Files
flynn/src/tools/builtin/minio-ingest.ts
T
2026-02-16 14:45:45 -08:00

219 lines
6.9 KiB
TypeScript

import { promisify } from 'node:util';
import { execFile } from 'node:child_process';
import { extname } from 'node:path';
import { mkdtempSync, rmSync } from 'node:fs';
import { join } from 'node:path';
import { tmpdir } from 'node:os';
import type { BackupConfig } from '../../config/schema.js';
import type { MemoryStore } from '../../memory/store.js';
import type { Tool, ToolResult } from '../types.js';
import { backupInternals } from '../../backup/index.js';
const execFileAsync = promisify(execFile);
type ExecRunner = (
file: string,
args: string[],
options?: { env?: NodeJS.ProcessEnv; maxBuffer?: number },
) => Promise<{ stdout: string | Buffer; stderr: string | Buffer }>;
const TEXT_EXTENSIONS = new Set([
'.txt',
'.md',
'.markdown',
'.csv',
'.tsv',
'.json',
'.jsonl',
'.yaml',
'.yml',
'.log',
'.xml',
'.html',
'.htm',
]);
const EXTRACTABLE_BINARY_EXTENSIONS = new Set([
'.pdf',
]);
export interface MinioIngestDeps {
execRunner?: ExecRunner;
now?: () => Date;
}
function isLikelyText(content: string): boolean {
return !content.includes('\u0000');
}
function isLikelyTextObject(objectKey: string): boolean {
const ext = extname(objectKey).toLowerCase();
if (!ext) {return true;}
return TEXT_EXTENSIONS.has(ext) || EXTRACTABLE_BINARY_EXTENSIONS.has(ext);
}
function isExtractableBinaryObject(objectKey: string): boolean {
return EXTRACTABLE_BINARY_EXTENSIONS.has(extname(objectKey).toLowerCase());
}
async function readObjectText(
runner: ExecRunner,
remotePath: string,
objectKey: string,
env: NodeJS.ProcessEnv,
): Promise<string> {
if (!isExtractableBinaryObject(objectKey)) {
const { stdout } = await runner('mc', ['cat', remotePath], { env, maxBuffer: 20 * 1024 * 1024 });
return toText(stdout);
}
const tempDir = mkdtempSync(join(tmpdir(), 'flynn-minio-ingest-'));
const localPath = join(tempDir, 'object.bin');
try {
await runner('mc', ['cp', remotePath, localPath], { env, maxBuffer: 20 * 1024 * 1024 });
const { stdout } = await runner('pdftotext', ['-q', localPath, '-'], { maxBuffer: 20 * 1024 * 1024 });
return toText(stdout);
} finally {
rmSync(tempDir, { recursive: true, force: true });
}
}
export const minioIngestInternals = {
isLikelyText,
isLikelyTextObject,
isExtractableBinaryObject,
readObjectText,
};
interface MinioIngestArgs {
object_key: string;
bucket?: string;
namespace?: string;
mode?: 'append' | 'replace';
max_chars?: number;
force?: boolean;
}
export function createMinioIngestTool(config: BackupConfig, store: MemoryStore, deps?: MinioIngestDeps): Tool {
return {
name: 'minio.ingest',
description: 'Read a text-like object from MinIO and ingest it into memory namespace for later retrieval/search.',
inputSchema: {
type: 'object',
properties: {
object_key: {
type: 'string',
description: 'Object key in MinIO bucket (for example: "knowledge/runbook.md")',
},
bucket: {
type: 'string',
description: 'Optional bucket override. Defaults to backup.minio.bucket.',
},
namespace: {
type: 'string',
description: 'Memory namespace to write to. Default: "global/knowledge".',
},
mode: {
type: 'string',
enum: ['append', 'replace'],
description: 'Write mode for memory namespace. Default: "append".',
},
max_chars: {
type: 'number',
description: 'Maximum characters to ingest. Default: 20000.',
},
force: {
type: 'boolean',
description: 'Ingest even if file extension/content look non-text.',
},
},
required: ['object_key'],
},
execute: async (rawArgs: unknown): Promise<ToolResult> => {
const args = rawArgs as MinioIngestArgs;
const minio = config.minio;
const objectKey = args.object_key?.trim();
const namespace = args.namespace ?? 'global/knowledge';
const mode = args.mode ?? 'append';
const maxChars = Math.max(1, Math.floor(args.max_chars ?? 20_000));
const force = args.force ?? false;
const bucket = args.bucket ?? minio.bucket;
if (!objectKey) {
return { success: false, output: '', error: 'object_key is required' };
}
if (!minio.enabled) {
return { success: false, output: '', error: 'MinIO ingestion requires backup.minio.enabled=true' };
}
if (!minio.endpoint || !minio.access_key || !minio.secret_key || !bucket) {
return {
success: false,
output: '',
error: 'Missing MinIO credentials in backup.minio (endpoint/access_key/secret_key/bucket)',
};
}
if (!force && !isLikelyTextObject(objectKey)) {
return {
success: false,
output: '',
error: `Unsupported object type for ingestion: ${objectKey}. Use force=true if you know it is text.`,
};
}
const alias = 'flynningest';
const host = backupInternals.buildMinioHost({
endpoint: minio.endpoint,
accessKey: minio.access_key,
secretKey: minio.secret_key,
secure: minio.secure,
});
const env = { ...process.env, [`MC_HOST_${alias}`]: host };
const runner = deps?.execRunner ?? (async (file: string, cmdArgs: string[], options?: { env?: NodeJS.ProcessEnv; maxBuffer?: number }) => {
return execFileAsync(file, cmdArgs, options);
});
const remotePath = `${alias}/${bucket}/${objectKey}`;
try {
const text = await readObjectText(runner, remotePath, objectKey, env);
if (!force && !isExtractableBinaryObject(objectKey) && !isLikelyText(text)) {
return {
success: false,
output: '',
error: `Object appears binary and cannot be ingested safely: ${objectKey}. Use force=true to override.`,
};
}
const trimmed = text.trim();
if (!trimmed) {
return {
success: false,
output: '',
error: `Object is empty: minio://${bucket}/${objectKey}`,
};
}
const clipped = trimmed.length > maxChars
? `${trimmed.slice(0, maxChars)}\n\n[truncated to ${maxChars} chars]`
: trimmed;
const importedAt = (deps?.now ? deps.now() : new Date()).toISOString();
const payload = `## MinIO Import\nsource: minio://${bucket}/${objectKey}\nimported_at: ${importedAt}\n\n${clipped}`;
store.write(namespace, payload, mode);
return {
success: true,
output: `Ingested MinIO object into memory.\nSource: minio://${bucket}/${objectKey}\nNamespace: ${namespace}\nMode: ${mode}`,
};
} catch (error) {
return {
success: false,
output: '',
error: error instanceof Error ? error.message : String(error),
};
}
},
};
}
function toText(value: string | Buffer): string {
return typeof value === 'string' ? value : value.toString('utf-8');
}