MediaMetz/gemini-cli
1
0
Fork 0
mirror of https://github.com/google-gemini/gemini-cli.git synced 2026-04-26 04:54:25 -07:00
Code Issues Packages Projects Releases Wiki Activity

refactor(core): address PR feedback on web_fetch nits and refactoring

Browse Source
This commit is contained in:
Aishanee Shah
2026-03-12 19:36:49 +00:00
parent 3fa0b81f87
commit 37880ef597
2 changed files with 62 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/core/src/telemetry/types.ts
+7 -2
View File
@@ -2129,12 +2129,17 @@ export class RecoveryAttemptEvent extends BaseAgentEvent {
export const EVENT_WEB_FETCH_FALLBACK_ATTEMPT =
'gemini_cli.web_fetch_fallback_attempt';
export type WebFetchFallbackReason =
| 'private_ip'
| 'primary_failed'
| 'private_ip_skipped';
export class WebFetchFallbackAttemptEvent implements BaseTelemetryEvent {
'event.name': 'web_fetch_fallback_attempt';
'event.timestamp': string;
reason: 'private_ip' | 'primary_failed' | 'private_ip_skipped';
reason: WebFetchFallbackReason;
constructor(reason: 'private_ip' | 'primary_failed' | 'private_ip_skipped') {
constructor(reason: WebFetchFallbackReason) {
this['event.name'] = 'web_fetch_fallback_attempt';
this['event.timestamp'] = new Date().toISOString();
this.reason = reason;
packages/core/src/tools/web-fetch.ts
+55 -44
View File
@@ -171,6 +171,10 @@ interface GroundingChunkItem {
web?: GroundingChunkWeb;
}
function isGroundingChunkItem(item: unknown): item is GroundingChunkItem {
return typeof item === 'object' && item !== null;
}
interface GroundingSupportSegment {
startIndex: number;
endIndex: number;
@@ -182,6 +186,10 @@ interface GroundingSupportItem {
groundingChunkIndices?: number[];
}
function isGroundingSupportItem(item: unknown): item is GroundingSupportItem {
return typeof item === 'object' && item !== null;
}
/**
* Parameters for the WebFetch tool
*/
@@ -320,22 +328,57 @@ class WebFetchToolInvocation extends BaseToolInvocation<
}
}
private filterAndValidateUrls(urls: string[]): {
toFetch: string[];
skipped: string[];
} {
const uniqueUrls = [...new Set(urls.map(normalizeUrl))];
const toFetch: string[] = [];
const skipped: string[] = [];
for (const url of uniqueUrls) {
if (this.isBlockedHost(url)) {
debugLogger.warn(
`[WebFetchTool] Skipped private or local host: ${url}`,
);
logWebFetchFallbackAttempt(
this.config,
new WebFetchFallbackAttemptEvent('private_ip_skipped'),
);
skipped.push(`[Blocked Host] ${url}`);
continue;
}
if (!checkRateLimit(url).allowed) {
debugLogger.warn(`[WebFetchTool] Rate limit exceeded for host: ${url}`);
skipped.push(`[Rate limit exceeded] ${url}`);
continue;
}
toFetch.push(url);
}
return { toFetch, skipped };
}
private async executeFallback(
urls: string[],
signal: AbortSignal,
): Promise<ToolResult> {
const contentBudget = Math.floor(MAX_CONTENT_LENGTH / urls.length);
const uniqueUrls = [...new Set(urls)];
const contentBudget = Math.floor(
MAX_CONTENT_LENGTH / (uniqueUrls.length || 1),
);
const results: string[] = [];
for (const url of urls) {
for (const url of uniqueUrls) {
results.push(
await this.executeFallbackForUrl(url, signal, contentBudget),
);
}
const randomDelimiter = `---${Math.random().toString(36).substring(2, 15)}---`;
const aggregatedContent = results
.map((content, i) => `URL: ${urls[i]}\nContent:\n${content}`)
.join('\n\n---\n\n');
.map((content, i) => `URL: ${uniqueUrls[i]}\nContent:\n${content}`)
.join(`\n\n${randomDelimiter}\n\n`);
try {
const geminiClient = this.config.getGeminiClient();
@@ -343,9 +386,9 @@ class WebFetchToolInvocation extends BaseToolInvocation<
I was unable to access the URL(s) directly using the primary fetch tool. Instead, I have fetched the raw content of the page(s). Please use the following content to answer the request. Do not attempt to access the URL(s) again.
---
${randomDelimiter}
${aggregatedContent}
---
${randomDelimiter}
`;
const result = await geminiClient.generateContent(
{ model: 'web-fetch-fallback' },
@@ -658,30 +701,7 @@ Response: ${truncateString(rawResponseText, 10000, '\n\n... [Error response trun
const userPrompt = this.params.prompt!;
const { validUrls } = parsePrompt(userPrompt);
// Unit 1: Normalization & Deduplication
const uniqueUrls = [...new Set(validUrls.map(normalizeUrl))];
const toFetch: string[] = [];
const skipped: string[] = [];
for (const url of uniqueUrls) {
if (this.isBlockedHost(url)) {
debugLogger.warn(
`[WebFetchTool] Skipped private or local host: ${url}`,
);
logWebFetchFallbackAttempt(
this.config,
new WebFetchFallbackAttemptEvent('private_ip_skipped'),
);
skipped.push(`[Blocked Host] ${url}`);
continue;
}
if (!checkRateLimit(url).allowed) {
debugLogger.warn(`[WebFetchTool] Rate limit exceeded for host: ${url}`);
skipped.push(`[Rate limit exceeded] ${url}`);
continue;
}
toFetch.push(url);
}
const { toFetch, skipped } = this.filterAndValidateUrls(validUrls);
// If everything was skipped, fail early
if (toFetch.length === 0 && skipped.length > 0) {
@@ -723,13 +743,9 @@ Response: ${truncateString(rawResponseText, 10000, '\n\n... [Error response trun
}
// 1. Apply Grounding Supports (Citations)
const rawGroundingSupports = groundingMetadata?.groundingSupports;
const groundingSupports = Array.isArray(rawGroundingSupports)
? rawGroundingSupports.filter(
(item): item is GroundingSupportItem =>
typeof item === 'object' && item !== null,
)
: undefined;
const groundingSupports = groundingMetadata?.groundingSupports?.filter(
isGroundingSupportItem,
);
if (groundingSupports && groundingSupports.length > 0) {
const insertions: Array<{ index: number; marker: string }> = [];
groundingSupports.forEach((support) => {
@@ -753,13 +769,8 @@ Response: ${truncateString(rawResponseText, 10000, '\n\n... [Error response trun
}
// 2. Append Source List
const rawSources = groundingMetadata?.groundingChunks;
const sources = Array.isArray(rawSources)
? rawSources.filter(
(item): item is GroundingChunkItem =>
typeof item === 'object' && item !== null,
)
: undefined;
const sources =
groundingMetadata?.groundingChunks?.filter(isGroundingChunkItem);
if (sources && sources.length > 0) {
const sourceListFormatted: string[] = [];
sources.forEach((source, index) => {