在工作流中deepseek-r1的think标签内部的内容，很容易让工作流其他的llm产生幻觉，导致不能良好的生成目标效果。

        我们通过代码的方式让deepseek-r1既有think思考链的效果，又不传递思考链。

        工作流的逻辑为上图

去除think中的代码为

import re
def main(answer: str) -> dict:cleaned_text = re.sub(r'<think[^>]*>.*?</think>', '', answer, flags=re.DOTALL)# 移除清理后可能在开头的多余换行符final_text = re.sub(r'^\n+', '', cleaned_text)return {"result": final_text,}

这样子生成的效果就将think内容替换了

整个工作流的代码为

app:description: ''icon: 🤖icon_background: '#FFEAD5'mode: workflowname: outThinkuse_icon_as_answer_icon: false
dependencies:
- current_identifier: nulltype: marketplacevalue:marketplace_plugin_unique_identifier: langgenius/ollama:0.0.3@9ded90ac00e8510119a24be7396ba77191c9610d5e1e29f59d68fa1229822fc7
kind: app
version: 0.1.5
workflow:conversation_variables: []environment_variables: []features:file_upload:allowed_file_extensions:- .JPG- .JPEG- .PNG- .GIF- .WEBP- .SVGallowed_file_types:- imageallowed_file_upload_methods:- local_file- remote_urlenabled: falsefileUploadConfig:audio_file_size_limit: 50batch_count_limit: 5file_size_limit: 15image_file_size_limit: 10video_file_size_limit: 100workflow_file_upload_limit: 10image:enabled: falsenumber_limits: 3transfer_methods:- local_file- remote_urlnumber_limits: 3opening_statement: ''retriever_resource:enabled: truesensitive_word_avoidance:enabled: falsespeech_to_text:enabled: falsesuggested_questions: []suggested_questions_after_answer:enabled: falsetext_to_speech:enabled: falselanguage: ''voice: ''graph:edges:- data:isInIteration: falseisInLoop: falsesourceType: starttargetType: llmid: 1743671825415-source-1743671830921-targetsource: '1743671825415'sourceHandle: sourcetarget: '1743671830921'targetHandle: targettype: customzIndex: 0- data:isInIteration: falseisInLoop: falsesourceType: llmtargetType: codeid: 1743671830921-source-1743671844725-targetsource: '1743671830921'sourceHandle: sourcetarget: '1743671844725'targetHandle: targettype: customzIndex: 0- data:isInIteration: falseisInLoop: falsesourceType: codetargetType: endid: 1743671844725-source-1743671861862-targetsource: '1743671844725'sourceHandle: sourcetarget: '1743671861862'targetHandle: targettype: customzIndex: 0nodes:- data:desc: ''selected: falsetitle: 开始type: startvariables:- label: querymax_length: 48options: []required: truetype: text-inputvariable: queryheight: 90id: '1743671825415'position:x: 80y: 282positionAbsolute:x: 80y: 282selected: falsesourcePosition: righttargetPosition: lefttype: customwidth: 244- data:context:enabled: falsevariable_selector: []desc: ''model:completion_params:temperature: 0.7mode: chatname: deepseek-r1:14bprovider: langgenius/ollama/ollamaprompt_template:- id: bf6b98c6-75dd-4513-b87f-21d6087490efrole: systemtext: 回答问题{{#1743671825415.query#}}selected: falsetitle: LLMtype: llmvariables: []vision:enabled: falseheight: 90id: '1743671830921'position:x: 384y: 282positionAbsolute:x: 384y: 282selected: falsesourcePosition: righttargetPosition: lefttype: customwidth: 244- data:code: "import re\ndef main(answer: str) -> dict:\n    cleaned_text = re.sub(r'<think[^>]*>.*?</think>',\\ '', answer, flags=re.DOTALL)\n    # 移除清理后可能在开头的多余换行符\n    final_text =\\ re.sub(r'^\\n+', '', cleaned_text)\n    return {\n        \"result\":\\ final_text,\n    }"code_language: python3desc: ''outputs:result:children: nulltype: stringselected: falsetitle: 去除thinktype: codevariables:- value_selector:- '1743671830921'- textvariable: answerheight: 54id: '1743671844725'position:x: 688y: 282positionAbsolute:x: 688y: 282selected: truesourcePosition: righttargetPosition: lefttype: customwidth: 244- data:desc: ''outputs:- value_selector:- '1743671844725'- resultvariable: resultselected: falsetitle: 结束type: endheight: 90id: '1743671861862'position:x: 992y: 282positionAbsolute:x: 992y: 282selected: falsesourcePosition: righttargetPosition: lefttype: customwidth: 244viewport:x: -39y: 94zoom: 1