GeneralBots/botserver
2
1
Fork 0
Code Issues Pull requests 1 Projects Releases 323 Packages Wiki Activity Actions

new(basic.gblib): New batch features.

Browse source
This commit is contained in:
Rodrigo Rodriguez 2024-08-13 21:12:58 -03:00
parent 139b28337e
commit 3b03cf4bcd
1 changed files with 26 additions and 26 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

52
packages/gpt.gblib/services/ChatServices.ts
View file

@ -64,13 +64,13 @@ import { DialogKeywords } from '../../basic.gblib/services/DialogKeywords.js';
import { GBVMService } from '../../basic.gblib/services/GBVMService.js';
import { GBLogEx } from '../../core.gbapp/services/GBLogEx.js';
import {
DEFAULT_SQL_DATABASE_PROMPT,
SQL_POSTGRES_PROMPT,
SQL_SQLITE_PROMPT,
SQL_MSSQL_PROMPT,
SQL_MYSQL_PROMPT
} from "langchain/chains/sql_db";
DEFAULT_SQL_DATABASE_PROMPT,
SQL_POSTGRES_PROMPT,
SQL_SQLITE_PROMPT,
SQL_MSSQL_PROMPT,
SQL_MYSQL_PROMPT
} from 'langchain/chains/sql_db';
export interface CustomOutputParserFields {}
export type ExpectedOutput = any;
@ -504,7 +504,7 @@ export class ChatServices {
appDataSource: dataSource
});
const prompt =
const prompt =
PromptTemplate.fromTemplate(`Based on the provided SQL table schema below, write a SQL query that would answer the user's question.
You are a SQL expert. Given an input question, first create a syntactically correct SQLite query to run, then look at the results of the query and return the answer to the input question.
Unless the user specifies in the question a specific number of examples to obtain, query for at most {top_k} results using the LIMIT clause as per SQL. You can order the results to return the most informative data in the database.
@ -525,9 +525,9 @@ export class ChatServices {
const sqlQueryChain = RunnableSequence.from([
{
schema: async () => db.getTableInfo(),
question: (input: { question: string }) => input.question ,
top_k: ()=>10,
table_info: ()=>'any'
question: (input: { question: string }) => input.question,
top_k: () => 10,
table_info: () => 'any'
},
prompt,
model,
@ -538,19 +538,19 @@ export class ChatServices {
* Create the final prompt template which is tasked with getting the natural
* language response to the SQL query.
*/
const finalResponsePrompt = SQL_SQLITE_PROMPT;
const finalResponsePrompt =
PromptTemplate.fromTemplate(`Based on the table schema below, question, SQL query, and SQL response, write a natural language response:
Optimize answers for KPI people.
------------
SCHEMA: {schema}
------------
QUESTION: {question}
------------
SQL QUERY: {query}
------------
SQL RESPONSE: {response}
------------
NATURAL LANGUAGE RESPONSE:`);
Optimize answers for KPI people. ${systemPrompt}
------------
SCHEMA: {schema}
------------
QUESTION: {question}
------------
SQL QUERY: {query}
------------
SQL RESPONSE: {response}
------------
NATURAL LANGUAGE RESPONSE:`);
/**
* Create a new RunnableSequence where we pipe the output from the previous chain, the users question,
@ -563,14 +563,15 @@ Optimize answers for KPI people.
const finalChain = RunnableSequence.from([
{
input: input => input.question,
query: sqlQueryChain,
query: sqlQueryChain
},
{
schema: async () => db.getTableInfo(),
input: input => input.question,
query: input => input.query,
response: input => db.run(input.query),
top_k: ()=>10, table_info: ()=>'any'
top_k: () => 10,
table_info: () => 'any'
},
{
result: finalResponsePrompt.pipe(model).pipe(new StringOutputParser()),
@ -582,7 +583,6 @@ Optimize answers for KPI people.
result = await finalChain.invoke({
question: question
});
} else if (LLMMode === 'nochain') {
result = await (tools.length > 0 ? modelWithTools : model).invoke(`
${systemPrompt}