[Bug]:创意生成无法显示结果

[ChrisMa789]

Do you need to file an issue?

• I have searched the existing issues and this bug is not already filed.
• I believe this is a legitimate bug, not just a question or feature request.

Describe the bug

在左侧输入框中输入想法后，创意生成里显示：Successfully generated 0 research ideas
日志显示如下：
INFO: 222.222.222.222:0 - "GET /api/v1/knowledge/list HTTP/1.1" 200 OK
INFO: 127.0.0.1:58758 - "GET / HTTP/1.1" 200 OK
INFO: 222.222.222.222:0 - "GET /api/v1/dashboard/recent?limit=50 HTTP/1.1" 200 OK
INFO: 222.222.222.222:0 - "GET /api/v1/chat/sessions?limit=20 HTTP/1.1" 200 OK
INFO: 222.222.222.222:0 - "GET /api/v1/solve/sessions?limit=20 HTTP/1.1" 200 OK
INFO: 222.222.222.222:0 - "GET /api/v1/notebook/list HTTP/1.1" 200 OK
INFO: 222.222.222.222:0 - "WebSocket /api/v1/ideagen/generate" [accepted]
[IdeaGen] INFO: ============================================================
[IdeaGen] INFO: WebSocket connection accepted
[IdeaGen] INFO: ============================================================
INFO: connection open
[IdeaGen] INFO: Received request: notebook_id=None, record_ids=None, direct_records_count=0
[IdeaGen] INFO: Task ID: ideagen_20260126_074104_a6c2548f
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [init] Initializing idea generation workflow...
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [extracting] Processing your research topic description...
[IdeaGen] INFO: Created virtual knowledge point from user thoughts (text-only mode)
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [knowledge_extracted] Extracted 1 knowledge points
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [filtering] Filtering 1 knowledge points (loose criteria)...
[Ideagen.idea_generation] DEBUG: Prompts loaded: idea_generation (zh)
[Ideagen.idea_generation] INFO: Calling LLM to filter 1 knowledge points...
[Ideagen.idea_generation] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping
[Ideagen.idea_generation] DEBUG: LLM Input [idea_generation:idea_generation] system=409chars, user=106chars
[Ideagen.idea_generation] DEBUG: LLM Output [idea_generation:idea_generation] response=144chars
[Ideagen.idea_generation] DEBUG: LLM response: model=deepseek-chat, duration=3.22s
[Ideagen.idea_generation] DEBUG: LLM response length: 144 chars
[Ideagen.idea_generation] ERROR: JSON decode error: Expecting value: line 1 column 1 (char 0)
[Ideagen.idea_generation] DEBUG: Raw response: ```json
{
"filtered_points": [
{
"knowledge_point": "User Research Topic",
"description": "Deeptutor你为啥不输出创意？"
}
]
}

[IdeaGen] INFO: Loose filter: 1 -> 1 knowledge points
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [filtered] Filtered to 1 knowledge points
[IdeaGen] INFO: Processing knowledge point 1/1: User Research Topic
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [exploring] Exploring research ideas for: User Research Topic (1/1)
[Ideagen.idea_generation] INFO: Generating ideas for: User Research Topic
[Ideagen.idea_generation] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping
[Ideagen.idea_generation] DEBUG: LLM Input [idea_generation:idea_generation] system=550chars, user=87chars
INFO:     127.0.0.1:49786 - "GET / HTTP/1.1" 200 OK
[Ideagen.idea_generation] DEBUG: LLM Output [idea_generation:idea_generation] response=1014chars
[Ideagen.idea_generation] DEBUG: LLM response: model=deepseek-chat, duration=23.13s
[Ideagen.idea_generation] DEBUG: LLM response length: 1014 chars
[Ideagen.idea_generation] ERROR: JSON decode error: Expecting value: line 1 column 1 (char 0)
[Ideagen.idea_generation] DEBUG: Raw response: ```json
{
  "research_ideas": [
    "研究AI助手创意生成机制的触发条件与抑制因素：系统研究影响AI助手（如Deeptutor）创意输出的内部算法逻辑（如模型架构、训练数据、提示工程）与外部交互因素（如用户提问方式、上下文、系统指令），识别导致其‘不输出创意’的具体技术瓶颈与设计约束。",
    "开发一种‘创意潜能评估与激活’框架用于教育类AI：针对教育辅导场景，设计一套评估AI助手当前创意输出水平的指标体系，并开发相应的‘创意激活’提示策略或微调方法，使类似Deeptutor的AI能在保证准确性的前提下，更主动、恰当地生成启发式问题、类比或拓展性学习路径。",
    "探索人机协作中用户引导对AI创意涌现的影响：以‘用户提问为何AI不创意’为切入点，研究不同引导策略（如提供示例、设定约束、角色扮演、分步引导）如何有效激发AI的创造性回应，并构建一个优化用户与教育AI互动以提升共创意效果的交互设计指南。",
    "基于对抗性提示测试AI助手的‘安全-创意’权衡边界：系统性地设计一系列测试提示，探究像Deeptutor这类注重安全与准确性的A...
[IdeaGen] INFO: Generated 0 research ideas
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [explored] Generated 0 research ideas for: User Research Topic
[IdeaGen] WARNING: No ideas generated, skipping
[IdeaGen] SUCCESS: Workflow complete: generated 0 ideas from 1 knowledge points
[IdeaGen] INFO: [ideagen_20260126_074104_a6c2548f] [complete] Successfully generated 0 research ideas
[Ideagen] INFO: ============================================================
[Ideagen] INFO: LLM Usage Summary for Ideagen
[Ideagen] INFO: ============================================================
[Ideagen] INFO: Model       : deepseek-chat
[Ideagen] INFO: API Calls   : 2
[Ideagen] INFO: Tokens      : 152 (Input: 113, Output: 39)
[Ideagen] INFO: Cost        : $0.000027 USD
[Ideagen] INFO: ============================================================
[IdeaGen] SUCCESS: Task ideagen_20260126_074104_a6c2548f completed
[IdeaGen] INFO: WebSocket closed
[IdeaGen] INFO: ============================================================
INFO:     connection closed


### Steps to reproduce

_No response_

### Expected Behavior

_No response_

### Related Module

Dashboard

### Configuration Used

_No response_

### Logs and screenshots

_No response_

### Additional Information

- DeepTutor Version:
- Operating System:
- Python Version:
- Node.js Version:
- Browser (if applicable):
- Related Issues:

[SuperManTouX]

同样遇到了这个问题：[Backend] [IdeaGen] INFO: ============================================================
[Backend] [IdeaGen] INFO: WebSocket connection accepted
[Backend] [IdeaGen] INFO: ============================================================
[Backend] INFO: connection open
[Backend] [IdeaGen] INFO: Received request: notebook_id=None, record_ids=None, direct_records_count=4
[Backend] [IdeaGen] INFO: Task ID: ideagen_20260126_190021_5a90da22
[Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [init] Initializing idea generation workflow...
[Backend] [IdeaGen] INFO: Using 4 direct records
[Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [extracting] Extracting knowledge points from 4 records...
[Backend] [Ideagen.material_organizer] DEBUG: Prompts loaded: material_organizer (zh)
[Backend] [Ideagen.material_organizer] INFO: Processing 4 records...
[Backend] [Ideagen.material_organizer] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping
[Backend] [Ideagen.material_organizer] DEBUG: LLM Input [material_organizer:material_organizer] system=654chars, user=3002chars
[Backend] [Ideagen.material_organizer] DEBUG: LLM Output [material_organizer:material_organizer] response=848chars
[Backend] [Ideagen.material_organizer] DEBUG: LLM response: model=deepseek-chat, duration=13.45s
[Backend] [Ideagen.material_organizer] DEBUG: LLM response length: 848 chars
[Backend] [Ideagen.material_organizer] ERROR: JSON decode error: Expecting value: line 1 column 1 (char 0)
[Backend] [Ideagen.material_organizer] DEBUG: Raw response: json [Backend] { [Backend] "knowledge_points": [ [Backend] { [Backend] "knowledge_point": "Python循环中的else子句", [Backend] "description": "在Python的while和for循环中，else子句是一个可选且行为独特的特性。它的执行与否取决于循环的终止方式：只有当循环正常结束（即循环条件最终变为False）时，else块中的代码才会被执行；反之，如果循环是通过break语句提前终止的，则else块会被完全跳过。这一机制与条件判断中的else不同，它提供了一种便捷的方式来区分循环是“自然结束”还是被“强制中断”。" [Backend] }, [Backend] { [Backend] "knowledge_point": "while循环的标准执行流程", [Backend] "description": "while循环是一种基于条件的控制流结构。其标准执行流程是：首先评估循环条件（一个布尔表达式），如果条件为True，则执行循环体内的所有语句；循环体执行完毕后，程序流程会再次跳转回循环开始处，重新评估条件。这个过程会不断重复，形成一个“判断-执行”的循... [Backend] [Ideagen.material_organizer] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping [Backend] [Ideagen.material_organizer] DEBUG: LLM Input [material_organizer:material_organizer] system=248chars, user=331chars [Backend] [Ideagen.material_organizer] DEBUG: LLM Output [material_organizer:material_organizer] response=201chars [Backend] [Ideagen.material_organizer] DEBUG: LLM response: model=deepseek-chat, duration=4.01s [Backend] [IdeaGen] INFO: Extracted 1 knowledge points [Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [knowledge_extracted] Extracted 1 knowledge points [Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [filtering] Filtering 1 knowledge points (loose criteria)... [Backend] [Ideagen.idea_generation] DEBUG: Prompts loaded: idea_generation (zh) [Backend] [Ideagen.idea_generation] INFO: Calling LLM to filter 1 knowledge points... [Backend] [Ideagen.idea_generation] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping [Backend] [Ideagen.idea_generation] DEBUG: LLM Input [idea_generation:idea_generation] system=409chars, user=205chars [Backend] [Ideagen.idea_generation] DEBUG: LLM Output [idea_generation:idea_generation] response=243chars [Backend] [Ideagen.idea_generation] DEBUG: LLM response: model=deepseek-chat, duration=2.81s [Backend] [Ideagen.idea_generation] DEBUG: LLM response length: 243 chars [Backend] [Ideagen.idea_generation] ERROR: JSON decode error: Expecting value: line 1 column 1 (char 0) [Backend] [Ideagen.idea_generation] DEBUG: Raw response: json
[Backend] {
[Backend] "filtered_points": [
[Backend] {
[Backend] "knowledge_point": "Comprehensive Research Topic",
[Backend] "description": "Based on the selected 4 notebook records, multiple research directions and knowledge points can be explored."
[Backend] }
[Backend] ]
[Backend] }
[Backend] ... [Backend] [IdeaGen] INFO: Loose filter: 1 -> 1 knowledge points [Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [filtered] Filtered to 1 knowledge points [Backend] [IdeaGen] INFO: Processing knowledge point 1/1: Comprehensive Research Topic [Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [exploring] Exploring research ideas for: Comprehensive Research Topic (1/1) [Backend] [Ideagen.idea_generation] INFO: Generating ideas for: Comprehensive Research Topic [Backend] [Ideagen.idea_generation] DEBUG: response_format not supported for deepseek/deepseek-chat, skipping [Backend] [Ideagen.idea_generation] DEBUG: LLM Input [idea_generation:idea_generation] system=550chars, user=186chars [Backend] [Ideagen.idea_generation] DEBUG: LLM Output [idea_generation:idea_generation] response=753chars [Backend] [Ideagen.idea_generation] DEBUG: LLM response: model=deepseek-chat, duration=15.48s [Backend] [Ideagen.idea_generation] DEBUG: LLM response length: 753 chars [Backend] [Ideagen.idea_generation] ERROR: JSON decode error: Expecting value: line 1 column 1 (char 0) [Backend] [Ideagen.idea_generation] DEBUG: Raw response: json
[Backend] {
[Backend] "research_ideas": [
[Backend] "研究如何利用多源、异构的笔记本记录数据，构建一个自动化的研究主题发现与关联图谱生成系统，该系统能够识别潜在的研究方向、核心知识点及其之间的逻辑关系，为研究者提供宏观的科研地图。",
[Backend] "探索基于大语言模型（LLM）的智能研究助手，该助手能够深度分析研究者的个人笔记本记录（如文献笔记、实验数据、灵感碎片），主动识别知识缺口，并生成个性化的、跨学科的研究问题建议列表。",
[Backend] "开发一种量化评估模型，用于衡量从分散的笔记本记录中“涌现”出的研究主题的成熟度、创新性和可行性，该模型可结合内部一致性（笔记间的关联强度）、外部新颖性（与现有文献的对比）以及资源可行性等多个维度。",
[Backend] "研究“笔记驱动”的协作科研新模式，设计一个支持多人异步协作的知识管理平台，该平台能自动聚合、对齐并升华团队成员的个人笔记，形成团队级的“超级研究主题”，并追踪每个主题的演化路径和贡献来源。",
[Backend] "从认知科学角度出发，研究高效的研究笔记记录方法与最终研究产出质量之间的相关性，通过设计对照实验，探索何种笔记结构（如概念图...
[Backend] [IdeaGen] INFO: Generated 0 research ideas
[Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [explored] Generated 0 research ideas for: Comprehensive Research Topic
[Backend] [IdeaGen] WARNING: No ideas generated, skipping
[Backend] [IdeaGen] SUCCESS: Workflow complete: generated 0 ideas from 1 knowledge points
[Backend] [IdeaGen] INFO: [ideagen_20260126_190021_5a90da22] [complete] Successfully generated 0 research ideas
[Backend] [Ideagen] INFO: ============================================================
[Backend] [Ideagen] INFO: LLM Usage Summary for Ideagen
[Backend] [Ideagen] INFO: ============================================================
[Backend] [Ideagen] INFO: Model : deepseek-chat
[Backend] [Ideagen] INFO: API Calls : 4
[Backend] [Ideagen] INFO: Tokens : 753 (Input: 648, Output: 105)
[Backend] [Ideagen] INFO: Cost : $0.000120 USD
[Backend] [Ideagen] INFO: ============================================================
[Backend] [IdeaGen] SUCCESS: Task ideagen_20260126_190021_5a90da22 completed
[Backend] [IdeaGen] INFO: WebSocket closed
[Backend] [IdeaGen] INFO: ============================================================

[ChrisMa789]

也是用的DeepSeek吗，我换个模型就行了。

[SuperManTouX]

是的，deepseek-chat。