github/node-DeepResearch
1
0
Fork 0
mirror of https://github.com/jina-ai/node-DeepResearch.git synced 2025-12-26 06:28:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: improve prompting

Browse Source
This commit is contained in:
Han Xiao 2025-02-05 12:37:23 +08:00
parent fd40c7b020
commit 7f208c5072
2 changed files with 36 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
src/agent.ts
View File

@ -111,7 +111,7 @@ function getPrompt(
allowRead: boolean = true,
allowSearch: boolean = true,
badContext?: { question: string, answer: string, evaluation: string, recap: string; blame: string; improvement: string; }[],
knowledge?: { question: string; answer: string; references: any[]}[],
knowledge?: { question: string; answer: string; references: any[] }[],
allURLs?: Record<string, string>,
beastMode?: boolean
): string {
@ -295,6 +295,7 @@ function removeHTMLtags(text: string) {
return text.replace(/<[^>]*>?/gm, '');
}
export async function getResponse(question: string, tokenBudget: number = 1_000_000,
maxBadAttempts: number = 3,
existingContext?: Partial<TrackerContext>): Promise<{ result: StepAction; context: TrackerContext }> {
@ -537,42 +538,51 @@ But then you realized you have asked them before. You decided to to think out of
const searchResults = [];
for (const query of keywordsQueries) {
console.log(`Search query: ${query}`);
let results;
if (SEARCH_PROVIDER === 'duck') {
results = await duckSearch(query, {
safeSearch: SafeSearchType.STRICT
});
} else {
const {response} = await braveSearch(query);
await sleep(STEP_SLEEP);
results = {
results: response.web.results.map(r => ({
title: r.title,
url: r.url,
description: r.description
}))
};
switch (SEARCH_PROVIDER) {
case 'duck':
results = await duckSearch(query, {safeSearch: SafeSearchType.STRICT});
break;
case 'brave':
try {
const {response} = await braveSearch(query);
results = {
results: response.web?.results?.map(r => ({
title: r.title,
url: r.url,
description: r.description
})) || []
};
} catch (error) {
console.error('Brave search failed:', error);
results = {results: []};
}
await sleep(STEP_SLEEP)
break;
default:
results = {results: []};
}
const minResults = results.results.map(r => ({
title: r.title,
url: r.url,
description: r.description,
description: r.description
}));
for (const r of minResults) {
allURLs[r.url] = r.title;
}
Object.assign(allURLs, Object.fromEntries(
minResults.map(r => [r.url, r.title])
));
searchResults.push({query, results: minResults});
allKeywords.push(query);
}
allKnowledge.push({
question: `What do Internet say about ${thisStep.searchQuery}?`,
answer: removeHTMLtags(searchResults.map(r => r.results.map(r => r.description).join('; ')).join('; ')),
// flatten into one url list, and take unique urls
references: searchResults.map(r => r.results.map(r => r.url)).flat().filter((v, i, a) => a.indexOf(v) === i),
type: 'side-info'
});
question: `What do Internet say about ${thisStep.searchQuery}?`,
answer: removeHTMLtags(searchResults.map(r => r.results.map(r => r.description).join('; ')).join('; ')),
// flatten into one url list, and take unique urls
references: searchResults.map(r => r.results.map(r => r.url)).flat().filter((v, i, a) => a.indexOf(v) === i),
type: 'side-info'
});
diaryContext.push(`
At step ${step}, you took the **search** action and look for external information for the question: "${currentQuestion}".

2
src/tools/dedup.ts
View File

@ -79,7 +79,7 @@ Function FilterSetA(setA, setB, threshold):
- Different title/body filters (intitle: vs inbody:)
</similarity-definition>
Now, run FilterSetA on the following:
Now with threshold set to 0.2; run FilterSetA on the following:
SetA: ${JSON.stringify(newQueries)}
SetB: ${JSON.stringify(existingQueries)}`;
}