Commit0dcffcb

authored

Merge pull requestex3ndr#5 from wrapss/maxSettings

✨feat: Add maxLines, maxTokens and temperature settings

2 parents9be091b +f7afdc8 commit0dcffcbCopy full SHA for 0dcffcb

File tree

+34

-4

lines changed

+34

-4

lines changed

Lines changed: 15 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -41,6 +41,21 @@`
`41`	`41`	`"default":"http://127.0.0.1:11434/",`
`42`	`42`	`"description":"Ollama Server Endpoint"`
`43`	`43`	`},`
	`44`	`+"inference.maxLines": {`
	`45`	`+"type":"number",`
	`46`	`+"default":16,`
	`47`	`+"description":"Max number of lines to be keep."`
	`48`	`+ },`
	`49`	`+"inference.maxTokens": {`
	`50`	`+"type":"number",`
	`51`	`+"default":256,`
	`52`	`+"description":"Max number of new tokens to be generated."`
	`53`	`+ },`
	`54`	`+"inference.temperature": {`
	`55`	`+"type":"number",`
	`56`	`+"default":0.2,`
	`57`	`+"description":"Temperature of the model. Increasing the temperature will make the model answer more creatively."`
	`58`	`+ },`
`44`	`59`	`"inference.model": {`
`45`	`60`	`"type":"string",`
`46`	`61`	`"enum": [`

Lines changed: 6 additions & 3 deletions

Original file line number	Diff line number	Diff line change
`@@ -8,6 +8,9 @@ export async function autocomplete(args: {`
`8`	`8`	`model:string,`
`9`	`9`	`prefix:string,`
`10`	`10`	`suffix:string,`
	`11`	`+maxLines:number,`
	`12`	`+maxTokens:number,`
	`13`	`+temperature:number,`
`11`	`14`	`canceled?:()=>boolean,`
`12`	`15`	`}):Promise<string>{`
`13`	`16`
`@@ -17,7 +20,8 @@ export async function autocomplete(args: {`
`17`	`20`	`prompt:adaptPrompt({prefix:args.prefix,suffix:args.suffix,model:args.model}),`
`18`	`21`	`raw:true,`
`19`	`22`	`options:{`
`20`		`-num_predict:256`
	`23`	`+num_predict:args.maxTokens,`
	`24`	`+temperature:args.temperature`
`21`	`25`	`}`
`22`	`26`	`};`
`23`	`27`
`@@ -75,9 +79,8 @@ export async function autocomplete(args: {`
`75`	`79`
`76`	`80`	`// Update total lines`
`77`	`81`	`totalLines+=countSymbol(tokens.response,'\n');`
`78`		`-`
`79`	`82`	`// Break if too many lines and on top level`
`80`		`-if(totalLines>16&&blockStack.length===0){`
	`83`	`+if(totalLines>args.maxLines&&blockStack.length===0){`
`81`	`84`	`info('Too many lines, breaking.');`
`82`	`85`	`break;`
`83`	`86`	`}`

Lines changed: 6 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -65,6 +65,9 @@ export class PromptProvider implements vscode.InlineCompletionItemProvider {`
`65`	`65`	`letconfig=vscode.workspace.getConfiguration('inference');`
`66`	`66`	`letendpoint=config.get('endpoint')asstring;`
`67`	`67`	`letmodel=config.get('model')asstring;`
	`68`	`+letmaxLines=config.get('maxLines')asnumber;`
	`69`	`+letmaxTokens=config.get('maxTokens')asnumber;`
	`70`	`+lettemperature=config.get('temperature')asnumber;`
`68`	`71`	`if(endpoint.endsWith('/')){`
`69`	`72`	`endpoint=endpoint.slice(0,endpoint.length-1);`
`70`	`73`	`}`
`@@ -98,6 +101,9 @@ export class PromptProvider implements vscode.InlineCompletionItemProvider {`
`98`	`101`	`suffix:prepared.suffix,`
`99`	`102`	`endpoint:endpoint,`
`100`	`103`	`model:model,`
	`104`	`+maxLines:maxLines,`
	`105`	`+maxTokens:maxTokens,`
	`106`	`+ temperature,`
`101`	`107`	`canceled:()=>token.isCancellationRequested,`
`102`	`108`	`});`
`103`	`109`	info(`AI completion completed:${res}`);

Lines changed: 7 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -17,12 +17,18 @@ suite('Extension Test Suite', () => {`
`17`	`17`	`test('should perform autocomplete',async()=>{`
`18`	`18`	`letendpoint='http://127.0.0.1:11434';`
`19`	`19`	`letmodel='codellama:7b-code-q4_K_S';// Lightweight llm for tests`
	`20`	`+letmaxLines=16;`
	`21`	`+letmaxTokens=256;`
	`22`	`+lettemperature=0.2;`
`20`	`23`	`letprompt='fun main(): ';`
`21`	`24`	`letresult=awaitautocomplete({`
`22`	`25`	`endpoint,`
`23`	`26`	`model,`
`24`	`27`	`prefix:prompt,`
`25`		`-suffix:''`
	`28`	`+suffix:'',`
	`29`	`+maxLines,`
	`30`	`+maxTokens,`
	`31`	`+temperature`
`26`	`32`	`});`
`27`	`33`	`console.warn(result);`
`28`	`34`	`});`

Comments

(0)