From fe39150d6278ff8fd97169f640cdbc30fac2c71a Mon Sep 17 00:00:00 2001 From: hub-gif <2487812171@qq.com> Date: Thu, 16 Apr 2026 15:36:29 +0800 Subject: [PATCH] =?UTF-8?q?refactor(demo):=20=E6=8E=A2=E9=92=88=20LLM=20?= =?UTF-8?q?=E6=94=B9=E4=B8=BA=E7=BB=86=E7=B1=BB=E5=BD=92=E7=BA=B3=EF=BC=88?= =?UTF-8?q?=E4=B8=8E=20comment=5Fgroups=20=E5=90=8C=E6=BA=90=EF=BC=89?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Made-with: Cursor --- .../demos/chapter8_text_mining_probe.py | 150 ++++++++++++------ 1 file changed, 100 insertions(+), 50 deletions(-) diff --git a/backend/pipeline/demos/chapter8_text_mining_probe.py b/backend/pipeline/demos/chapter8_text_mining_probe.py index 44f736c..98ac544 100644 --- a/backend/pipeline/demos/chapter8_text_mining_probe.py +++ b/backend/pipeline/demos/chapter8_text_mining_probe.py @@ -2,7 +2,7 @@ 第八章「文本挖掘探针」独立脚本（**不修改**主报告代码）。 流程（按细类分组）：清洗（jieba 分词 + 停用词）→ **词云图（可选）** → 词频 / TF-IDF → 共现对 → LDA 主题 -→ 规则化叙事小结 → 可选 LLM 解读（需 ``--live-llm`` 且配置好 ``AI_crawler``）。 +→ 规则化叙事小结 → 文末可选 **细类 LLM 归纳**（与正式报告 ``generate_comment_group_summaries_llm`` / ``llm_comment_group_summaries`` 同源，非 §8.2 情感）。 依赖（请自行安装）:: @@ -12,7 +12,8 @@ python -m pipeline.demos.chapter8_text_mining_probe --run-dir \"../data/JD/pipeline_runs/20260413_104252_低GI\" python -m pipeline.demos.chapter8_text_mining_probe --run-dir \"...\" --out chapter8_probe.md - python -m pipeline.demos.chapter8_text_mining_probe --run-dir \"...\" --live-llm --max-llm-groups 2 + python -m pipeline.demos.chapter8_text_mining_probe --run-dir \"...\" --live-llm + python -m pipeline.demos.chapter8_text_mining_probe --run-dir \"...\" --live-llm --llm-chunked 输出：默认写入 ``/chapter8_text_mining_probe.md``。 """ @@ -348,39 +349,70 @@ def _narrative_stub( return "\n".join(lines) -def _maybe_llm_block( - texts: list[str], - scores: list[int | None] | None, +def _effective_focus_words(run_dir: Path) -> tuple[str, ...]: + """与 ``runner.write_competitor_analysis_for_run_dir`` 一致：优先 ``effective_report_config.json``。""" + p = run_dir / "effective_report_config.json" + if p.is_file(): + try: + eff = json.loads(p.read_text(encoding="utf-8")) + except json.JSONDecodeError: + eff = None + if isinstance(eff, dict): + fw_src = eff.get("comment_focus_words") or list(jcr.COMMENT_FOCUS_WORDS) + fw_tuple = tuple( + str(x).strip() for x in fw_src if str(x).strip() + ) or jcr.COMMENT_FOCUS_WORDS + return fw_tuple + return jcr.COMMENT_FOCUS_WORDS + + +def _run_comment_groups_llm_section( + *, + run_dir: Path, keyword: str, - live: bool, + merged_rows: list[dict[str, str]], + comment_rows: list[dict[str, str]], + llm_chunked: bool, ) -> str: - if not live: + """ + 与正式报告 **§8 末** 一致：``build_comment_groups_llm_payload`` + + ``generate_comment_group_summaries_llm``（或 chunked 变体）。 + """ + sku_h = MERGED_FIELD_TO_CSV_HEADER["sku_id"] + title_h = MERGED_FIELD_TO_CSV_HEADER["title"] + fb = jcr._consumer_feedback_by_matrix_group( + merged_rows=merged_rows, + comment_rows=comment_rows, + sku_header=sku_h, + ) + fw = _effective_focus_words(run_dir) + pl = jcr.build_comment_groups_llm_payload( + feedback_groups=fb, + focus_words=fw, + merged_rows=merged_rows, + sku_header=sku_h, + title_h=title_h, + ) + if not pl: return ( - "> **LLM 解读**：未启用（请加 ``--live-llm``；需本机 ``AI_crawler`` 等与大模型调用环境可用）。" + "> **细类 LLM 归纳**：payload 为空（无可用评价分组或全部被过滤），跳过。" ) try: - from pipeline.llm.generate import generate_comment_sentiment_analysis_llm # noqa: WPS433 + if llm_chunked: + from pipeline.llm.generate import ( # noqa: WPS433 + generate_comment_group_summaries_llm_chunked, + ) + + body = generate_comment_group_summaries_llm_chunked(pl, keyword=keyword) + else: + from pipeline.llm.generate import ( # noqa: WPS433 + generate_comment_group_summaries_llm, + ) + + body = generate_comment_group_summaries_llm(pl, keyword=keyword) except Exception as e: - return f"> **LLM 解读**：导入失败：{e}" - if len(texts) < 2: - return "> **LLM 解读**：有效评论不足 2 条，跳过。" - try: - pl = jcr.build_comment_sentiment_llm_payload( - texts, - scores=scores, - max_samples_positive=10, - max_samples_negative=12, - max_samples_mixed=6, - semantic_pool_max=24, - max_chars_per_review=320, - shuffle_seed=keyword or "probe", - ) - pl["keyword"] = keyword - pl["probe_note"] = "chapter8_text_mining_probe 脚本生成，非生产流水线。" - body = generate_comment_sentiment_analysis_llm(pl) - except Exception as e: - return f"> **LLM 解读**调用失败：{e}" - return "#### LLM 深入解读（探针）\n\n" + body.strip() + return f"> **细类 LLM 归纳**调用失败：{e}" + return body.strip() def build_markdown( @@ -392,7 +424,7 @@ def build_markdown( cooc_vocab: int, cooc_pairs: int, live_llm: bool, - max_llm_groups: int, + llm_chunked: bool, wordcloud_enabled: bool, wordcloud_max: int, ) -> str: @@ -403,9 +435,6 @@ def build_markdown( comment_rows=comments, sku_header=sku_h, ) - # 与正文同序评分，便于 LLM payload - all_texts, all_scores = jcr._iter_comment_text_units_and_scores(comments, merged) - text_to_score: dict[str, int | None] = dict(zip(all_texts, all_scores)) lines: list[str] = [ "# 八、消费者反馈与用户画像（文本挖掘探针 · 实验稿）", @@ -416,7 +445,8 @@ def build_markdown( "", "## 8.0 说明", "", - "本稿为**独立探针**，流程参考「清洗 → 词云（可选）→ 词频/TF-IDF → 共现 → LDA → 叙事小结 →（可选）LLM」。" + "本稿为**独立探针**，流程参考「清洗 → 词云（可选）→ 词频/TF-IDF → 共现 → LDA → 叙事小结」；" + "文末可选 **细类 LLM 归纳**（与正式 ``llm_comment_group_summaries`` / ``generate_comment_group_summaries_llm`` 同源）。" "与线上一致的部分：**细类划分与 SKU 归因**复用 ``jd_competitor_report._consumer_feedback_by_matrix_group``；" "其余为 **jieba + sklearn** 的开放词表分析，**不替代**正式报告中的规则统计。", "", @@ -440,7 +470,6 @@ def build_markdown( ] ) - llm_used = 0 wc_n = 0 for gname, _cr_rows, texts in groups: n_raw = len([t for t in texts if (t or "").strip()]) @@ -510,15 +539,34 @@ def build_markdown( )) lines.extend(["", "---", ""]) - if live_llm and llm_used < max_llm_groups: - sub_scores = [text_to_score.get(t) for t in texts] - block = _maybe_llm_block(texts, sub_scores, kw, live=True) - lines.append(block) - lines.extend(["", "---", ""]) - llm_used += 1 - if not live_llm: - lines.append("") - lines.append(_maybe_llm_block([], None, kw, live=False)) + lines.extend( + [ + "", + "---", + "", + "## 细类评论要点归纳（大模型 · 与正式报告 §8 末同源）", + "", + "> 输入为 ``build_comment_groups_llm_payload``：关注词子串命中、有效文本行、带 SKU/品名/店铺前缀的评价摘录；" + "模型为 ``COMMENT_GROUPS_SYSTEM``（与 ``generate_comment_group_summaries_llm`` 一致）。" + "上方 jieba / TF-IDF / LDA / 词云为**独立探针**，**未**注入本段 payload。", + "", + ] + ) + if live_llm: + lines.append( + _run_comment_groups_llm_section( + run_dir=run_dir, + keyword=kw, + merged_rows=merged, + comment_rows=comments, + llm_chunked=llm_chunked, + ) + ) + else: + lines.append( + "> **细类 LLM 归纳**：未启用。请使用 ``--live-llm``（需 ``AI_crawler`` 等可用）；" + "细类很多、单次 JSON 易超上下文时可加 ``--llm-chunked``（逐细类调用后拼接）。" + ) lines.append("") lines.append("*（完）*") @@ -547,13 +595,15 @@ def main() -> None: ap.add_argument( "--live-llm", action="store_true", - help="对前若干个细类调用 §8.2 同款 LLM（需环境可用）", + help=( + "文末调用与正式管线一致的细类归纳：" + "``generate_comment_group_summaries_llm``（需 AI_crawler 等可用）" + ), ) ap.add_argument( - "--max-llm-groups", - type=int, - default=1, - help="最多对几个细类调用 LLM（避免费用与时间）", + "--llm-chunked", + action="store_true", + help="细类逐条调用 ``generate_comment_group_summaries_llm_chunked``，防上下文过长", ) ap.add_argument( "--no-wordcloud", @@ -576,7 +626,7 @@ def main() -> None: cooc_vocab=args.cooc_vocab, cooc_pairs=args.cooc_pairs, live_llm=args.live_llm, - max_llm_groups=max(0, args.max_llm_groups), + llm_chunked=args.llm_chunked, wordcloud_enabled=not args.no_wordcloud, wordcloud_max=max(0, args.wordcloud_max), )