fix(charts): 探针模式下跳过 §8.3 关注词场景条图并清理旧 PNG

Made-with: Cursor

backend/crawler_copy/jd_pc_search/jd_competitor_report.py

@@ -2733,7 +2733,7 @@ def build_competitor_markdown(
         "- **归因**：每条评价按其 SKU 对应到深入样本，再映射到该 SKU 所属细类；SKU 不在合并表中的评价单独归入说明性分组；**在合并表中但该 SKU 缺 ``detail_category_path`` 或路径无法解析为可读细类的，该评价不进入按细类统计**（与 §5 **同一条排除规则**）。",
         "- **正负面粗判（§8.2）**：若评价含有效「评分」列则**先按星级**粗分正负与中评，再在对应子集内统计口语短语；无评分时仍按关键词子串；若任务开启 **llm_comment_sentiment**，可附**大模型对抽样原文的主题归因**，与条形图互补。",
         (
-            "- **文本挖掘探针（§8.3）**：本任务已启用 **jieba + sklearn** 的开放词表分析（词频 / TF-IDF / 共现 / LDA，可选词云），与 §8.2 规则词表条形图**口径不同**、**互补**；**不再**输出原「关注词次数 + 场景占比」左右并列条图。"
+            "- **文本挖掘探索（§8.3）**：本任务已启用 **jieba + sklearn** 的开放词表分析（词频 / TF-IDF / 共现 / LDA，可选词云），与 §8.2 规则词表条形图**不同**、**互补**；**不再**输出原「关注词次数 + 场景占比」左右并列条图。"
             if _ch8_probe_sec
             else "- **关注词与使用场景（§8.3）**：对组内评价正文做关注词子串计数（左栏条形图）；对每条有效文本独立扫描**本次任务生效的场景词组**（来自报告调参或系统默认），一条可属多场景，右栏为**占该细类有效文本比例 %**（多标签下可相加 **>** 100%）。二者在 **同一张图左右并列**，与 §5 矩阵细类一一对应。"
         ),
@@ -2744,7 +2744,7 @@ def build_competitor_markdown(
     ]
     if _sm_score:
         _sec82_block.append(
-            "- **四象限口径**：本批存在有效「评分」时——**1～2 星**计为偏负向，**4～5 星**计为偏正向，**3 星**计为中评，**空文本**计为中性；"
+            "- **正负面粗判规模**：本批存在有效「评分」时——**1～2 星**计为偏负向，**4～5 星**计为偏正向，**3 星**计为中评，**空文本**计为中性；"
             "无评分的条仍按关键词子串划分；「混合」仅在**无评分**且同条兼含正/负关键词时出现。"
         )
     _sec82_block.extend(
@@ -2778,7 +2778,7 @@ def build_competitor_markdown(
         _embed_chart(
             run_dir,
             "chart_sentiment_overview_pie.png",
-            "评价语气占比（扇形图；与上表条数一致）",
+            "评价正负面粗判规模（扇形图；与上表条数一致）",
         )
     )
     lines.extend(
@@ -2844,6 +2844,7 @@ def build_competitor_markdown(
             ]
         )
     else:
+        # 仅当未嵌入文本挖掘探针（_ch8_probe_sec 为空）时：原「关注词 + 场景」条图与逐细类段落
         lines.extend(
             [
                 "### 8.3 关注词与使用场景（按细类）",

backend/pipeline/jd/runner.py

@@ -356,7 +356,7 @@ def write_competitor_analysis_for_run_dir(
     )
     from ..reporting.charts import generate_report_charts
 
-    generate_report_charts(run_dir, brief_final)
+    generate_report_charts(run_dir, brief_final, report_config=eff_rc)
 
     llm_sentiment_md = ""
     sentiment_llm_record: dict[str, Any] = {

backend/pipeline/reporting/charts.py

@@ -256,8 +256,17 @@ def _cleanup_obsolete_report_assets(out_dir: Path) -> None:
                 pass
 
 
-def generate_report_charts(run_dir: Path, brief: dict[str, Any]) -> list[str]:
+def generate_report_charts(
-    """生成扇形/条形 PNG。返回已写入的文件名列表（不含路径）。"""
+    run_dir: Path,
+    brief: dict[str, Any],
+    *,
+    report_config: dict[str, Any] | None = None,
+) -> list[str]:
+    """生成扇形/条形 PNG。返回已写入的文件名列表（不含路径）。
+
+    若 ``report_config["chapter8_text_mining_probe"]`` 为真，**不**生成 ``chart_focus_and_scenarios_bar__*.png``
+   （与竞品报告 §8.3 文本挖掘探针互斥，避免无效产出）。
+    """
     _setup_matplotlib_cjk()
     import matplotlib.pyplot as plt
     from matplotlib.ticker import FuncFormatter
@@ -543,76 +552,87 @@ def generate_report_charts(run_dir: Path, brief: dict[str, Any]) -> list[str]:
             core = "group"
         return f"i{index:02d}_{core}"
 
-    scen_by_slug: dict[str, tuple[list[str], list[float], int]] = {}
+    _skip_focus_scenario_combo = bool(
-    by_grp = brief.get("usage_scenarios_by_matrix_group") or []
+        isinstance(report_config, dict)
-    if isinstance(by_grp, list):
+        and report_config.get("chapter8_text_mining_probe")
-        for item in by_grp:
+    )
-            if not isinstance(item, dict):
+    if _skip_focus_scenario_combo:
-                continue
+        for fp in out_dir.glob("chart_focus_and_scenarios_bar__*.png"):
-            slug = (item.get("chart_slug") or "").strip()
+            try:
-            gname = str(item.get("group") or "").strip()[:24]
+                fp.unlink()
-            idx = item.get("matrix_group_index")
+            except OSError:
-            if not slug and gname != "" and isinstance(idx, int):
+                pass
-                slug = scenario_group_asset_slug(gname, idx)
+    if not _skip_focus_scenario_combo:
-            if not slug:
+        scen_by_slug: dict[str, tuple[list[str], list[float], int]] = {}
-                continue
+        by_grp = brief.get("usage_scenarios_by_matrix_group") or []
-            scen_rows = item.get("scenarios") or []
+        if isinstance(by_grp, list):
-            n_unit = int(item.get("effective_text_units") or 0)
+            for item in by_grp:
-            gpairs: list[tuple[str, float]] = []
+                if not isinstance(item, dict):
-            if isinstance(scen_rows, list):
+                    continue
-                for r in scen_rows:
+                slug = (item.get("chart_slug") or "").strip()
-                    if not isinstance(r, dict):
+                gname = str(item.get("group") or "").strip()[:24]
-                        continue
+                idx = item.get("matrix_group_index")
-                    lb = str(r.get("scenario") or "").strip()[:48]
+                if not slug and gname != "" and isinstance(idx, int):
-                    c = r.get("count")
+                    slug = scenario_group_asset_slug(gname, idx)
-                    if lb and isinstance(c, (int, float)) and c > 0:
+                if not slug:
-                        gpairs.append((lb, float(c)))
+                    continue
-            gpairs = _merge_labeled_counts_tail(gpairs, max_items=14)
+                scen_rows = item.get("scenarios") or []
-            if gpairs and n_unit > 0:
+                n_unit = int(item.get("effective_text_units") or 0)
-                scen_by_slug[slug] = (
+                gpairs: list[tuple[str, float]] = []
-                    [p[0] for p in gpairs],
+                if isinstance(scen_rows, list):
-                    [p[1] for p in gpairs],
+                    for r in scen_rows:
-                    n_unit,
+                        if not isinstance(r, dict):
-                )
+                            continue
+                        lb = str(r.get("scenario") or "").strip()[:48]
+                        c = r.get("count")
+                        if lb and isinstance(c, (int, float)) and c > 0:
+                            gpairs.append((lb, float(c)))
+                gpairs = _merge_labeled_counts_tail(gpairs, max_items=14)
+                if gpairs and n_unit > 0:
+                    scen_by_slug[slug] = (
+                        [p[0] for p in gpairs],
+                        [p[1] for p in gpairs],
+                        n_unit,
+                    )
 
-    fb = brief.get("consumer_feedback_by_matrix_group") or []
+        fb = brief.get("consumer_feedback_by_matrix_group") or []
-    if isinstance(fb, list):
+        if isinstance(fb, list):
-        for item in fb:
+            for item in fb:
-            if not isinstance(item, dict):
+                if not isinstance(item, dict):
-                continue
+                    continue
-            slug = (item.get("chart_slug") or "").strip()
+                slug = (item.get("chart_slug") or "").strip()
-            gname = str(item.get("group") or "").strip()[:24]
+                gname = str(item.get("group") or "").strip()[:24]
-            idx = item.get("matrix_group_index")
+                idx = item.get("matrix_group_index")
-            if not slug and gname != "" and isinstance(idx, int):
+                if not slug and gname != "" and isinstance(idx, int):
-                slug = scenario_group_asset_slug(gname, idx)
+                    slug = scenario_group_asset_slug(gname, idx)
-            if not slug:
+                if not slug:
-                continue
+                    continue
-            hk = item.get("focus_keyword_hits") or []
+                hk = item.get("focus_keyword_hits") or []
-            wl: list[str] = []
+                wl: list[str] = []
-            vl: list[float] = []
+                vl: list[float] = []
-            if isinstance(hk, list):
+                if isinstance(hk, list):
-                for row in hk[:20]:
+                    for row in hk[:20]:
-                    if not isinstance(row, dict):
+                        if not isinstance(row, dict):
-                        continue
+                            continue
-                    w = str(row.get("word") or "").strip()[:32]
+                        w = str(row.get("word") or "").strip()[:32]
-                    c = row.get("count")
+                        c = row.get("count")
-                    if w and isinstance(c, (int, float)) and c > 0:
+                        if w and isinstance(c, (int, float)) and c > 0:
-                        wl.append(w)
+                            wl.append(w)
-                        vl.append(float(c))
+                            vl.append(float(c))
-            wl = wl[:18]
+                wl = wl[:18]
-            vl = vl[:18]
+                vl = vl[:18]
-            gl, gv, n_scen = scen_by_slug.get(slug, ([], [], 0))
+                gl, gv, n_scen = scen_by_slug.get(slug, ([], [], 0))
-            n_unit_fb = int(item.get("effective_comment_text_units") or 0)
+                n_unit_fb = int(item.get("effective_comment_text_units") or 0)
-            n_texts = n_scen if n_scen > 0 else n_unit_fb
+                n_texts = n_scen if n_scen > 0 else n_unit_fb
-            save_combo_focus_scenario_bar(
+                save_combo_focus_scenario_bar(
-                gname=gname,
+                    gname=gname,
-                slug=slug,
+                    slug=slug,
-                wl=wl,
+                    wl=wl,
-                vl=vl,
+                    vl=vl,
-                gl=gl,
+                    gl=gl,
-                gv=gv,
+                    gv=gv,
-                n_texts=n_texts,
+                    n_texts=n_texts,
-            )
+                )
 
     sent = brief.get("comment_sentiment_lexicon") or {}
     if isinstance(sent, dict):