Files
claudecodeui/server/modules/providers/list/gemini/gemini-session-synchronizer.provider.ts
Haileyesus f289ce8419 fix(gemini): align headless session/auth flow with async CLI behavior
Why:
- Gemini does not expose a new session id synchronously.
  It emits the canonical id in the init stream event.
- Creating temporary ids in web mode introduced identity drift,
  extra mapping logic, and harder resume/debug behavior.
- Headless server runs often miss shell-inherited auth vars,
  while users configure Gemini through user-level env files.
- Gemini mirrors session artifacts across folders,
  which caused duplicate sync events and duplicate session rows.

What changed:
- Removed temporary Gemini ids for new sessions.
- New Gemini sessions are now created only after init provides session_id.
- Persisted cliSessionId from the discovered canonical id,
  keeping one identifier across stream, storage, and resume.
- Built Gemini spawn env from process env plus user-level fallback files:
  ~/.gemini/.env then ~/.env, honoring GEMINI_CLI_HOME.
- Added --skip-trust for headless runs,
  because web flows cannot answer interactive trust prompts.
- Improved terminal error mapping and rejection reasons,
  especially for auth exit code 41 with actionable context.
- Limited Gemini synchronization to tmp JSONL chat artifacts,
  and disabled duplicate watcher/index paths that mirror the same sessions.
- Added gemini-2.5-flash-lite to shared model constants.

Result:
- Gemini session identity is canonical and provider-consistent.
- Headless auth now matches practical Gemini CLI configuration patterns.
- Duplicate Gemini session indexing is reduced at the source.
- Operators get clearer, actionable failure messages.
2026-05-07 12:50:33 +03:00

406 lines
11 KiB
TypeScript

import crypto from 'node:crypto';
import os from 'node:os';
import path from 'node:path';
import { readFile } from 'node:fs/promises';
import { projectsDb, sessionsDb } from '@/modules/database/index.js';
import {
findFilesRecursivelyCreatedAfter,
normalizeProjectPath,
normalizeSessionName,
readFileTimestamps,
} from '@/shared/utils.js';
import type { IProviderSessionSynchronizer } from '@/shared/interfaces.js';
import type { AnyRecord } from '@/shared/types.js';
type ParsedSession = {
sessionId: string;
projectPath: string;
sessionName?: string;
};
type GeminiJsonlMetadata = {
sessionId: string;
projectPath?: string;
projectHash?: string;
firstUserMessage?: string;
};
/**
* Session indexer for Gemini transcript artifacts.
*/
export class GeminiSessionSynchronizer implements IProviderSessionSynchronizer {
private readonly provider = 'gemini' as const;
private readonly geminiHome = path.join(os.homedir(), '.gemini');
/**
* Scans Gemini legacy JSON and new JSONL artifacts and upserts sessions into DB.
*/
async synchronize(since?: Date): Promise<number> {
const projectHashLookup = this.buildProjectHashLookup();
// const legacySessionFiles = await findFilesRecursivelyCreatedAfter(
// path.join(this.geminiHome, 'sessions'),
// '.json',
// since ?? null
// );
// Gemini creates overlapping artifacts across `sessions/` and `tmp/`.
// We currently index only `tmp/*/chats/*.jsonl` because those files are the
// live transcript source and avoid duplicate session rows from mirrored files.
// const legacyTempFiles = await findFilesRecursivelyCreatedAfter(
// path.join(this.geminiHome, 'tmp'),
// '.json',
// since ?? null
// );
// const jsonlSessionFiles = await findFilesRecursivelyCreatedAfter(
// path.join(this.geminiHome, 'sessions'),
// '.jsonl',
// since ?? null
// );
const jsonlTempFiles = await findFilesRecursivelyCreatedAfter(
path.join(this.geminiHome, 'tmp'),
'.jsonl',
since ?? null
);
// Current strategy: index only temp chat JSONL artifacts.
const files = [
// ...legacySessionFiles,
// Intentionally disabled to avoid duplicate indexing from mirrored
// `sessions/*.json` and `sessions/*.jsonl` artifacts.
// ...legacyTempFiles,
// ...jsonlSessionFiles,
...jsonlTempFiles,
];
let processed = 0;
for (const filePath of files) {
if (this.shouldSkipTempArtifact(filePath)) {
continue;
}
const parsed = filePath.endsWith('.jsonl')
? await this.processJsonlSessionFile(filePath, projectHashLookup)
: await this.processLegacySessionFile(filePath);
if (!parsed) {
continue;
}
const timestamps = await readFileTimestamps(filePath);
sessionsDb.createSession(
parsed.sessionId,
this.provider,
parsed.projectPath,
parsed.sessionName,
timestamps.createdAt,
timestamps.updatedAt,
filePath
);
processed += 1;
}
return processed;
}
/**
* Parses and upserts one Gemini legacy JSON or JSONL artifact.
*/
async synchronizeFile(filePath: string): Promise<string | null> {
if (!filePath.endsWith('.json') && !filePath.endsWith('.jsonl')) {
return null;
}
if (this.shouldSkipTempArtifact(filePath)) {
return null;
}
const parsed = filePath.endsWith('.jsonl')
? await this.processJsonlSessionFile(filePath, this.buildProjectHashLookup())
: await this.processLegacySessionFile(filePath);
if (!parsed) {
return null;
}
const timestamps = await readFileTimestamps(filePath);
return sessionsDb.createSession(
parsed.sessionId,
this.provider,
parsed.projectPath,
parsed.sessionName,
timestamps.createdAt,
timestamps.updatedAt,
filePath
);
}
/**
* Extracts session metadata from one Gemini legacy JSON artifact.
*/
private async processLegacySessionFile(filePath: string): Promise<ParsedSession | null> {
try {
const content = await readFile(filePath, 'utf8');
const data = JSON.parse(content) as AnyRecord;
const sessionId =
typeof data.sessionId === 'string'
? data.sessionId
: typeof data.id === 'string'
? data.id
: undefined;
if (!sessionId) {
return null;
}
const workspaceProjectPath = await this.resolveProjectPathFromChatWorkspace(filePath);
const projectPath = typeof data.projectPath === 'string' && data.projectPath.trim().length > 0
? data.projectPath
: workspaceProjectPath;
if (!projectPath) {
return null;
}
const messages = Array.isArray(data.messages) ? data.messages : [];
const firstMessage = messages[0] as AnyRecord | undefined;
let rawName: string | undefined;
if (Array.isArray(firstMessage?.content) && typeof firstMessage.content[0]?.text === 'string') {
rawName = firstMessage.content[0].text;
} else if (typeof firstMessage?.content === 'string') {
rawName = firstMessage.content;
}
return {
sessionId,
projectPath,
sessionName: normalizeSessionName(rawName, 'New Gemini Chat'),
};
} catch {
return null;
}
}
/**
* Extracts session metadata from one Gemini JSONL artifact.
*/
private async processJsonlSessionFile(
filePath: string,
projectHashLookup: Map<string, string>
): Promise<ParsedSession | null> {
const metadata = await this.extractJsonlMetadata(filePath);
if (!metadata) {
return null;
}
let projectPath = typeof metadata.projectPath === 'string' ? metadata.projectPath.trim() : '';
if (!projectPath) {
const workspaceProjectPath = await this.resolveProjectPathFromChatWorkspace(filePath);
if (workspaceProjectPath) {
projectPath = workspaceProjectPath;
}
}
if (!projectPath && typeof metadata.projectHash === 'string') {
projectPath = projectHashLookup.get(metadata.projectHash.trim().toLowerCase()) ?? '';
}
if (!projectPath) {
return null;
}
// Once we resolve a project hash/path pair, keep it in-memory for this sync run.
if (typeof metadata.projectHash === 'string' && metadata.projectHash.trim()) {
projectHashLookup.set(metadata.projectHash.trim().toLowerCase(), projectPath);
}
return {
sessionId: metadata.sessionId,
projectPath,
sessionName: normalizeSessionName(metadata.firstUserMessage, 'New Gemini Chat'),
};
}
/**
* Reads first useful metadata from Gemini JSONL files.
*/
private async extractJsonlMetadata(filePath: string): Promise<GeminiJsonlMetadata | null> {
try {
const content = await readFile(filePath, 'utf8');
const lines = content.split('\n');
let sessionId: string | undefined;
let projectPath: string | undefined;
let projectHash: string | undefined;
let firstUserMessage: string | undefined;
for (const line of lines) {
const trimmed = line.trim();
if (!trimmed) {
continue;
}
let parsed: AnyRecord;
try {
parsed = JSON.parse(trimmed) as AnyRecord;
} catch {
continue;
}
if (!sessionId && typeof parsed.sessionId === 'string') {
sessionId = parsed.sessionId;
}
if (!projectPath && typeof parsed.projectPath === 'string') {
projectPath = parsed.projectPath;
}
if (!projectHash && typeof parsed.projectHash === 'string') {
projectHash = parsed.projectHash;
}
if (!firstUserMessage && parsed.type === 'user') {
firstUserMessage = this.extractGeminiTextContent(parsed.content);
}
if (sessionId && (projectPath || projectHash) && firstUserMessage) {
break;
}
}
if (!sessionId) {
return null;
}
return {
sessionId,
projectPath,
projectHash,
firstUserMessage,
};
} catch {
return null;
}
}
/**
* Tries to resolve project root from Gemini tmp chat workspaces.
*/
private async resolveProjectPathFromChatWorkspace(filePath: string): Promise<string> {
if (!filePath.includes(`${path.sep}chats${path.sep}`)) {
return '';
}
const chatsDir = path.dirname(filePath);
const workspaceDir = path.dirname(chatsDir);
const projectRootPath = path.join(workspaceDir, '.project_root');
try {
const rootContent = await readFile(projectRootPath, 'utf8');
return rootContent.trim();
} catch {
return '';
}
}
/**
* Builds a hash->path lookup for Gemini JSONL metadata that stores projectHash.
*/
private buildProjectHashLookup(): Map<string, string> {
const lookup = new Map<string, string>();
const knownPaths = new Set<string>();
for (const project of projectsDb.getProjectPaths()) {
if (typeof project.project_path === 'string' && project.project_path.trim()) {
knownPaths.add(project.project_path.trim());
}
}
for (const session of sessionsDb.getAllSessions()) {
if (session.provider === this.provider && typeof session.project_path === 'string' && session.project_path.trim()) {
knownPaths.add(session.project_path.trim());
}
}
for (const knownPath of knownPaths) {
this.addProjectHashCandidates(lookup, knownPath);
}
return lookup;
}
/**
* Adds likely Gemini hash variants for one project path.
*/
private addProjectHashCandidates(lookup: Map<string, string>, projectPath: string): void {
const trimmed = projectPath.trim();
if (!trimmed) {
return;
}
const normalized = normalizeProjectPath(trimmed);
const resolved = path.resolve(trimmed);
const resolvedNormalized = normalizeProjectPath(resolved);
const candidates = new Set<string>([
trimmed,
normalized,
resolved,
resolvedNormalized,
]);
if (process.platform === 'win32') {
for (const candidate of [...candidates]) {
candidates.add(candidate.toLowerCase());
}
}
for (const candidate of candidates) {
if (!candidate) {
continue;
}
const hash = this.sha256(candidate);
if (!lookup.has(hash)) {
lookup.set(hash, trimmed);
}
}
}
/**
* Returns first user text from Gemini content payload shapes.
*/
private extractGeminiTextContent(content: unknown): string | undefined {
if (typeof content === 'string' && content.trim().length > 0) {
return content;
}
if (!Array.isArray(content)) {
return undefined;
}
for (const part of content) {
if (typeof part === 'string' && part.trim().length > 0) {
return part;
}
if (part && typeof part === 'object' && typeof (part as AnyRecord).text === 'string') {
const text = (part as AnyRecord).text;
if (text.trim().length > 0) {
return text;
}
}
}
return undefined;
}
/**
* Keeps tmp scanning scoped to chat artifacts only.
*/
private shouldSkipTempArtifact(filePath: string): boolean {
return (
filePath.startsWith(path.join(this.geminiHome, 'tmp'))
&& !filePath.includes(`${path.sep}chats${path.sep}`)
);
}
private sha256(value: string): string {
return crypto.createHash('sha256').update(value).digest('hex');
}
}