wangliwen
/
langchain4j-aideepin
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
langchain4j-aideepin/docs/create.sql

655 lines
30 KiB
MySQL
Raw Normal View History

增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
-- 需要先安装pgvector这个扩展（https://github.com/pgvector/pgvector）
-- CREATE EXTENSION vector;
SET client_encoding = 'UTF8';
CREATE SCHEMA public;
CREATE TABLE public.adi_ai_image
(
id bigserial primary key,
user_id bigint DEFAULT '0'::bigint NOT NULL,
uuid character varying(32) DEFAULT ''::character varying NOT NULL,
prompt character varying(1024) DEFAULT ''::character varying NOT NULL,
generate_size character varying(20) DEFAULT ''::character varying NOT NULL,
generate_number integer DEFAULT 1 NOT NULL,
original_image character varying(1000) DEFAULT ''::character varying NOT NULL,
mask_image character varying(1000) DEFAULT ''::character varying NOT NULL,
resp_images_path character varying(2048) DEFAULT ''::character varying NOT NULL,
generated_images character varying(2048) DEFAULT ''::character varying NOT NULL,
interacting_method smallint DEFAULT '1'::smallint NOT NULL,
process_status smallint DEFAULT '1'::smallint NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL,
CONSTRAINT adi_ai_image_generate_number_check CHECK (((generate_number >= 1) AND (generate_number <= 10))),
CONSTRAINT adi_ai_image_interacting_method_check CHECK ((interacting_method = ANY (ARRAY [1, 2, 3]))),
CONSTRAINT adi_ai_image_process_status_check CHECK ((process_status = ANY (ARRAY [1, 2, 3]))),
CONSTRAINT adi_ai_image_user_id_check CHECK ((user_id >= 0))
);
ALTER TABLE ONLY public.adi_ai_image
ADD CONSTRAINT udx_uuid UNIQUE (uuid);
COMMENT ON TABLE public.adi_ai_image IS 'Images generated by ai';
COMMENT ON COLUMN public.adi_ai_image.user_id IS 'The user who generated the image';
COMMENT ON COLUMN public.adi_ai_image.uuid IS 'The uuid of the request of generated images';
COMMENT ON COLUMN public.adi_ai_image.prompt IS 'The prompt for generating images';
COMMENT ON COLUMN public.adi_ai_image.generate_size IS 'The size of the generated images. Must be one of "256x256", "512x512", or "1024x1024"';
COMMENT ON COLUMN public.adi_ai_image.generate_number IS 'The number of images to generate. Must be between 1 and 10. Defaults to 1.';
COMMENT ON COLUMN public.adi_ai_image.original_image IS 'The path of the original image (local path or http path), interacting_method must be 2/3';
COMMENT ON COLUMN public.adi_ai_image.mask_image IS 'The path of the mask image (local path or http path), interacting_method must be 2';
COMMENT ON COLUMN public.adi_ai_image.resp_images_path IS 'The url of the generated images which from openai response, separated by commas';
COMMENT ON COLUMN public.adi_ai_image.generated_images IS 'The path of the generated images, separated by commas';
COMMENT ON COLUMN public.adi_ai_image.interacting_method IS '1: Creating images from scratch based on a text prompt; 2: Creating edits of an existing image based on a new text prompt; 3: Creating variations of an existing image';
COMMENT ON COLUMN public.adi_ai_image.process_status IS 'Generate image status, 1: doing, 2: fail, 3: success';
COMMENT ON COLUMN public.adi_ai_image.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_ai_image.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_ai_image.is_deleted IS 'Flag indicating whether the record is deleted (0: not deleted, 1: deleted)';
CREATE TABLE public.adi_ai_model
(
模型配置放到表中
2024-05-05 18:40:32 +08:00
id bigserial primary key,
name varchar(45) default '' not null,
type varchar(45) default 'llm' not null,
remark varchar(1000) default '',
platform varchar(45) default '' not null,
max_tokens int default 0 not null,
is_enable boolean default false NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
);
COMMENT ON TABLE public.adi_ai_model IS 'ai模型';
模型配置放到表中
2024-05-05 18:40:32 +08:00
COMMENT ON COLUMN public.adi_ai_model.type IS 'The type of the AI model,eg: text,image,embedding,rerank';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
COMMENT ON COLUMN public.adi_ai_model.name IS 'The name of the AI model';
COMMENT ON COLUMN public.adi_ai_model.remark IS 'Additional remarks about the AI model';
模型配置放到表中
2024-05-05 18:40:32 +08:00
COMMENT ON COLUMN public.adi_ai_model.platform IS 'eg: openai,dashscope,qianfan,ollama';
COMMENT ON COLUMN public.adi_ai_model.max_tokens IS 'The maximum number of tokens that can be generated';
COMMENT ON COLUMN public.adi_ai_model.is_enable IS '1: Normal usage, 0: Not available';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
COMMENT ON COLUMN public.adi_ai_model.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_ai_model.update_time IS 'Timestamp of record last update, automatically updated on each update';
模型配置放到表中
2024-05-05 18:40:32 +08:00
INSERT INTO adi_ai_model (name, type, platform, max_tokens, is_enable)
VALUES ('gpt-3.5-turbo', 'text', 'openai', 2048, false);
INSERT INTO adi_ai_model (name, type, platform, is_enable)
VALUES ('dall-e-2', 'image', 'openai', false);
INSERT INTO adi_ai_model (name, type, platform, is_enable)
VALUES ('qwen-turbo', 'text', 'dashscope', false);
INSERT INTO adi_ai_model (name, type, platform, is_enable)
VALUES ('ernie-3.5-8k-0205', 'text', 'qianfan', false);
INSERT INTO adi_ai_model (name, type, platform, is_enable)
VALUES ('tinydolphin', 'text', 'ollama', false);
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
CREATE TABLE public.adi_conversation
(
id bigserial primary key,
user_id bigint DEFAULT '0'::bigint NOT NULL,
uuid character varying(32) DEFAULT ''::character varying NOT NULL,
title character varying(45) DEFAULT ''::character varying NOT NULL,
openai_conversation_id character varying(32) DEFAULT ''::character varying NOT NULL,
tokens integer DEFAULT 0 NOT NULL,
ai_system_message character varying(1000) DEFAULT ''::character varying NOT NULL,
ai_model character varying(45) DEFAULT ''::character varying NOT NULL,
understand_context_enable boolean DEFAULT false NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
);
COMMENT ON TABLE public.adi_conversation IS '对话表';
COMMENT ON COLUMN public.adi_conversation.user_id IS '用户id';
COMMENT ON COLUMN public.adi_conversation.ai_model IS '模型名称';
COMMENT ON COLUMN public.adi_conversation.title IS '对话标题';
CREATE TABLE public.adi_conversation_message
(
id bigserial primary key,
parent_message_id bigint DEFAULT '0'::bigint NOT NULL,
conversation_id bigint DEFAULT '0'::bigint NOT NULL,
conversation_uuid character varying(32) DEFAULT ''::character varying NOT NULL,
remark text NOT NULL,
uuid character varying(32) DEFAULT ''::character varying NOT NULL,
message_role integer DEFAULT 1 NOT NULL,
tokens integer DEFAULT 0 NOT NULL,
user_id bigint DEFAULT '0'::bigint NOT NULL,
模型配置放到表中
2024-05-05 18:40:32 +08:00
ai_model_id bigint default 0 not null,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
secret_key_type integer DEFAULT 1 NOT NULL,
understand_context_msg_pair_num integer DEFAULT 0 NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
);
COMMENT ON TABLE public.adi_conversation_message IS '对话消息表';
COMMENT ON COLUMN public.adi_conversation_message.parent_message_id IS '父级消息id';
COMMENT ON COLUMN public.adi_conversation_message.conversation_id IS '对话id';
COMMENT ON COLUMN public.adi_conversation_message.conversation_uuid IS 'conversation''s uuid';
COMMENT ON COLUMN public.adi_conversation_message.remark IS 'ai回复的消息';
COMMENT ON COLUMN public.adi_conversation_message.uuid IS '唯一标识消息的UUID';
COMMENT ON COLUMN public.adi_conversation_message.message_role IS '产生该消息的角色：1: 用户, 2: 系统, 3: 助手';
COMMENT ON COLUMN public.adi_conversation_message.tokens IS '消耗的token数量';
COMMENT ON COLUMN public.adi_conversation_message.user_id IS '用户ID';
模型配置放到表中
2024-05-05 18:40:32 +08:00
COMMENT ON COLUMN public.adi_conversation_message.ai_model_id IS 'adi_ai_model id';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
COMMENT ON COLUMN public.adi_conversation_message.secret_key_type IS '加密密钥类型';
COMMENT ON COLUMN public.adi_conversation_message.understand_context_msg_pair_num IS '上下文消息对数量';
CREATE TABLE public.adi_file
(
id bigserial primary key,
name character varying(36) DEFAULT ''::character varying NOT NULL,
uuid character varying(32) DEFAULT ''::character varying NOT NULL,
ext character varying(36) DEFAULT ''::character varying NOT NULL,
user_id bigint DEFAULT '0'::bigint NOT NULL,
path character varying(250) DEFAULT ''::character varying NOT NULL,
ref_count integer DEFAULT 0 NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL,
md5 character varying(128) DEFAULT ''::character varying NOT NULL
);
COMMENT ON TABLE public.adi_file IS '文件';
COMMENT ON COLUMN public.adi_file.name IS 'File name';
COMMENT ON COLUMN public.adi_file.uuid IS 'UUID of the file';
COMMENT ON COLUMN public.adi_file.ext IS 'File extension';
COMMENT ON COLUMN public.adi_file.user_id IS '0: System; Other: User';
COMMENT ON COLUMN public.adi_file.path IS 'File path';
COMMENT ON COLUMN public.adi_file.ref_count IS 'The number of references to this file';
COMMENT ON COLUMN public.adi_file.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_file.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_file.is_deleted IS '0: Normal; 1: Deleted';
COMMENT ON COLUMN public.adi_file.md5 IS 'MD5 hash of the file';
CREATE TABLE public.adi_prompt
(
id bigserial primary key,
user_id bigint DEFAULT '0'::bigint NOT NULL,
act character varying(120) DEFAULT ''::character varying NOT NULL,
prompt text NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
);
COMMENT ON TABLE public.adi_prompt IS '提示词';
COMMENT ON COLUMN public.adi_prompt.user_id IS '所属用户(0: system)';
COMMENT ON COLUMN public.adi_prompt.act IS '提示词标题';
COMMENT ON COLUMN public.adi_prompt.prompt IS '提示词内容';
COMMENT ON COLUMN public.adi_prompt.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_prompt.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_prompt.is_deleted IS '0:未删除；1：已删除';
CREATE TABLE public.adi_sys_config
first commit
2023-08-09 20:35:59 +08:00
(
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
id bigserial primary key,
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
name character varying(100) DEFAULT ''::character varying NOT NULL,
增加支持通义千问
2024-02-27 23:55:53 +08:00
value character varying(1000) DEFAULT ''::character varying NOT NULL,
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
create_time timestamp DEFAULT localtimestamp NOT NULL,
update_time timestamp DEFAULT localtimestamp NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
);
COMMENT ON TABLE public.adi_sys_config IS '系统配置表';
COMMENT ON COLUMN public.adi_sys_config.name IS '配置项名称';
COMMENT ON COLUMN public.adi_sys_config.value IS '配置项值';
COMMENT ON COLUMN public.adi_sys_config.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_sys_config.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_sys_config.is_deleted IS '0：未删除；1：已删除';
CREATE TABLE public.adi_user
first commit
2023-08-09 20:35:59 +08:00
(
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
id bigserial primary key,
name character varying(45) DEFAULT ''::character varying NOT NULL,
password character varying(120) DEFAULT ''::character varying NOT NULL,
uuid character varying(32) DEFAULT ''::character varying NOT NULL,
email character varying(120) DEFAULT ''::character varying NOT NULL,
active_time timestamp,
user_status smallint DEFAULT '1'::smallint NOT NULL,
is_admin boolean DEFAULT false NOT NULL,
quota_by_token_daily integer DEFAULT 0 NOT NULL,
quota_by_token_monthly integer DEFAULT 0 NOT NULL,
quota_by_request_daily integer DEFAULT 0 NOT NULL,
quota_by_request_monthly integer DEFAULT 0 NOT NULL,
secret_key character varying(120) DEFAULT ''::character varying NOT NULL,
understand_context_enable smallint DEFAULT '0'::smallint NOT NULL,
update create.sql
2024-04-17 09:12:01 +08:00
understand_context_msg_pair_num integer DEFAULT 3 NOT NULL,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
quota_by_image_daily integer DEFAULT 0 NOT NULL,
quota_by_image_monthly integer DEFAULT 0 NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
);
COMMENT ON TABLE public.adi_user IS '用户表';
COMMENT ON COLUMN public.adi_user.name IS '用户名';
COMMENT ON COLUMN public.adi_user.password IS '密码';
COMMENT ON COLUMN public.adi_user.uuid IS 'UUID of the user';
COMMENT ON COLUMN public.adi_user.email IS '用户邮箱';
COMMENT ON COLUMN public.adi_user.active_time IS '激活时间';
COMMENT ON COLUMN public.adi_user.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_user.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_user.user_status IS '用户状态，1：待验证；2：正常；3：冻结';
COMMENT ON COLUMN public.adi_user.is_admin IS '是否管理员，0：否；1：是';
COMMENT ON COLUMN public.adi_user.is_deleted IS '0：未删除；1：已删除';
COMMENT ON COLUMN public.adi_user.quota_by_token_daily IS '每日token配额';
COMMENT ON COLUMN public.adi_user.quota_by_token_monthly IS '每月token配额';
COMMENT ON COLUMN public.adi_user.quota_by_request_daily IS '每日请求配额';
COMMENT ON COLUMN public.adi_user.quota_by_request_monthly IS '每月请求配额';
COMMENT ON COLUMN public.adi_user.secret_key IS '用户密钥';
COMMENT ON COLUMN public.adi_user.understand_context_enable IS '上下文理解开关';
COMMENT ON COLUMN public.adi_user.understand_context_msg_pair_num IS '上下文消息对数量';
COMMENT ON COLUMN public.adi_user.quota_by_image_daily IS '每日图片配额';
COMMENT ON COLUMN public.adi_user.quota_by_image_monthly IS '每月图片配额';
CREATE TABLE public.adi_user_day_cost
(
id bigserial primary key,
user_id bigint DEFAULT '0'::bigint NOT NULL,
day integer DEFAULT 0 NOT NULL,
requests integer DEFAULT 0 NOT NULL,
tokens integer DEFAULT 0 NOT NULL,
create_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
update_time timestamp DEFAULT CURRENT_TIMESTAMP NOT NULL,
secret_key_type integer DEFAULT 0 NOT NULL,
images_number integer DEFAULT 0 NOT NULL,
is_deleted boolean DEFAULT false NOT NULL
);
COMMENT ON TABLE public.adi_user_day_cost IS '用户每天消耗总量表';
COMMENT ON COLUMN public.adi_user_day_cost.user_id IS '用户ID';
COMMENT ON COLUMN public.adi_user_day_cost.day IS '日期，用7位整数表示，如20230901';
COMMENT ON COLUMN public.adi_user_day_cost.requests IS '请求数量';
COMMENT ON COLUMN public.adi_user_day_cost.tokens IS '消耗的token数量';
COMMENT ON COLUMN public.adi_user_day_cost.create_time IS 'Timestamp of record creation';
COMMENT ON COLUMN public.adi_user_day_cost.update_time IS 'Timestamp of record last update, automatically updated on each update';
COMMENT ON COLUMN public.adi_user_day_cost.secret_key_type IS '加密密钥类型';
COMMENT ON COLUMN public.adi_user_day_cost.images_number IS '图片数量';
-- update_time trigger
CREATE OR REPLACE FUNCTION update_modified_column()
RETURNS TRIGGER AS
$$
BEGIN
NEW.update_time = CURRENT_TIMESTAMP;
RETURN NEW;
END;
$$ language 'plpgsql';
CREATE TRIGGER trigger_ai_image_update_time
BEFORE UPDATE
ON adi_ai_image
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_ai_model_update_time
BEFORE UPDATE
ON adi_ai_model
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_conv_update_time
BEFORE UPDATE
ON adi_conversation
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_conv_message_update_time
BEFORE UPDATE
ON adi_conversation_message
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_file_update_time
BEFORE UPDATE
ON adi_file
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_prompt_update_time
BEFORE UPDATE
ON adi_prompt
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_sys_config_update_time
BEFORE UPDATE
ON adi_sys_config
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_user_update_time
BEFORE UPDATE
ON adi_user
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
CREATE TRIGGER trigger_user_day_cost_update_time
BEFORE UPDATE
ON adi_user_day_cost
FOR EACH ROW
EXECUTE PROCEDURE update_modified_column();
模型配置放到表中
2024-05-05 18:40:32 +08:00
create trigger trigger_ai_model
before update
on adi_ai_model
for each row
execute procedure update_modified_column();
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
add ollama
2024-03-14 20:19:53 +08:00
VALUES ('openai_setting', '{"secret_key":"","models":[]}');
增加支持通义千问
2024-02-27 23:55:53 +08:00
INSERT INTO adi_sys_config (name, value)
add ollama
2024-03-14 20:19:53 +08:00
VALUES ('dashscope_setting', '{"api_key":"","models":[]}');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
add ollama
2024-03-14 20:19:53 +08:00
VALUES ('qianfan_setting', '{"api_key":"","secret_key":"","models":[]}');
INSERT INTO adi_sys_config (name, value)
VALUES ('ollama_setting', '{"base_url":"","models":[]}');
feat:增加æ千帆支持文心一言
2024-03-05 00:20:04 +08:00
INSERT INTO adi_sys_config (name, value)
add ai search
2024-04-08 00:17:23 +08:00
VALUES ('google_setting',
'{"url":"https://www.googleapis.com/customsearch/v1","key":"","cx":""}');
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('request_text_rate_limit', '{"times":24,"minutes":3}');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('request_image_rate_limit', '{"times":6,"minutes":3}');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('conversation_max_num', '50');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_token_daily', '10000');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_token_monthly', '200000');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_request_daily', '150');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_request_monthly', '3000');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_image_daily', '30');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
first commit
2023-08-09 20:35:59 +08:00
VALUES ('quota_by_image_monthly', '300');
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
INSERT INTO adi_sys_config (name, value)
VALUES ('quota_by_qa_ask_daily', '50');
INSERT INTO adi_sys_config (name, value)
VALUES ('quota_by_qa_item_monthly', '100');
first commit
2023-08-09 20:35:59 +08:00
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
create table adi_knowledge_base
first commit
2023-08-09 20:35:59 +08:00
(
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
id bigserial primary key,
uuid varchar(32) default ''::character varying not null,
title varchar(250) default ''::character varying not null,
remark text default ''::character varying not null,
is_public boolean default false not null,
知识库展示列表拆分成两列
2024-03-18 00:36:03 +08:00
star_count int default 0 not null,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
owner_id bigint default 0 not null,
update create.sql
2024-04-17 09:12:01 +08:00
owner_uuid varchar(32) default ''::character varying not null,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
owner_name varchar(45) default ''::character varying not null,
create_time timestamp default CURRENT_TIMESTAMP not null,
update_time timestamp default CURRENT_TIMESTAMP not null,
is_deleted boolean default false not null
);
comment on table adi_knowledge_base is '知识库';
comment on column adi_knowledge_base.title is '知识库名称';
comment on column adi_knowledge_base.remark is '知识库描述';
comment on column adi_knowledge_base.is_public is '是否公开';
知识库展示列表拆分成两列
2024-03-18 00:36:03 +08:00
comment on column adi_knowledge_base.star_count is '点赞数';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
comment on column adi_knowledge_base.owner_id is '所属人id';
update create.sql
2024-04-17 09:12:01 +08:00
comment on column adi_knowledge_base.owner_uuid is '所属人uuid';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
comment on column adi_knowledge_base.owner_name is '所属人名称';
comment on column adi_knowledge_base.create_time is '创建时间';
comment on column adi_knowledge_base.update_time is '更新时间';
comment on column adi_knowledge_base.is_deleted is '0：未删除；1：已删除';
create trigger trigger_kb_update_time
before update
on adi_knowledge_base
for each row
execute procedure update_modified_column();
create table adi_knowledge_base_item
first commit
2023-08-09 20:35:59 +08:00
(
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
id bigserial primary key,
uuid varchar(32) default ''::character varying not null,
kb_id bigint DEFAULT '0'::bigint NOT NULL,
kb_uuid varchar(32) default ''::character varying not null,
source_file_id bigint DEFAULT '0'::bigint NOT NULL,
title varchar(250) default ''::character varying not null,
brief varchar(250) default ''::character varying not null,
remark text default ''::character varying not null,
is_embedded boolean default false not null,
create_time timestamp default CURRENT_TIMESTAMP not null,
update_time timestamp default CURRENT_TIMESTAMP not null,
is_deleted boolean default false not null
);
comment on table adi_knowledge_base_item is '知识库-条目';
comment on column adi_knowledge_base_item.kb_id is '所属知识库id';
comment on column adi_knowledge_base_item.source_file_id is '来源文件id';
comment on column adi_knowledge_base_item.title is '条目标题';
comment on column adi_knowledge_base_item.brief is '条目内容摘要';
comment on column adi_knowledge_base_item.remark is '条目内容';
comment on column adi_knowledge_base_item.is_embedded is '是否已向量化,0:否,1:是';
comment on column adi_knowledge_base_item.create_time is '创建时间';
comment on column adi_knowledge_base_item.update_time is '更新时间';
comment on column adi_knowledge_base_item.is_deleted is '0：未删除；1：已删除';
create trigger trigger_kb_item_update_time
before update
on adi_knowledge_base_item
for each row
execute procedure update_modified_column();
update create.sql
2024-04-17 09:12:01 +08:00
create table adi_knowledge_base_star_record
(
id bigserial primary key,
kb_id bigint DEFAULT '0'::bigint NOT NULL,
kb_uuid varchar(32) default ''::character varying not null,
user_id bigint default '0' NOT NULL,
user_uuid varchar(32) default ''::character varying not null,
create_time timestamp default CURRENT_TIMESTAMP not null,
update_time timestamp default CURRENT_TIMESTAMP not null,
is_deleted boolean default false not null,
UNIQUE (kb_id, user_id)
);
comment on table adi_knowledge_base_star_record is '知识库-点赞记录';
comment on column adi_knowledge_base_star_record.kb_id is 'adi_knowledge_base id';
comment on column adi_knowledge_base_star_record.kb_uuid is 'adi_knowledge_base uuid';
comment on column adi_knowledge_base_star_record.user_id is 'adi_user id';
comment on column adi_knowledge_base_star_record.user_uuid is 'adi_user uuid';
comment on column adi_knowledge_base_star_record.create_time is '创建时间';
comment on column adi_knowledge_base_star_record.update_time is '更新时间';
comment on column adi_knowledge_base_star_record.is_deleted is '0:normal; 1:deleted';
create trigger trigger_kb_star_record_update_time
before update
on adi_knowledge_base_star_record
for each row
execute procedure update_modified_column();
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
create table adi_knowledge_base_qa_record
first commit
2023-08-09 20:35:59 +08:00
(
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
id bigserial primary key,
uuid varchar(32) default ''::character varying not null,
kb_id bigint DEFAULT '0'::bigint NOT NULL,
kb_uuid varchar(32) default ''::character varying not null,
question varchar(1000) default ''::character varying not null,
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
prompt text default ''::character varying not null,
prompt_tokens integer DEFAULT 0 NOT NULL,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
answer text default ''::character varying not null,
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
answer_tokens integer DEFAULT 0 NOT NULL,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
source_file_ids varchar(500) default ''::character varying not null,
user_id bigint default '0' NOT NULL,
模型配置放到表中
2024-05-05 18:40:32 +08:00
ai_model_id bigint default 0 not null,
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
create_time timestamp default CURRENT_TIMESTAMP not null,
update_time timestamp default CURRENT_TIMESTAMP not null,
is_deleted boolean default false not null
);
comment on table adi_knowledge_base_qa_record is '知识库-提问记录';
comment on column adi_knowledge_base_qa_record.kb_id is '所属知识库id';
comment on column adi_knowledge_base_qa_record.kb_uuid is '所属知识库uuid';
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
comment on column adi_knowledge_base_qa_record.question is '用户的原始问题';
comment on column adi_knowledge_base_qa_record.prompt is '提供给LLM的提示词';
comment on column adi_knowledge_base_qa_record.prompt_tokens is '提示词消耗的token';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
comment on column adi_knowledge_base_qa_record.answer is '答案';
fix:知识库token计算问题
2024-03-12 00:41:49 +08:00
comment on column adi_knowledge_base_qa_record.answer_tokens is '答案消耗的token';
增加RAG大模型知识库
2024-02-15 14:29:57 +08:00
comment on column adi_knowledge_base_qa_record.source_file_ids is '来源文档id,以逗号隔开';
comment on column adi_knowledge_base_qa_record.user_id is '提问用户id';
comment on column adi_knowledge_base_qa_record.create_time is '创建时间';
comment on column adi_knowledge_base_qa_record.update_time is '更新时间';
comment on column adi_knowledge_base_qa_record.is_deleted is '0：未删除；1：已删除';
create trigger trigger_kb_qa_record_update_time
before update
on adi_knowledge_base_qa_record
for each row
execute procedure update_modified_column();
add ai search
2024-04-08 00:17:23 +08:00
-- ai search
create table adi_ai_search_record
(
id bigserial primary key,
uuid varchar(32) default ''::character varying not null,
question varchar(1000) default ''::character varying not null,
search_engine_response jsonb not null,
prompt text default ''::character varying not null,
prompt_tokens integer DEFAULT 0 NOT NULL,
answer text default ''::character varying not null,
answer_tokens integer DEFAULT 0 NOT NULL,
user_id bigint default '0' NOT NULL,
user_uuid varchar(32) default ''::character varying not null,
fix: model platform
2024-05-11 09:02:42 +08:00
ai_model_id bigint default 0 not null,
add ai search
2024-04-08 00:17:23 +08:00
create_time timestamp default CURRENT_TIMESTAMP not null,
update_time timestamp default CURRENT_TIMESTAMP not null,
is_deleted boolean default false not null
);
comment on table adi_ai_search_record is 'Search record';
comment on column adi_ai_search_record.question is 'User original question';
comment on column adi_ai_search_record.search_engine_response is 'Search engine''s response content';
comment on column adi_ai_search_record.prompt is 'Prompt of LLM';
comment on column adi_ai_search_record.prompt_tokens is 'prompt消耗的token数量';
comment on column adi_ai_search_record.answer is 'LLM response';
comment on column adi_ai_search_record.answer_tokens is 'LLM响应消耗的token数量';
comment on column adi_ai_search_record.user_id is 'Id from adi_user';
fix: model platform
2024-05-11 09:02:42 +08:00
COMMENT ON COLUMN adi_ai_search_record.ai_model_id IS 'adi_ai_model id';
add ai search
2024-04-08 00:17:23 +08:00
comment on column adi_ai_search_record.create_time is '创建时间';
comment on column adi_ai_search_record.update_time is '更新时间';
comment on column adi_ai_search_record.is_deleted is '0: Normal; 1: Deleted';
create trigger trigger_ai_search_record
before update
on adi_ai_search_record
for each row
execute procedure update_modified_column();