topdog/supersonic
1
0
Fork 0
mirror of https://github.com/tencentmusic/supersonic.git synced 2025-12-10 11:07:06 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

(improvement)(common) add sys parameter description (#502)

Browse Source 
Co-authored-by: jolunoluo
This commit is contained in:
LXW
2023-12-13 21:02:56 +08:00
committed by GitHub
parent 2ad0553f6c
commit 2c7afd0d55
1 changed files with 39 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

70
common/src/main/java/com/tencent/supersonic/common/pojo/SysParameter.java
View File

@@ -45,57 +45,65 @@ public class SysParameter {
public void init() {
parameters = Lists.newArrayList();
admins = Lists.newArrayList("admin");
//llm config
Parameter parameter = new Parameter("llm.model.name", "gpt3.5",
"模型名称", "list", "大语言模型相关配置");
parameter.setCandidateValues(Lists.newArrayList("gpt3.5", "gpt3.5-16k"));
parameters.add(parameter);
parameters.add(new Parameter("llm.api.key", "sk-secret",
"模型密钥", "string", "大语言模型相关配置"));
parameters.add(new Parameter("llm.temperature", "0.0",
"温度值", "number", "大语言模型相关配置"));
//s2SQL config
parameters.add(new Parameter("s2SQL.generation", "2-steps",
"S2SQL生成方式", "string", "S2SQL相关配置"));
parameters.add(new Parameter("s2SQL.linking.value.switch", "true",
"是否将linkingValues提供给大模型", "bool", "S2SQL相关配置"));
//detect config
parameters.add(new Parameter("one.detection.size", "8",
"一次探测个数", "number", "[mapper]hanlp相关配置"));
"hanlp一次探测返回结果个数", "在每次探测后, 将前后缀匹配的结果合并, 并根据相似度阈值过滤后的结果个数",
"number", "Mapper相关配置"));
parameters.add(new Parameter("one.detection.max.size", "20",
"一次探测最大个数", "number", "[mapper]hanlp相关配置"));
"hanlp一次探测前后缀匹配结果返回个数", "单次前后缀匹配返回的结果个数", "number", "Mapper相关配置"));
//mapper config
parameters.add(new Parameter("metric.dimension.min.threshold", "0.3",
"指标名、维度名最小文本相似度", "number", "[mapper]模糊匹配相关配置"));
parameters.add(new Parameter("metric.dimension.threshold", "0.3",
"指标名、维度名文本相似度", "number", "[mapper]模糊匹配相关配置"));
"指标名、维度名文本相似度阈值", "文本片段和匹配到的指标、维度名计算出来的编辑距离阈值, 若超出该阈值, 则舍弃",
"number", "Mapper相关配置"));
parameters.add(new Parameter("metric.dimension.min.threshold", "0.3",
"指标名、维度名最小文本相似度阈值",
"最小编辑距离阈值, 在FuzzyNameMapper中, 如果上面设定的编辑距离阈值的1/2大于该最小编辑距离, 则取上面设定阈值的1/2作为阈值, 否则取该阈值",
"number", "Mapper相关配置"));
parameters.add(new Parameter("dimension.value.threshold", "0.5",
"维度值最小文本相似度", "number", "[mapper]模糊匹配相关配置"));
"维度值最小文本相似度阈值", "文本片段和匹配到的维度值计算出来的编辑距离阈值, 若超出该阈值, 则舍弃",
"number", "Mapper相关配置"));
//embedding mapper config
parameters.add(new Parameter("embedding.mapper.word.min",
"4", "用于向量召回最小的文本长度", "number", "[mapper]向量召回相关配置"));
"4", "用于向量召回最小的文本长度", "为提高向量召回效率, 小于该长度的文本不进行向量语义召回", "number", "Mapper相关配置"));
parameters.add(new Parameter("embedding.mapper.word.max", "5",
"用于向量召回最大的文本长度", "number", "[mapper]向量召回相关配置"));
"用于向量召回最大的文本长度", "为提高向量召回效率, 大于该长度的文本不进行向量语义召回", "number", "Mapper相关配置"));
parameters.add(new Parameter("embedding.mapper.batch", "50",
"批量向量召回文本请求个数", "number", "[mapper]向量召回相关配置"));
"批量向量召回文本请求个数", "每次进行向量语义召回的原始文本片段个数", "number", "Mapper相关配置"));
parameters.add(new Parameter("embedding.mapper.number", "5",
"批量向量召回文本返回结果个数", "number", "[mapper]向量召回相关配置"));
"批量向量召回文本返回结果个数", "每个文本进行向量语义召回的文本结果个数", "number", "Mapper相关配置"));
parameters.add(new Parameter("embedding.mapper.distance.threshold",
"0.58", "向量召回相似度阈值", "number", "[mapper]向量召回相关配置"));
"0.58", "向量召回相似度阈值", "相似度大于该阈值的则舍弃", "number", "Mapper相关配置"));
//llm config
Parameter parameter = new Parameter("llm.model.name", "gpt3.5",
"模型名称", "","list", "Parser相关配置");
parameter.setCandidateValues(Lists.newArrayList("gpt3.5", "gpt3.5-16k"));
parameters.add(parameter);
parameters.add(new Parameter("llm.api.key", "sk-secret",
"模型密钥", "string", "Parser相关配置"));
parameters.add(new Parameter("llm.temperature", "0.0",
"温度值", "number", "Parser相关配置"));
Parameter s2SQLParameter = new Parameter("s2SQL.generation", "2_pass_auto_cot_self_consistency",
"S2SQL生成方式", "list", "Parser相关配置");
s2SQLParameter.setCandidateValues(Lists.newArrayList("1_pass_auto_cot", "1_pass_auto_cot_self_consistency",
"2_pass_auto_cot", "2_pass_auto_cot_self_consistency"));
parameters.add(s2SQLParameter);
parameters.add(new Parameter("s2SQL.linking.value.switch", "true",
"是否将Mapper探测识别到的维度值提供给大模型", "为了数据安全考虑, 这里可进行开关选择",
"bool", "Parser相关配置"));
//skip config
parameters.add(new Parameter("query.text.length.threshold", "10",
"文本长短阈值", "文本超过该阈值为长文本", "number", "是否跳过当前parser相关配置"));
"用户输入文本长短阈值", "文本超过该阈值为长文本", "number", "Parser相关配置"));
parameters.add(new Parameter("short.text.threshold", "0.5",
"短文本匹配阈值", "如果是短文本, 若query得分/文本长度>该阈值, 则跳过当前parser",
"number", "是否跳过当前parser相关配置"));
"短文本匹配阈值", "由于请求大模型耗时较长, 因此如果有规则类型的Query得分达到阈值,则跳过大模型的调用,\n如果是短文本, 若query得分/文本长度>该阈值, 则跳过当前parser",
"number", "Parser相关配置"));
parameters.add(new Parameter("long.text.threshold", "0.8",
"长文本匹配阈值", "如果是长文本, 若query得分/文本长度>该阈值, 则跳过当前parser",
"number", "是否跳过当前parser相关配置"));
"number", "Parser相关配置"));
}
}