t
18c9354095
- 注入数据库、运行时配置、统一响应、文件存储和平台 AI 能力，补充业务读写工具与客户快捷操作契约。 - 移除模块内重复的组织、客户、工单、标签、技能、旧工作流、MCP 和迁移实现，将身份权限与业务主体交由宿主管理。 - 使用 libSQL 重构向量存储，并完善图片消息、访客身份、排队调度、企业微信和支持聊天页面。 - 统一 HTTP、DTO 与 WebSocket 的 snake_case 协议，补齐模块初始化、业务动作和公共载荷等回归测试。
77 lines
2.7 KiB
YAML
77 lines
2.7 KiB
YAML
# To use this file, copy it to cmd/testdata/aiconfig/ and rename it to ai_config.yaml.
|
|
#
|
|
# Configuration guide:
|
|
# 1. name: Unique identifier for the configuration. The system uses this field for deduplication and updates.
|
|
# 2. provider: Model provider. Always set this to openai because the current AI API calls use the OpenAI-compatible protocol.
|
|
# 3. baseUrl: API service endpoint.
|
|
# 4. apiKey: API key.
|
|
# 5. modelType: Model type. Must be one of the following:
|
|
# - llm: Large language model for chat and reasoning.
|
|
# - embedding: Text embedding model for generating vector representations of text.
|
|
# - rerank: Ranking model for reordering search results.
|
|
# 6. modelName: Specific model name, as defined by the provider.
|
|
# 7. dimension: Vector dimension.
|
|
# - llm and rerank: Usually set to 0.
|
|
# - embedding: Number of vector dimensions. For example, OpenAI text-embedding-3-small uses 1536.
|
|
# 8. maxContextTokens: Maximum context tokens. Use 0 for no limit.
|
|
# - llm and embedding: The maximum number of input tokens supported by the model.
|
|
# - rerank: Usually set to 0.
|
|
# 9. maxOutputTokens: Maximum output tokens.
|
|
# - llm: The maximum number of output tokens supported by the model.
|
|
# - embedding and rerank: Usually set to 0.
|
|
# 10. timeoutMs: Request timeout in milliseconds.
|
|
# 11. maxRetryCount: Number of retry attempts after a failed request.
|
|
# 12. rpmLimit: Requests per minute limit. Use 0 for no limit.
|
|
# 13. tpmLimit: Tokens per minute limit. Use 0 for no limit.
|
|
# 15. sortNo: Sort order for prioritizing models of the same type. Lower numbers have higher priority.
|
|
# 16. remark: Notes describing the purpose of this configuration or any special considerations.
|
|
#
|
|
items:
|
|
- name: LLM
|
|
provider: openai
|
|
baseUrl: https://dashscope.aliyuncs.com/compatible-mode/v1
|
|
apiKey: <REPLACE_WITH_REAL_KEY>
|
|
modelType: llm
|
|
modelName: qwen3.5-plus
|
|
dimension: 0
|
|
maxContextTokens: 128000
|
|
maxOutputTokens: 4096
|
|
timeoutMs: 30000
|
|
maxRetryCount: 1
|
|
rpmLimit: 0
|
|
tpmLimit: 0
|
|
sortNo: 10
|
|
remark: LLM
|
|
|
|
- name: embedding
|
|
provider: openai
|
|
baseUrl: https://dashscope.aliyuncs.com/compatible-mode/v1
|
|
apiKey: <REPLACE_WITH_REAL_KEY>
|
|
modelType: embedding
|
|
modelName: qwen3.7-text-embedding
|
|
dimension: 1536
|
|
maxContextTokens: 0
|
|
maxOutputTokens: 0
|
|
timeoutMs: 30000
|
|
maxRetryCount: 1
|
|
rpmLimit: 0
|
|
tpmLimit: 0
|
|
sortNo: 20
|
|
remark: embedding
|
|
|
|
- name: rerank
|
|
provider: openai
|
|
baseUrl: https://dashscope.aliyuncs.com/compatible-mode/v1
|
|
apiKey: <REPLACE_WITH_REAL_KEY>
|
|
modelType: rerank
|
|
modelName: qwen3-vl-rerank
|
|
dimension: 0
|
|
maxContextTokens: 0
|
|
maxOutputTokens: 0
|
|
timeoutMs: 30000
|
|
maxRetryCount: 1
|
|
rpmLimit: 0
|
|
tpmLimit: 0
|
|
sortNo: 30
|
|
remark: rerank
|