From faee7c879c808300ace18eb028f02c06c7d3f04e Mon Sep 17 00:00:00 2001 From: jinye_huang Date: Tue, 15 Jul 2025 11:25:58 +0800 Subject: [PATCH] =?UTF-8?q?=E4=BF=AE=E6=94=B9=E4=BA=86=E9=83=A8=E5=88=86?= =?UTF-8?q?=E9=80=89=E9=A2=98=E5=AD=97=E6=AE=B5?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- resource/prompt/generateTopics/system.txt | 24 +++++++++--------- .../__pycache__/topic_parser.cpython-312.pyc | Bin 2617 -> 2612 bytes tweet/topic_parser.py | 4 +-- 3 files changed, 14 insertions(+), 14 deletions(-) diff --git a/resource/prompt/generateTopics/system.txt b/resource/prompt/generateTopics/system.txt index d84b41b..1053062 100644 --- a/resource/prompt/generateTopics/system.txt +++ b/resource/prompt/generateTopics/system.txt @@ -15,11 +15,11 @@ - `logic`: 选定逻辑内容 (描述性文本) - `object`: 选定对象 (例如 "泰宁古城"，只能从用户提示词中提到的object列表中选择, 如果用户没有提到，则选择"None", 不能自己创造, 不能选择object列表之外的景区) - `product`: 选定产品内容 (如果没有提供单独产品，则为"None") - - `product_logic`: 选定产品的逻辑内容 (描述性文本) + - `productLogic`: 选定产品的逻辑内容 (描述性文本) - `style`: 选题风格的文件名。**必须是从 Style 文件夹中选择的完整文件名，例如 "攻略风文案"。** - - `style_logic`: 选题风格的逻辑内容 (描述性文本) - - `target_audience`: 选题面向人群的文件名。**必须是从 Demand 文件夹中选择的完整文件名，例如 "亲子向"。** - - `target_audience_logic`: 选题面向人群的逻辑内容 (描述性文本) + - `styleLogic`: 选题风格的逻辑内容 (描述性文本) + - `targetAudience`: 选题面向人群的文件名。**必须是从 Demand 文件夹中选择的完整文件名，例如 "亲子向"。** + - `targetAudienceLogic`: 选题面向人群的逻辑内容 (描述性文本) 4. 请确保生成的 JSON 数组包含用户要求的准确数量的选题对象。 5. **不要虚构景区信息、活动信息或价格**。任何活动信息以用户所给明确资料为准。 6. 选题策划时，按长假节日优先，并结合景区所在地情况（气候、习俗等）。 @@ -34,11 +34,11 @@ "logic": "暑假初期，针对家庭出游进行预热宣传", "object": "泰宁古城", "product": "...", - "product_logic": "结合住宿和导览，提供便捷的家庭游解决方案", + "productLogic": "结合住宿和导览，提供便捷的家庭游解决方案", "style": "攻略风", - "style_logic": "强调家庭共享时光和文化体验", - "target_audience": "亲子向", - "target_audience_logic": "满足家长带娃出游，寓教于乐的需求" + "styleLogic": "强调家庭共享时光和文化体验", + "targetAudience": "亲子向", + "targetAudienceLogic": "满足家长带娃出游，寓教于乐的需求" }, { "index": "2", @@ -46,11 +46,11 @@ "logic": "...", "object": "...", "product": "...", - "product_logic": "...", + "productLogic": "...", "style": "...", - "style_logic": "...", - "target_audience": "...", - "target_audience_logic": "..." + "styleLogic": "...", + "targetAudience": "...", + "targetAudienceLogic": "..." } ] ``` diff --git a/tweet/__pycache__/topic_parser.cpython-312.pyc b/tweet/__pycache__/topic_parser.cpython-312.pyc index 8d199bf34d506a605a8c42f7015de7baa169f836..3cb181c584c8717cfbf056c4181e1f8ec59aaffa 100644 GIT binary patch delta 103 zcmdlfvPFdVG%qg~0}xy|TAC5Fk@p6xJzq&;QF>~LV`)leYF=`x9Wzk&7FTgeWlpM3 petKr|EuMm+{FKt<5(rNiRk_b*O*S4TwxkZ;TN1vTtvUWN0sts2Bn1Ef delta 108 zcmdlYvQvcjG%qg~0}#wDFUd&U$a{m;g})@RC_S|#KCv_tjaD~hHgK4-HE8xIp(atH4%iNMWf9RC;r)iEXr diff --git a/tweet/topic_parser.py b/tweet/topic_parser.py index 4be3fb8..0c225ee 100644 --- a/tweet/topic_parser.py +++ b/tweet/topic_parser.py @@ -46,8 +46,8 @@ class TopicParser: # 验证每个选题是否包含所有必需的键 valid_topics = [] - required_keys = {"index", "date", "logic", "object", "product", "style", "target_audience"} - optional_keys = {"product_logic", "style_logic", "target_audience_logic"} + required_keys = {"index", "date", "logic", "object", "product", "style", "targetAudience"} + optional_keys = {"productLogic", "styleLogic", "targetAudienceLogic"} for i, item in enumerate(parsed_json): if isinstance(item, dict) and required_keys.issubset(item.keys()):