functions": res.get("functions") or data["functions"],
+        "pending": data["pending"],
+        "errors": res.get("errors") or [],
+        "markdown": mod.to_markdown(res, project_name) if hasattr(mod, "to_markdown") else "",
+    }
+
+
+async def _cluster_ctx(sor, cluster_id, ctx):
+    """取类别+批次+覆盖率，校验 org/项目可见性。返回 (ok, dict|err)。
+
+    项目隔离口径与 opp_mining_flow._check_batch_visible 一致：批次挂了项目的，
+    仅同项目会话可见；平台级批次（project_id 空）本机构可见。
+    """
+    recs = await sor.sqlExe(
+        "SELECT c.*, b.org_id AS b_org, b.scope, b.project_id AS b_project "
+        "FROM opp_clusters c "
+        "LEFT JOIN opp_mining_batches b ON b.id=c.batch_id WHERE c.id=${c}$",
+        {"c": cluster_id})
+    await sor.sqlExe("COMMIT", {})
+    if not recs:
+        return False, "类别不存在"
+    cl = rows_to_dicts(recs, limit=1)[0]
+    org_id = str(cl.get("org_id") or cl.get("b_org") or "")
+    if org_id and org_id != (ctx.get("org_id") or ""):
+        return False, "类别不属于当前机构"
+    bproj = str(cl.get("b_project") or cl.get("project_id") or "")
+    if bproj and bproj != str(ctx.get("project_id") or ""):
+        return False, "批次属于其他项目，当前会话不可见"
+    ok, cov = await coverage_report(sor, ctx, cluster_id)
+    if not ok:
+        return False, cov
+    return True, {"cluster": cl, "coverage": cov}
+
+
+async def draft_feasibility(sor, cluster_id, ctx):
+    """起草八节可行性研究报告（report_type=feasibility）。返回 (ok, msg|report_id)。"""
+    ok, info = await _cluster_ctx(sor, cluster_id, ctx)
+    if not ok:
+        return False, info
+    cl = info["cluster"]
+    cov = info["coverage"]
+    if cov["cluster"].get("normalize_status") != "done":
+        return False, "类别未完成共性提取，先 opp_normalize_cluster"
+    from .opp_report_capability import create_report
+    stds = cov["std_demands"]
+    core = [s for s in stds if s["tier"] == "core"]
+    ext = [s for s in stds if s["tier"] == "ext"]
+    if not core and not ext:
+        return False, ("类别「%s」无核心/扩展共性（全部标准需求覆盖率<40%%），"
+                       "可行性样本不足——换一个共性更高的类别，"
+                       "或改用 opp_create_report 写调研报告" % (cl.get("name") or ""))
+
+    # FP 输入：核心+扩展标准需求 → 规则映射功能点（每标准需求≈1 ILF + 2 EI + 1 EO/EQ，
+    # 保守经验映射；DET 按证据原子数粗估）——LLM 不参与 FP 数值
+    funcs = []
+    for s in core + ext:
+        ev = json.loads(s.get("evidence_json") or "{}")
+        n_atoms = len(ev.get("atoms") or []) or 1
+        funcs.append({"name": s["name"] + "档案", "type": "ILF", "desc": s["name"],
+                      "evidence": "覆盖率%.0f%% 标准需求" % (float(s["coverage"] or 0) * 100),
+                      "det": min(19, 5 + n_atoms), "ret": 1, "ftr": 0, "confidence": "medium"})
+        funcs.append({"name": s["name"] + "维护", "type": "EI", "desc": "新增/修改" + s["name"],
+                      "evidence": "标准需求 CRUD", "det": min(15, 4 + n_atoms), "ret": 0,
+                      "ftr": 1, "confidence": "medium"})
+        funcs.append({"name": s["name"] + "查询", "type": "EQ", "desc": "检索" + s["name"],
+                      "evidence": "标准需求查询", "det": min(19, 4 + n_atoms), "ret": 0,
+                      "ftr": 1, "confidence": "medium"})
+    ok_fp, fp = await fp_estimate(sor, funcs, project_name=cl.get("name") or "")
+    if not ok_fp:
+        return False, "FP 估算失败: %s" % fp
+    cfg = await get_fp_params(sor)
+    pm = round(float(fp["total_fp"]) * cfg["opp_fp_pm_per_fp"], 1)
+    cost = round(pm * cfg["opp_cost_wan_per_pm"], 1)
+
+    # LLM 写三节叙述（架构/商业价值/风险），其余节用真实数据拼装
+    arch = cov_txt = biz = risk = ""
+    try:
+        arch = await _llm(
+            "为「%s」软件写技术架构与可行性分析（≤200字）：推荐技术栈、架构分层、"
+            "关键技术风险点。核心功能：%s。只输出正文。" % (
+                cl.get("name"), "、".join(s["name"] for s in core[:6]) or "、".join(
+                    s["name"] for s in stds[:6])), ctx, timeout=90)
+    except Exception as e:
+        arch = "（架构分析生成失败：%s）" % str(e)[:60]
+    try:
+        biz = await _llm(
+            "为「%s」软件写商业价值分析（≤150字）：目标客户、付费意愿依据（众包需求 %s 条、"
+            "核心共性 %d 项）、定价建议。只输出正文。" % (
+                cl.get("name"), cl.get("doc_count"), len(core)), ctx, timeout=90)
+    except Exception as e:
+        biz = "（商业价值分析生成失败：%s）" % str(e)[:60]
+    try:
+        risk = await _llm(
+            "为「%s」研发项目写风险清单（≤150字，3-5条，格式：- 风险：应对）。只输出正文。"
+            % cl.get("name"), ctx, timeout=90)
+    except Exception as e:
+        risk = "（风险分析生成失败：%s）" % str(e)[:60]
+
+    cov_txt = "\n".join("- %s（覆盖率 %.0f%%，%s，覆盖 %d 文档）" % (
+        s["name"], float(s["coverage"] or 0) * 100, s["tier"], s["doc_count"]) for s in stds[:20])
+    content = "\n\n".join([
+        "## 市场需求",
+        "类别「%s」来自需求挖掘批次 %s（scope=%s），类内众包需求 %s 条，占批次 %.1f%%。"
+        "数据来源：数据爬取平台众包需求（猪八戒/一品威客/Freelancer/PPH），明细可经"
+        " opp_cluster_detail 逐条查证。" % (
+            cl.get("name"), cl.get("batch_id"), cl.get("scope"), cl.get("doc_count"),
+            float(cl.get("share") or 0) * 100),
+        "## 共性需求",
+        cov_txt or "（无标准需求）",
+        "## 需求规格",
+        "核心共性（≥60%%）：%s\n扩展共性（≥40%%）：%s\n能力域：%s" % (
+            "、".join(s["name"] for s in core) or "无",
+            "、".join(s["name"] for s in ext) or "无",
+            "、".join(d["name"] for d in cov["domains"])),
+        "## 架构与可行性",
+        arch,
+        "## FP与成本",
+        "功能点合计 %s FP（规则引擎 fp_calc 计算，IFPUG 口径）；按 %.2f 人月/FP 估 %s 人月；"
+        "按 %.1f 万元/人月估研发成本 %s 万元。\n%s" % (
+            fp["total_fp"], cfg["opp_fp_pm_per_fp"], pm, cfg["opp_cost_wan_per_pm"], cost,
+            (fp.get("markdown") or "")[:1500]),
+        "## 商业价值",
+        biz,
+        "## 风险",
+        risk,
+        "## 结论建议",
+        "建议立项优先级见评分（opp_score_candidate）；本报告为草稿，需人工确认门禁后进入审批。",
+    ])
+    ok_r, rid = await create_report(
+        ctx.get("project_id") or "", cl.get("name") or "未命名类别",
+        title="%s 研发可行性研究" % (cl.get("name") or ""), analysis=content,
+        created_by=ctx.get("user_id") or "agent.opportunity", report_type="feasibility")
+    if not ok_r:
+        return False, rid
+    # 评分落库到报告 content 前置 JSON？不——评分独立工具按需算。记录 FP 摘要进批次无关处：
+    return True, rid
+
+
+async def score_candidate(sor, cluster_id, ctx):
+    """规则评分：市场需求量×共性集中度×规模适配→优先级分。返回 (ok, dict|err)。"""
+    ok, info = await _cluster_ctx(sor, cluster_id, ctx)
+    if not ok:
+        return False, info
+    cl = info["cluster"]
+    cov = info["coverage"]
+    stds = cov["std_demands"]
+    n_docs = int(cl.get("doc_count") or 0)
+    core = [s for s in stds if s["tier"] == "core"]
+    ext = [s for s in stds if s["tier"] == "ext"]
+    # 需求热度分（log 规模，防大簇垄断）
+    import math
+    heat = min(100.0, 20 * math.log10(n_docs + 1))
+    # 共性集中度：核心+扩展覆盖文档占比
+    covered = sum(s["doc_count"] for s in core + ext)
+    conc = (covered / n_docs * 100) if n_docs else 0
+    # 规模适配：标准需求 8-25 个为最佳研发规模（过小不值得/过大需分期）
+    n_std = len(stds)
+    size_fit = 100.0 if 8 <= n_std <= 25 else (70.0 if 5 <= n_std < 8 or 25 < n_std <= 40 else 40.0)
+    score = round(heat * 0.4 + conc * 0.35 + size_fit * 0.25, 1)
+    stars = max(1, min(5, int(score // 20) + 1))
+    return True, {
+        "cluster": cl.get("name"), "score": score, "priority": "★" * stars,
+        "factors": {"需求热度(40%)": round(heat, 1), "共性集中度(35%)": round(conc, 1),
+                    "规模适配(25%)": size_fit},
+        "detail": {"doc_count": n_docs, "core": len(core), "ext": len(ext), "std": n_std},
+    }
+
+
+async def promote_to_project(sor, report_id, ctx):
+    """审批通过的可行性报告 → 立项（sd_projects，pipeline=opportunity_general）。
+    门禁：报告必须 approved（人工审批回流），agent 不能自批自审。"""
+    recs = await sor.sqlExe(
+        "SELECT r.*, a.status AS appr_status FROM opp_reports r "
+        "LEFT JOIN opp_approvals a ON a.report_id=r.id "
+        "WHERE r.id=${r}$ ORDER BY a.created_at DESC LIMIT 1", {"r": report_id})
+    await sor.sqlExe("COMMIT", {})
+    if not recs:
+        return False, "报告不存在"
+    rep = rows_to_dicts(recs, limit=1)[0]
+    if rep.get("report_type") != "feasibility":
+        return False, "仅可行性研究报告可立项"
+    # 门禁链：draft→confirmed（人工确认）→approval_initiated→approved（审批回流）。
+    # resolve_approval 会把报告 status 置 approved，所以 confirmed 只是中间态——
+    # 状态在 confirmed/approval_initiated/approved 都说明已过人工确认门禁。
+    if rep.get("status") not in ("confirmed", "approval_initiated", "approved"):
+        return False, "报告未过人工确认门禁（status=%s），不能立项" % (rep.get("status") or "draft")
+    if (rep.get("appr_status") or "") != "approved":
+        return False, "研发审批未通过（status=%s），不能立项" % (rep.get("appr_status") or "无审批单")
+    from pipeline_service.project_capability import create_project
+    ok, pid = await create_project(
+        name="%s 研发项目" % (rep.get("software") or "未命名"),
+        project_type="demand_mining", description="由需求挖掘可行性报告立项：%s" % rep.get("title"),
+        org_id=ctx.get("org_id") or "0", created_by=ctx.get("user_id") or "",
+        pipeline_id="opportunity_general")
+    if not ok:
+        return False, "立项失败: %s" % pid
+    return True, pid

pipeline_opportunity/opp_ppt.py

@@ -41,6 +41,10 @@ MAX_BULLETS_PER_PAGE = 7
 # 四节模板顺序（与 opp_report_capability.REPORT_SECTIONS 一致）
 SECTION_ORDER = ("研发场景", "输入", "输出", "成果")
 
+# 可行性研究八节模板（与 opp_report_capability.FEASIBILITY_SECTIONS 一致，P3）
+FEASIBILITY_ORDER = ("市场需求", "共性需求", "需求规格", "架构与可行性",
+                     "FP与成本", "商业价值", "风险", "结论建议")
+
 
 def _rgb(t):
     from pptx.dml.color import RGBColor
@@ -140,18 +144,23 @@ def _split_lead(text):
     return "", _clean_inline(text)
 
 
-def _parse_report(content):
+def _parse_report(content, report_type="research"):
-    """按四节模板解析（与详情页同口径）：返回 [(节名, markdown)]。
+    """按节模板解析（与详情页同口径）：返回 [(节名, markdown)]。
 
-    无四节标题的旧报告 → 整篇归「输出」节（如实呈现，不编造分节）。
+    research=四节（研发场景/输入/输出/成果）；feasibility=八节。
+    无节标题的旧报告 → 整篇归「输出」节（如实呈现，不编造分节）。
     """
     content = content or ""
+    is_feas = report_type == "feasibility"
+    names = FEASIBILITY_ORDER if is_feas else SECTION_ORDER
+    fallback = "需求规格" if is_feas else "输出"
     pat = re.compile(
-        r"^##\s*(?:[一二三四五六七八九十\d]+[、.．]?\s*)?[*]*\s*"
+        r"^##\s*(?:[一二三四五六七八九十\d]+[、.．]?\s*)?[*]*\s*("
-        r"(研发场景|输入|输出|成果)\s*[*]*\s*[:：]?\s*$", re.MULTILINE)
+        + "|".join(re.escape(n) for n in names)
+        + r")\s*[*]*\s*[:：]?\s*$", re.MULTILINE)
     marks = list(pat.finditer(content))
     if not marks:
-        return [("输出", content.strip())]
+        return [(fallback, content.strip())]
     out = []
     preamble = content[:marks[0].start()].strip()
     if preamble:
@@ -161,7 +170,7 @@ def _parse_report(content):
         body = content[m.end():end].strip()
         if body:
             out.append((m.group(1), body))
-    return out or [("输出", content.strip())]
+    return out or [(fallback, content.strip())]
 
 
 # ══════════════════════ 版式基元 ══════════════════════
@@ -432,10 +441,11 @@ def _closing(prs, page_no):
 
 # ══════════════════════ 主入口 ══════════════════════
 
-def build_report_ppt(title, content, software="", out_path=""):
+def build_report_ppt(title, content, software="", out_path="", report_type="research"):
     """生成 PPT 并保存。返回 (True, 文件路径) 或 (False, 错误)。
 
-    签名保持不变（opp_report_capability._build_report_ppt 调用方零改动）。
+    report_type: research=四节章节页 / feasibility=八节章节页（P3）。
+    其余参数签名保持不变（旧调用方零改动）。
     """
     try:
         from pptx import Presentation
@@ -456,14 +466,15 @@ def build_report_ppt(title, content, software="", out_path=""):
     # 预扫描：算出要渲染的页面，便于一次性算总页数（页脚 x/y）
     pages = []          # [('divider', idx, name) | ('bullets', title, kicker, items)
                         #  | ('table', title, kicker, rows, extra)]
-    sections = _parse_report(content)
+    sections = _parse_report(content, report_type)
+    sec_names = FEASIBILITY_ORDER if report_type == "feasibility" else SECTION_ORDER
     sec_idx = 0
     for sec_name, sec_md in sections:
-        if sec_name in SECTION_ORDER:
+        if sec_name in sec_names:
             sec_idx += 1
             pages.append(("divider", sec_idx, sec_name))
         blocks = _split_blocks(sec_md)
-        cur_title = sec_name if sec_name not in SECTION_ORDER else sec_name
+        cur_title = sec_name
         buf_items = []          # [(lvl, lead, text)]
         pending_table = None
 

pipeline_opportunity/opp_report_capability.py

@@ -28,12 +28,19 @@ logger = logging.getLogger("pipeline.opportunity.report")
 
 
 async def create_report(project_id, software, title="", analysis="",
-                        created_by="agent.opportunity"):
+                        created_by="agent.opportunity", report_type="research"):
-    """创建研发报告草稿。返回 (True, report_id) 或 (False, 错误)。"""
+    """创建研发报告草稿。返回 (True, report_id) 或 (False, 错误)。
+
+    report_type（2026-09-12 P3）：research=调研报告（四节模板）/
+    feasibility=可行性研究（八节模板，opp_feasibility 起草）。
+    """
     if not software:
         return False, "缺少 software（软件/主题名）"
+    if report_type not in ("research", "feasibility"):
+        return False, "report_type 必须是 research 或 feasibility"
     if not title:
-        title = "%s 研发机会分析报告" % software
+        title = ("%s 研发可行性研究" % software) if report_type == "feasibility" \
+            else ("%s 研发机会分析报告" % software)
     db, dbname = get_db()
     rid = new_id()
     async with db.sqlorContext(dbname) as sor:
@@ -44,6 +51,7 @@ async def create_report(project_id, software, title="", analysis="",
             "title": title,
             "content": analysis or "",
             "status": RP_DRAFT,
+            "report_type": report_type,
             "created_by": created_by,
         })
         await sor.sqlExe("COMMIT", {})
@@ -89,6 +97,16 @@ async def list_reports(project_id="", status=""):
 # 报告正文四节模板（二级标题，agent 写报告必须按此结构；旧报告无节时全文归「输出」）
 REPORT_SECTIONS = ("研发场景", "输入", "输出", "成果")
 
+# 可行性研究八节模板（report_type=feasibility，P3 2026-09-14；
+# draft_feasibility 起草，详情页/PPT 按此分节渲染）
+FEASIBILITY_SECTIONS = ("市场需求", "共性需求", "需求规格", "架构与可行性",
+                        "FP与成本", "商业价值", "风险", "结论建议")
+
+
+def sections_for_type(report_type):
+    """按报告类型选节模板（单一规则源：详情页/PPT/解析复用）。"""
+    return FEASIBILITY_SECTIONS if report_type == "feasibility" else REPORT_SECTIONS
+
 
 async def _user_org(sor, user_id):
     recs = await sor.sqlExe(
@@ -114,7 +132,8 @@ async def list_visible_reports(sor, user_id):
     #   （实测 ValueError: unsupported format character；rag_client.py 的
     #   NOT LIKE 'agent.%' 是同一颗潜伏雷）。参数值里的 % 安全。
     sql = (
-        "SELECT r.id, r.project_id, r.software, r.title, r.status, r.ppt_path, "
+        "SELECT r.id, r.project_id, r.software, r.title, r.status, r.report_type, "
+        "r.ppt_path, "
         "r.created_by, r.created_at, r.updated_at, "
         "p.name AS project_name, p.created_by AS project_owner "
         "FROM opp_reports r LEFT JOIN sd_projects p ON r.project_id = p.id "
@@ -164,27 +183,32 @@ async def check_report_visible(sor, report_id, user_id):
     return False, rep
 
 
-def parse_report_sections(content):
+def parse_report_sections(content, report_type="research"):
-    """按四节模板解析报告正文 → {'研发场景': md, '输入': md, '输出': md, '成果': md}。
+    """按节模板解析报告正文 → {节名: markdown}。
 
-    模板（agent 写报告约定）：二级标题 `## 研发场景` / `## 输入` / `## 输出` / `## 成果`。
+    research（默认）：四节 `## 研发场景/输入/输出/成果`；
+    feasibility：八节 `## 市场需求/共性需求/需求规格/架构与可行性/FP与成本/商业价值/风险/结论建议`。
     兼容标题带序号/冒号/加粗（如 `## 一、研发场景：`、`## **输出**`）。
     旧报告（无任何匹配节，如实测的市场调研报告）→ 全文归「输出」节，其余节为空，
     详情页对空节如实显示「（本报告未填写此节）」，不编造。
     """
     import re as _re
     content = content or ""
+    sec_names = sections_for_type(report_type)
     # 节标题行：## [可选序号/标点] 节名 [可选冒号/加粗残留]
     pattern = _re.compile(
         r"^##\s*(?:[一二三四五六七八九十\d]+[、.．]?\s*)?[**]*\s*"
-        r"(研发场景|输入|输出|成果)\s*[**]*\s*[:：]?\s*$",
+        r"(" + "|".join(_re.escape(n) for n in sec_names) + r")\s*[**]*\s*[:：]?\s*$",
         _re.MULTILINE)
     marks = list(pattern.finditer(content))
-    sections = {k: "" for k in REPORT_SECTIONS}
+    sections = {k: "" for k in sec_names}
     if not marks:
-        sections["输出"] = content.strip()
+        # 无匹配节标题：四节模板归「输出」，八节模板归「需求规格」（承载正文主体）
+        fallback = "输出" if report_type != "feasibility" else "需求规格"
+        sections[fallback] = content.strip()
         return sections
-    # 首个节标题之前的前言（如报告大标题/数据来源说明）归「输出」节头部展示
+    # 首个节标题之前的前言（如报告大标题/数据来源说明）归兜底节头部展示
+    fallback = "输出" if report_type != "feasibility" else "需求规格"
     preamble = content[:marks[0].start()].strip()
     for i, m in enumerate(marks):
         name = m.group(1)
@@ -193,7 +217,7 @@ def parse_report_sections(content):
         # 同名节重复出现时拼接（容错，不丢内容）
         sections[name] = (sections[name] + "\n\n" + body).strip() if sections[name] else body
     if preamble:
-        sections["输出"] = (preamble + ("\n\n" + sections["输出"] if sections["输出"] else "")).strip()
+        sections[fallback] = (preamble + ("\n\n" + sections[fallback] if sections[fallback] else "")).strip()
     return sections
 
 
@@ -261,11 +285,13 @@ async def _build_report_ppt(sor, rep):
         project_dir = os.path.join(os.getcwd(), "files", "opp_reports")
     rid = rep.get("id", "")
     software = re.sub(r"[^\w\u4e00-\u9fff-]", "_", rep.get("software", "") or "report")[:40]
-    fname = "research_report_%s_%s.pptx" % (software, rid[:8])
+    prefix = "feasibility_report" if rep.get("report_type") == "feasibility" else "research_report"
+    fname = "%s_%s_%s.pptx" % (prefix, software, rid[:8])
     out_path = os.path.join(project_dir, "deliverables", fname)
     ok, res = build_report_ppt(
         rep.get("title", ""), rep.get("content", ""),
-        software=rep.get("software", ""), out_path=out_path)
+        software=rep.get("software", ""), out_path=out_path,
+        report_type=rep.get("report_type") or "research")
     if not ok:
         logger.warning("build_report_ppt 失败: %s", str(res)[:200])
         return ""

skill/SKILL.md

@@ -64,6 +64,16 @@ appbase `params` 表两个键（`opp_common.get_crawler_config` 读取，默认
 | `opp_mining_status(batch_id)` | 批次状态轮询（pulling/embedding/clustering/naming/done/failed+原因） |
 | `opp_list_clusters(batch_id)` | 类别排名表（热度/需求数/占比/命名证据）= TopX 热门类别 |
 | `opp_cluster_detail(cluster_id,limit)` | 类内需求明细样例（每条带来源URL可查证） |
+| `opp_normalize_cluster(cluster_id)` | 类内共性提取（原子化→归一→能力域→覆盖率，LLM多耗时正比文档数） |
+| `opp_coverage_report(cluster_id)` | 共性覆盖率报告（core≥60%/ext≥40%/个性三层+能力域分组） |
+| `opp_fp_estimate(functions_json,project_name)` | 功能点估算：LLM只给功能清单，复杂度/FP加权走规则引擎fp_calc（禁心算） |
+| `opp_draft_feasibility(cluster_id)` | 起草八节可行性报告（report_type=feasibility；需先normalize且有core/ext共性） |
+| `opp_score_candidate(cluster_id)` | 候选评分：热度40%+集中度35%+规模适配25%→★优先级 |
+| `opp_promote_to_project(report_id)` | 立项门禁：仅feasibility+人工确认+审批approved→sd_projects(demand_mining) |
+
+可行性报告八节模板（详情页/PPT 按 report_type 分节渲染，`sections_for_type` 单一规则源）：
+`市场需求/共性需求/需求规格/架构与可行性/FP与成本/商业价值/风险/结论建议`；
+调研报告（research）仍是四节模板，存量报告解析口径不变。
 
 角色：`agent.opp_analyst` / `agent.opp_writer`；slash：`/hot` `/ai` `/reports` `/oppdiag`。
 

wwwroot/agent/index.ui

@@ -6,8 +6,8 @@
             "widgettype": "HBox",
             "options": {"width": "100%", "alignItems": "center", "padding": "16px 24px 8px 24px", "cheight": 6, "gap": "12px"},
             "subwidgets": [
-                {"widgettype": "Title2", "options": {"text": "商机产线"}},
+                {"widgettype": "Title2", "options": {"otext": "商机产线", "text": "商机产线", "i18n": true}},
-                {"widgettype": "Text", "options": {"text": "热点软件 → 今日AI招标 → 检索 → 编写研发报告 → 人工确认 → 研发审批", "cfontsize": 0.9, "color": "#94a3b8"}},
+                {"widgettype": "Text", "options": {"otext": "热点软件 → 今日AI招标 → 检索 → 编写研发报告 → 人工确认 → 研发审批", "text": "热点软件 → 今日AI招标 → 检索 → 编写研发报告 → 人工确认 → 研发审批", "i18n": true, "cfontsize": 0.9, "color": "#94a3b8"}},
                 {"widgettype": "Filler"},
                 {
                     "widgettype": "Button",

wwwroot/api/opp_demand_items.dspy

@@ -27,13 +27,13 @@ async with DBPools().sqlorContext(dbname) as sor:
 rows = []
 if not ok:
 	rows.append({"widgettype": "Text", "options": {
-		"text": "查询失败：" + str(res)[:200],
+		"otext": "查询失败：${p0}", "text": "查询失败：${p0}", "i18n": True, "i18n_params": {"p0": str(str(res)[:200])},
 		"color": "#dc2626", "padding": "12px", "halign": "left"}})
 else:
 	items = res.get("items") or []
 	if not items:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "该主题近 %d 天无众包需求分项" % days, "color": "#64748b", "padding": "12px", "halign": "left"}})
+			"otext": "该主题近 ${p0} 天无众包需求分项", "text": "该主题近 ${p0} 天无众包需求分项", "i18n": True, "i18n_params": {"p0": str('%d' % (days,))}, "color": "#64748b", "padding": "12px", "halign": "left"}})
 	for it in items:
 		item_id = str(it.get("id") or "")
 		budget = it.get("budget_wan") or 0
@@ -90,7 +90,7 @@ return {
 			"padding": "12px 16px", "gap": "0px", "overflow": "auto"},
 		"subwidgets": [
 			{"widgettype": "Text", "options": {
-				"text": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。",
+				"otext": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。", "text": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。", "i18n": True,
 				"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}}
 		] + rows
 	}]

wwwroot/api/opp_item_detail.dspy

@@ -27,7 +27,7 @@ dl_base = entire_url("/pipeline-opportunity/api/opp_attachment_dl.dspy")
 body = []
 if not ok or not isinstance(det, dict):
 	body.append({"widgettype": "Text", "options": {
-		"text": "获取详情失败：" + str(det)[:200],
+		"otext": "获取详情失败：${p0}", "text": "获取详情失败：${p0}", "i18n": True, "i18n_params": {"p0": str(str(det)[:200])},
 		"color": "#dc2626", "padding": "12px", "halign": "left"}})
 else:
 	# ── 标题 ──
@@ -111,7 +111,7 @@ else:
 	# ── 公告/需求全文（markdown 渲染，链接可点击；2026-09-09 用户定夺改 MdWidget）──
 	detail_text = str(det.get("detail_text") or "").strip()
 	body.append({"widgettype": "Text", "options": {
-		"text": "全文内容", "cfontsize": 0.95, "color": "#0f172a",
+		"otext": "全文内容", "text": "全文内容", "i18n": True, "cfontsize": 0.95, "color": "#0f172a",
 		"halign": "left", "margin": "0 0 4px 0"}})
 	if detail_text:
 		shown = detail_text[:20000]
@@ -123,7 +123,7 @@ else:
 			"borderRadius": "8px", "margin": "0 0 10px 0", "bgcolor": "#ffffff"}})
 	else:
 		body.append({"widgettype": "Text", "options": {
-			"text": "爬取平台未保存该条全文，可点「打开原文页」查看。",
+			"otext": "爬取平台未保存该条全文，可点「打开原文页」查看。", "text": "爬取平台未保存该条全文，可点「打开原文页」查看。", "i18n": True,
 			"cfontsize": 0.82, "color": "#94a3b8", "halign": "left",
 			"margin": "0 0 10px 0"}})
 
@@ -140,7 +140,7 @@ else:
 			"margin": "0 0 10px 0"}})
 	elif not atts:
 		body.append({"widgettype": "Text", "options": {
-			"text": "该条暂无附件", "cfontsize": 0.8, "color": "#94a3b8",
+			"otext": "该条暂无附件", "text": "该条暂无附件", "i18n": True, "cfontsize": 0.8, "color": "#94a3b8",
 			"halign": "left", "margin": "0 0 10px 0"}})
 	else:
 		for a in atts:

wwwroot/api/opp_mining_clusters.dspy

@@ -48,7 +48,7 @@ else:
 	clusters = _res if isinstance(_res, list) else []
 	if not clusters:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "该批次没有聚类结果（可能范围内无需求数据）",
+			"otext": "该批次没有聚类结果（可能范围内无需求数据）", "text": "该批次没有聚类结果（可能范围内无需求数据）", "i18n": True,
 			"color": "#64748b", "padding": "12px", "halign": "left"}})
 	for c in clusters:
 		cid = str(c.get("id") or "")
@@ -93,7 +93,7 @@ else:
 					                                   "wrap": True}},
 				]},
 				{"widgettype": "Filler"},
-				{"widgettype": "Text", "options": {"text": "%d 条 ｜ %.1f%%" % (cnt, share),
+				{"widgettype": "Text", "options": {"otext": "${p0} 条 ｜ ${p1}%", "text": "${p0} 条 ｜ ${p1}%", "i18n": True, "i18n_params": {"p0": str('%d' % (cnt,)), "p1": str('%.1f' % (share,))},
 				                                   "cfontsize": 0.85, "color": "#64748b"}},
 				{"widgettype": "Text", "options": {"text": "›", "cfontsize": 1.1, "color": "#94a3b8"}},
 			]})
@@ -109,7 +109,7 @@ return {
 		            "gap": "4px", "overflow": "auto"},
 		"subwidgets": [
 			{"widgettype": "Text", "options": {
-				"text": "点击任一类别，查看类内需求明细（每条带来源链接可查证）。",
+				"otext": "点击任一类别，查看类内需求明细（每条带来源链接可查证）。", "text": "点击任一类别，查看类内需求明细（每条带来源链接可查证）。", "i18n": True,
 				"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}},
 		] + rows,
 	}],

wwwroot/api/opp_mining_items.dspy

@@ -57,7 +57,7 @@ else:
 	items = _res.get("samples") or []
 	if not items:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "该类别暂无明细", "color": "#64748b", "padding": "12px", "halign": "left"}})
+			"otext": "该类别暂无明细", "text": "该类别暂无明细", "i18n": True, "color": "#64748b", "padding": "12px", "halign": "left"}})
 	for it in items:
 		src_id = str(it.get("src_id") or "")
 		title = str(it.get("title") or "")
@@ -113,7 +113,7 @@ return {
 		            "padding": "12px 16px", "gap": "0px", "overflow": "auto"},
 		"subwidgets": [
 			{"widgettype": "Text", "options": {
-				"text": "点击任一条目弹出需求详情（数据来自数据爬取平台快照，每条标注来源平台）。",
+				"otext": "点击任一条目弹出需求详情（数据来自数据爬取平台快照，每条标注来源平台）。", "text": "点击任一条目弹出需求详情（数据来自数据爬取平台快照，每条标注来源平台）。", "i18n": True,
 				"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}},
 		] + rows,
 	}],

wwwroot/api/opp_mining_popup.dspy

@@ -77,8 +77,7 @@ def _start_script(scope):
 rows = []
 # ── 顶部：启动按钮 + 说明 ──
 rows.append({"widgettype": "Text", "options": {
-	"text": "需求挖掘：把众包平台（猪八戒/一品威客/Freelancer/PPH）的软件需求做语义聚类，"
+	"otext": "需求挖掘：把众包平台（猪八戒/一品威客/Freelancer/PPH）的软件需求做语义聚类，发现市场热门软件类别（TopX）或分析你指定的需求类型。批次后台执行，完成后点击行查看类别排名。", "text": "需求挖掘：把众包平台（猪八戒/一品威客/Freelancer/PPH）的软件需求做语义聚类，发现市场热门软件类别（TopX）或分析你指定的需求类型。批次后台执行，完成后点击行查看类别排名。", "i18n": True,
-	        "发现市场热门软件类别（TopX）或分析你指定的需求类型。批次后台执行，完成后点击行查看类别排名。",
 	"cfontsize": 0.8, "color": "#64748b", "halign": "left", "wrap": True, "margin": "0 0 8px 0"}})
 rows.append({"widgettype": "HBox", "options": {"width": "100%", "gap": "10px", "margin": "0 0 12px 0"},
 	"subwidgets": [
@@ -96,13 +95,12 @@ rows.append({"widgettype": "HBox", "options": {"width": "100%", "gap": "10px", "
 
 if not _ok:
 	rows.append({"widgettype": "Text", "options": {
-		"text": "查询失败：" + str(_res)[:200], "color": "#dc2626", "padding": "12px", "halign": "left"}})
+		"otext": "查询失败：${p0}", "text": "查询失败：${p0}", "i18n": True, "i18n_params": {"p0": str(str(_res)[:200])}, "color": "#dc2626", "padding": "12px", "halign": "left"}})
 else:
 	batches = _res if isinstance(_res, list) else []
 	if not batches:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "还没有挖掘批次。点击上方「⛏ 启动全量挖掘」开始（首次全量约3-5分钟；"
+			"otext": "还没有挖掘批次。点击上方「⛏ 启动全量挖掘」开始（首次全量约3-5分钟；embedding 有全机构共享缓存，二次批次显著更快）。", "text": "还没有挖掘批次。点击上方「⛏ 启动全量挖掘」开始（首次全量约3-5分钟；embedding 有全机构共享缓存，二次批次显著更快）。", "i18n": True,
-			        "embedding 有全机构共享缓存，二次批次显著更快）。",
 			"color": "#64748b", "padding": "12px", "halign": "left", "wrap": True}})
 	for b in batches:
 		bid = str(b.get("id") or "")

wwwroot/api/opp_overseas_items.dspy

@@ -27,13 +27,13 @@ async with DBPools().sqlorContext(dbname) as sor:
 rows = []
 if not ok:
 	rows.append({"widgettype": "Text", "options": {
-		"text": "查询失败：" + str(res)[:200],
+		"otext": "查询失败：${p0}", "text": "查询失败：${p0}", "i18n": True, "i18n_params": {"p0": str(str(res)[:200])},
 		"color": "#dc2626", "padding": "12px", "halign": "left"}})
 else:
 	items = res.get("items") or []
 	if not items:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "该主题近 %d 天无海外分项" % days, "color": "#64748b", "padding": "12px", "halign": "left"}})
+			"otext": "该主题近 ${p0} 天无海外分项", "text": "该主题近 ${p0} 天无海外分项", "i18n": True, "i18n_params": {"p0": str('%d' % (days,))}, "color": "#64748b", "padding": "12px", "halign": "left"}})
 	for it in items:
 		item_id = str(it.get("id") or "")
 		budget = it.get("budget_wan") or 0
@@ -90,7 +90,7 @@ return {
 			"padding": "12px 16px", "gap": "0px", "overflow": "auto"},
 		"subwidgets": [
 			{"widgettype": "Text", "options": {
-				"text": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。",
+				"otext": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。", "text": "点击任一条目弹出具体内容（数据来自数据爬取平台）；每条说明中已标注来源平台。", "i18n": True,
 				"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}}
 		] + rows
 	}]

wwwroot/api/opp_references_items.dspy

@@ -27,13 +27,13 @@ async with DBPools().sqlorContext(dbname) as sor:
 rows = []
 if not ok:
 	rows.append({"widgettype": "Text", "options": {
-		"text": "查询失败：" + str(res)[:200],
+		"otext": "查询失败：${p0}", "text": "查询失败：${p0}", "i18n": True, "i18n_params": {"p0": str(str(res)[:200])},
 		"color": "#dc2626", "padding": "12px", "halign": "left"}})
 else:
 	items = res.get("items") or []
 	if not items:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "该主题近 %d 天无分项" % days, "color": "#64748b", "padding": "12px", "halign": "left"}})
+			"otext": "该主题近 ${p0} 天无分项", "text": "该主题近 ${p0} 天无分项", "i18n": True, "i18n_params": {"p0": str('%d' % (days,))}, "color": "#64748b", "padding": "12px", "halign": "left"}})
 	for it in items:
 		item_id = str(it.get("id") or "")
 		budget = it.get("budget_wan") or 0
@@ -95,7 +95,7 @@ return {
 			"padding": "12px 16px", "gap": "0px", "overflow": "auto"},
 		"subwidgets": [
 			{"widgettype": "Text", "options": {
-				"text": "点击任一条目弹出具体内容（公告全文来自数据爬取平台，有附件的可直接下载）；每条说明中已标注数据来源。",
+				"otext": "点击任一条目弹出具体内容（公告全文来自数据爬取平台，有附件的可直接下载）；每条说明中已标注数据来源。", "text": "点击任一条目弹出具体内容（公告全文来自数据爬取平台，有附件的可直接下载）；每条说明中已标注数据来源。", "i18n": True,
 				"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}}
 		] + rows
 	}]

wwwroot/api/opp_references_popup.dspy

@@ -8,7 +8,7 @@ import json as _json
 
 uid = await get_user()
 if not uid:
-	return {"widgettype": "Text", "options": {"text": "请先登录", "color": "#dc2626", "padding": "12px"}}
+	return {"widgettype": "Text", "options": {"otext": "请先登录", "text": "请先登录", "i18n": True, "color": "#dc2626", "padding": "12px"}}
 
 dbname = get_module_dbname('pipeline-opportunity')
 
@@ -37,7 +37,7 @@ def _topic_row(cat, cnt, budget, target_url, popup_id, extra_txt=""):
 		{"widgettype": "Text", "options": {"text": cat, "cfontsize": 1.0,
 			"color": "#0f172a", "halign": "left"}},
 		{"widgettype": "Filler"},
-		{"widgettype": "Text", "options": {"text": "%d 条" % cnt,
+		{"widgettype": "Text", "options": {"otext": "${p0} 条", "text": "${p0} 条", "i18n": True, "i18n_params": {"p0": str('%d' % (cnt,))},
 			"cfontsize": 0.85, "color": "#64748b"}},
 	]
 	if extra_txt:
@@ -45,7 +45,7 @@ def _topic_row(cat, cnt, budget, target_url, popup_id, extra_txt=""):
 			"cfontsize": 0.75, "color": "#94a3b8"}})
 	else:
 		subs.append({"widgettype": "Text", "options": {
-			"text": "预算 %s 万" % ("{:,.0f}".format(budget) if budget else "—"),
+			"otext": "预算 ${p0} 万", "text": "预算 ${p0} 万", "i18n": True, "i18n_params": {"p0": str("{:,.0f}".format(budget) if budget else "—")},
 			"cfontsize": 0.85, "color": "#64748b"}})
 	subs.append({"widgettype": "Text", "options": {"text": "›",
 		"cfontsize": 1.1, "color": "#94a3b8"}})
@@ -70,7 +70,7 @@ else:
 	ranking = res.get("ranking") or []
 	if not ranking:
 		rows.append({"widgettype": "Text", "options": {
-			"text": "近30天暂无招标数据", "color": "#64748b", "padding": "12px", "halign": "left"}})
+			"otext": "近30天暂无招标数据", "text": "近30天暂无招标数据", "i18n": True, "color": "#64748b", "padding": "12px", "halign": "left"}})
 	for r in ranking:
 		rows.append(_topic_row(str(r.get("category", "")), r.get("tender_count", 0),
 			r.get("total_budget_wan", 0), items_url, "opp_refs_items_pw"))
@@ -81,7 +81,7 @@ if ok_dm:
 	dm_ranking = res_dm.get("ranking") or []
 	if dm_ranking:
 		dm_rows.append({"widgettype": "Text", "options": {
-			"text": "国内众包软件需求（猪八戒/一品威客，近30天）",
+			"otext": "国内众包软件需求（猪八戒/一品威客，近30天）", "text": "国内众包软件需求（猪八戒/一品威客，近30天）", "i18n": True,
 			"cfontsize": 0.95, "color": "#0f172a", "halign": "left",
 			"margin": "14px 0 6px 0", "padding": "6px 0 0 0",
 			"border": "1px solid #f1f5f9", "borderRadius": "6px", "bgcolor": "#f8fafc"}})
@@ -95,7 +95,7 @@ if ok_ov:
 	ov_ranking = res_ov.get("ranking") or []
 	if ov_ranking:
 		ov_rows.append({"widgettype": "Text", "options": {
-			"text": "海外众包软件需求（Freelancer.com / PeoplePerHour，近30天）",
+			"otext": "海外众包软件需求（Freelancer.com / PeoplePerHour，近30天）", "text": "海外众包软件需求（Freelancer.com / PeoplePerHour，近30天）", "i18n": True,
 			"cfontsize": 0.95, "color": "#0f172a", "halign": "left",
 			"margin": "14px 0 6px 0", "padding": "6px 0 0 0",
 			"border": "1px solid #f1f5f9", "borderRadius": "6px", "bgcolor": "#f8fafc"}})
@@ -112,7 +112,7 @@ return {
 		"gap": "6px", "overflow": "auto"},
 	"subwidgets": [
 		{"widgettype": "Text", "options": {
-			"text": "点击任一主题条目，弹出该主题的全部分项明细；明细中再点击任一条，弹出具体内容（数据来自数据爬取平台，含公告全文与附件下载）。",
+			"otext": "点击任一主题条目，弹出该主题的全部分项明细；明细中再点击任一条，弹出具体内容（数据来自数据爬取平台，含公告全文与附件下载）。", "text": "点击任一主题条目，弹出该主题的全部分项明细；明细中再点击任一条，弹出具体内容（数据来自数据爬取平台，含公告全文与附件下载）。", "i18n": True,
 			"cfontsize": 0.8, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}}
 	] + rows + dm_rows + ov_rows
 }

wwwroot/api/opp_report_detail.dspy

@@ -18,7 +18,8 @@ dbname = get_module_dbname('pipeline-opportunity')
 
 async with DBPools().sqlorContext(dbname) as sor:
 	from pipeline_opportunity.opp_report_capability import (
-		check_report_visible, parse_report_sections, list_report_files, REPORT_SECTIONS)
+		check_report_visible, parse_report_sections, list_report_files,
+		sections_for_type)
 	visible, rep = await check_report_visible(sor, report_id, uid)
 	await sor.sqlExe("COMMIT", {})
 	if not visible:
@@ -33,6 +34,8 @@ STATUS_TXT = {
 	"approved": "审批通过", "rejected": "已驳回",
 }
 
+TYPE_TXT = {"research": "调研", "feasibility": "可行性研究"}
+
 body = []
 
 # ── 标题 ──
@@ -43,19 +46,22 @@ body.append({"widgettype": "Text", "options": {
 
 # ── 元信息 ──
 st = str(rep.get("status") or "")
+rt = str(rep.get("report_type") or "research")
 proj = str(rep.get("project_name") or "")
-meta = "%s ｜ %s ｜ %s ｜ 创建人 %s" % (
+meta = "%s ｜ %s ｜ %s ｜ %s ｜ 创建人 %s" % (
 	str(rep.get("created_at") or "")[:16],
 	("项目：" + proj) if proj else "平台级调研（未挂项目）",
+	TYPE_TXT.get(rt, rt or "调研"),
 	STATUS_TXT.get(st, st or "—"),
 	str(rep.get("created_by") or "—"))
 body.append({"widgettype": "Text", "options": {
 	"text": meta, "cfontsize": 0.78, "color": "#64748b", "halign": "left",
 	"wrap": True, "margin": "0 0 10px 0"}})
 
-# ── 四节：研发场景 / 输入 / 输出 / 成果（按节解析正文；空节如实标注，不编造）──
+# ── 分节渲染：research 四节 / feasibility 八节（按 report_type 选模板；
-sections = parse_report_sections(str(rep.get("content") or ""))
+#    空节如实标注，不编造）──
-for name in REPORT_SECTIONS:
+sections = parse_report_sections(str(rep.get("content") or ""), rt)
+for name in sections_for_type(rt):
 	md = (sections.get(name) or "").strip()
 	body.append({"widgettype": "Text", "options": {
 		"text": name, "cfontsize": 0.95, "color": "#0f172a", "halign": "left",
@@ -70,7 +76,7 @@ for name in REPORT_SECTIONS:
 			"margin": "0 0 10px 0", "bgcolor": "#ffffff"}})
 	else:
 		body.append({"widgettype": "Text", "options": {
-			"text": "（本报告未填写此节）", "cfontsize": 0.82, "color": "#94a3b8",
+			"otext": "（本报告未填写此节）", "text": "（本报告未填写此节）", "i18n": True, "cfontsize": 0.82, "color": "#94a3b8",
 			"halign": "left", "margin": "0 0 10px 0"}})
 
 # ── 成果文件（多个列出，逐个下载）──
@@ -79,7 +85,7 @@ body.append({"widgettype": "Text", "options": {
 	"halign": "left", "margin": "0 0 4px 0"}})
 if not files:
 	body.append({"widgettype": "Text", "options": {
-		"text": "暂无成果文件（报告提交人工确认时会自动生成 PPT 存入项目 deliverables/）",
+		"otext": "暂无成果文件（报告提交人工确认时会自动生成 PPT 存入项目 deliverables/）", "text": "暂无成果文件（报告提交人工确认时会自动生成 PPT 存入项目 deliverables/）", "i18n": True,
 		"cfontsize": 0.82, "color": "#94a3b8", "halign": "left", "margin": "0 0 10px 0"}})
 for f in files:
 	size = int(f.get("size") or 0)

wwwroot/api/opp_reports_list.dspy

@@ -45,7 +45,7 @@ if not _cur_pid:
 			"options": {"width": "100%", "gap": "12px", "padding": "24px"},
 			"subwidgets": [
 				{"widgettype": "Text",
-				 "options": {"text": "当前会话还没有项目，请先选择或新建项目",
+				 "options": {"otext": "当前会话还没有项目，请先选择或新建项目", "text": "当前会话还没有项目，请先选择或新建项目", "i18n": True,
 							 "cfontsize": 1, "color": "#64748b"}},
 				{"widgettype": "HBox",
 				 "options": {"width": "100%", "gap": "10px"},
@@ -72,24 +72,37 @@ STATUS_TXT = {
 	"rejected": ("已驳回", "#dc2626"),
 }
 
+TYPE_TXT = {
+	"research": ("调研", "#0369a1"),
+	"feasibility": ("可行性研究", "#7c3aed"),
+}
+
 rows = []
 if not reports:
 	rows.append({"widgettype": "Text", "options": {
-		"text": "暂无可见的研发报告（列表按项目 owner 过滤：仅显示你 own 的项目及平台级调研报告）",
+		"otext": "暂无可见的研发报告（列表按项目 owner 过滤：仅显示你 own 的项目及平台级调研报告）", "text": "暂无可见的研发报告（列表按项目 owner 过滤：仅显示你 own 的项目及平台级调研报告）", "i18n": True,
 		"cfontsize": 0.85, "color": "#64748b", "padding": "12px", "halign": "left"}})
 for r in reports:
 	rid = str(r.get("id") or "")
 	st = str(r.get("status") or "")
 	st_txt, st_color = STATUS_TXT.get(st, (st or "—", "#64748b"))
+	rt = str(r.get("report_type") or "research")
+	rt_txt, rt_color = TYPE_TXT.get(rt, ("调研", "#0369a1"))
 	proj = str(r.get("project_name") or "")
 	meta = "%s ｜ %s ｜ %s" % (
 		str(r.get("created_at") or "")[:16],
 		("项目：" + proj) if proj else "平台级调研（未挂项目）",
 		st_txt)
 	row_subs = [
-		{"widgettype": "Text", "options": {
+		{"widgettype": "HBox", "options": {"width": "100%", "gap": "8px",
-			"text": str(r.get("title") or r.get("software") or "（无标题）"),
+										  "alignItems": "center"},
-			"cfontsize": 0.95, "color": "#0f172a", "halign": "left", "wrap": True}},
+		 "subwidgets": [
+			 {"widgettype": "Text", "options": {"text": rt_txt,
+				  "cfontsize": 0.7, "color": "#ffffff", "bgcolor": rt_color,
+				  "padding": "1px 8px", "borderRadius": "10px"}},
+			 {"widgettype": "Text", "options": {
+				  "text": str(r.get("title") or r.get("software") or "（无标题）"),
+				  "cfontsize": 0.95, "color": "#0f172a", "halign": "left", "wrap": True}}]},
 		{"widgettype": "HBox", "options": {"width": "100%", "gap": "10px",
 			"alignItems": "center"},
 			"subwidgets": [
@@ -122,7 +135,7 @@ return {
 		"padding": "12px 16px", "gap": "0px", "overflow": "auto"},
 	"subwidgets": [
 		{"widgettype": "Text", "options": {
-			"text": "研发报告（%d 份）｜按项目 owner 过滤；点击任一条查看研发场景/输入/输出/成果与成果文件下载" % len(reports),
+			"otext": "研发报告（${p0} 份）｜按项目 owner 过滤；点击任一条查看研发场景/输入/输出/成果与成果文件下载", "text": "研发报告（${p0} 份）｜按项目 owner 过滤；点击任一条查看研发场景/输入/输出/成果与成果文件下载", "i18n": True, "i18n_params": {"p0": str('%d' % (len(reports),))},
 			"cfontsize": 0.78, "color": "#64748b", "halign": "left", "margin": "0 0 8px 0"}}
 	] + rows
 }

wwwroot/opp_reports/get_opp_reports.dspy

@@ -61,6 +61,13 @@ fields_str=r'''[
         "nullable": "no",
         "default": "draft"
     },
+    {
+        "name": "report_type",
+        "title": "报告类型(research调研/feasibility可行性研究)",
+        "type": "str",
+        "length": 16,
+        "default": "research"
+    },
     {
         "name": "confirm_task_id",
         "title": "确认任务ID",

wwwroot/opp_reports/index.ui

@@ -121,6 +121,17 @@
         "datatype": "str",
         "label": "报告状态"
     },
+    {
+        "name": "report_type",
+        "title": "报告类型(research调研/feasibility可行性研究)",
+        "type": "str",
+        "length": 16,
+        "default": "research",
+        "cwidth": 16,
+        "uitype": "str",
+        "datatype": "str",
+        "label": "报告类型(research调研/feasibility可行性研究)"
+    },
     {
         "name": "confirm_task_id",
         "title": "确认任务ID",