@byted-las/contextlake-openclaw 1.0.4 → 1.0.5

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@byted-las/contextlake-openclaw",
-  "version": "1.0.4",
+  "version": "1.0.5",
   "description": "ContextLake OpenClaw Plugin for managing knowledge base",
   "main": "index.ts",
   "files": [
@@ -34,6 +34,7 @@
   },
   "dependencies": {
     "@aws-sdk/client-s3": "^3.1014.0",
+    "@aws-sdk/s3-request-presigner": "^3.1014.0",
     "@lancedb/lancedb": "^0.26.2",
     "@volcengine/tos-sdk": "^2.9.0",
     "commander": "^14.0.3",

package/src/commands/cli.ts

@@ -37,29 +37,23 @@ function parseMetadata(metadata: any): Record<string, any> {
 
 export function getCliCommands(pluginConfig: ContextLakeConfig, logger: any) {
     return {
-        connectAction: async (datasource_name: string, options: any) => {
-            logger.info(`[${new Date().toISOString()}] [ContextLake] CLI connect started`, { datasource_name, options });
+        connectAction: async (datasource_name: string, url: string, options: any) => {
+            logger.info(`[${new Date().toISOString()}] [ContextLake] CLI connect started`, { datasource_name, url, options });
             try {
                 const params: ConnectParams = {
                     datasource_name,
-                    vendor: options.vendor,
+                    url,
                     endpoint: options.endpoint,
                     access_key: options.ak,
                     secret_key: options.sk,
                     region: options.region,
-                    bucket: options.bucket,
-                    prefix: options.prefix,
                     sample_rows: parseInt(options.sampleRows),
                 };
 
                 // eslint-disable-next-line no-console
                 console.log(`[contextlake connect] Connecting to datasource "${datasource_name}"...`);
                 // eslint-disable-next-line no-console
-                console.log(`  vendor: ${params.vendor}`);
-                // eslint-disable-next-line no-console
-                console.log(`  bucket: ${params.bucket}`);
-                // eslint-disable-next-line no-console
-                console.log(`  prefix: ${params.prefix}`);
+                console.log(`  url: ${params.url}`);
 
                 const result = await connectDataSource(params);
                 // eslint-disable-next-line no-console

package/src/commands/index.ts

@@ -35,6 +35,12 @@ export function registerAll(ctx: OpenClawPluginApi, logger: PluginLogger) {
 
         ctx.registerTool(tools.writeLanceCatalogTool );
         logger.info(`[${new Date().toISOString()}] [ContextLake] Tool registered: ${tools.writeLanceCatalogTool.name}`);
+
+        ctx.registerTool(tools.readLanceCatalogTool );
+        logger.info(`[${new Date().toISOString()}] [ContextLake] Tool registered: ${tools.readLanceCatalogTool.name}`);
+
+        ctx.registerTool(tools.generatePresignedUrlTool );
+        logger.info(`[${new Date().toISOString()}] [ContextLake] Tool registered: ${tools.generatePresignedUrlTool.name}`);
         
         ctx.registerTool(tools.listDatasourceTool );
         logger.info(`[${new Date().toISOString()}] [ContextLake] Tool registered: ${tools.listDatasourceTool.name}`);
@@ -60,15 +66,12 @@ export function registerAll(ctx: OpenClawPluginApi, logger: PluginLogger) {
             const commands = getCliCommands(pluginConfig, logger);
             
             // connect -- data source profiling (las-data-profiler)
-            contextlake.command('connect <datasource_name>')
+            contextlake.command('connect <datasource_name> <url>')
                 .description('Connect to a data source and profile its structure, schemas, and media metadata into LanceDB')
-                .requiredOption('--vendor <vendor>', 'Data source type: volcengine | alibaba | tencent | aws | local')
                 .option('--endpoint <url>', 'S3 Endpoint URL (not needed for local)')
                 .option('--ak <credential_id>', 'Credential ID for the data source')
                 .option('--sk <credential_value>', 'Credential value for the data source')
                 .option('--region <region>', 'Region identifier (e.g. cn-beijing)')
-                .requiredOption('--bucket <bucket>', 'Bucket name (or local root directory for local vendor)')
-                .requiredOption('--prefix <prefix>', 'Path prefix to limit scan scope')
                 .option('--sample-rows <number>', 'Number of rows to sample per structured file', '100')
                 .action(commands.connectAction);
 

package/src/commands/slashcmd.ts

@@ -108,21 +108,15 @@ export function getSlashCommands(pluginConfig: ContextLakeConfig, logger: any) {
 
             logger.info(`[${new Date().toISOString()}] [ContextLake] Slash command profiler started`, { args });
             try {
-                if (args.length < 4) {
-                    return { text: `**Error:** Missing arguments. Usage: /contextlake-profiler <datasource_name> <vendor> <bucket> <prefix> [endpoint] [ak] [sk] [region]` };
+                if (args.length < 2) {
+                    return { text: `**Error:** Missing arguments. Usage: /contextlake-profiler <datasource_name> <url> [endpoint] [ak] [sk] [region]` };
                 }
                 
-                const [datasource_name, vendor, bucket, prefix, endpoint, access_key, secret_key, region] = args;
-                
-                if (!['volcengine', 'alibaba', 'tencent', 'aws', 'local'].includes(vendor)) {
-                     return { text: `**Error:** Invalid vendor. Must be one of: volcengine, alibaba, tencent, aws, local` };
-                }
+                const [datasource_name, url, endpoint, access_key, secret_key, region] = args;
                 
                 const params: ConnectParams = {
                     datasource_name,
-                    vendor: vendor as ConnectParams['vendor'],
-                    bucket,
-                    prefix,
+                    url,
                     endpoint,
                     access_key,
                     secret_key,

package/src/commands/tools.ts

@@ -3,8 +3,8 @@ import { retrieveAssets } from '../lib/actions/retrieve';
 import { listAssets, deleteAssets } from '../lib/actions/manage';
 import { connectDataSource, listDataSources } from '../lib/actions/profiler';
 import { getLasTools } from '../lib/actions/las-tools';
-import { listS3Objects, readS3Object } from '../lib/actions/s3-tools';
-import { writeLanceCatalog } from '../lib/actions/lance-tools';
+import { listS3Objects, readS3Object, getPresignedUrl } from '../lib/actions/s3-tools';
+import { writeLanceCatalog, readLanceCatalog } from '../lib/actions/lance-tools';
 import { ContextLakeConfig } from '../utils/config';
 // @ts-ignore
 import type { AnyAgentTool } from 'openclaw/plugin-sdk';
@@ -18,7 +18,9 @@ export function getAgentTools(pluginConfig: ContextLakeConfig, logger: any): {
     listDatasourceTool: AnyAgentTool;
     listS3ObjectsTool: AnyAgentTool;
     readS3ObjectTool: AnyAgentTool;
+    generatePresignedUrlTool: AnyAgentTool;
     writeLanceCatalogTool: AnyAgentTool;
+    readLanceCatalogTool: AnyAgentTool;
     lasTools: AnyAgentTool[];
 } {
     const lasTools = getLasTools(pluginConfig, logger);
@@ -274,16 +276,10 @@ Example User Queries:
                 type: 'object',
                 properties: {
                     datasource_name: { type: 'string', description: 'Name of the data source' },
-                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'], description: 'Data source type' },
-                    endpoint: { type: 'string', description: 'S3 Endpoint URL (not needed for local)' },
-                    access_key: { type: 'string', description: 'Credential ID for the data source' },
-                    secret_key: { type: 'string', description: 'Credential value for the data source' },
-                    region: { type: 'string', description: 'Region identifier (e.g. cn-beijing)' },
-                    bucket: { type: 'string', description: 'Bucket name (or local root directory for local vendor)' },
-                    prefix: { type: 'string', description: 'Path prefix to limit scan scope' },
+                    url: { type: 'string', description: 'Data source URL (e.g. tos://bucket/prefix, oss://..., s3://..., file:///path)' },
                     sample_rows: { type: 'integer', description: 'Number of rows to sample per structured file' }
                 },
-                required: ['datasource_name', 'vendor', 'bucket', 'prefix'],
+                required: ['datasource_name', 'url'],
                 additionalProperties: false
             },
             
@@ -317,17 +313,15 @@ Example User Queries:
             parameters: {
                 type: 'object',
                 properties: {
-                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'] },
-                    bucket: { type: 'string' },
+                    url: { type: 'string', description: 'Data source URL (e.g. tos://bucket/prefix, oss://..., file:///path)' },
+                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'], description: 'Required if url is not provided' },
+                    bucket: { type: 'string', description: 'Required if url is not provided' },
                     prefix: { type: 'string' },
                     endpoint: { type: 'string' },
-                    access_key: { type: 'string' },
-                    secret_key: { type: 'string' },
-                    region: { type: 'string' },
                     maxKeys: { type: 'integer' },
                     continuationToken: { type: 'string' }
                 },
-                required: ['vendor', 'bucket'],
+                required: [],
                 additionalProperties: false
             },
             async execute(toolCallId: string, params: any) {
@@ -347,22 +341,38 @@ Example User Queries:
             parameters: {
                 type: 'object',
                 properties: {
-                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'] },
-                    bucket: { type: 'string' },
-                    key: { type: 'string' },
+                    url: { type: 'string', description: 'Full URL to the object (e.g. tos://bucket/path/to/key.txt)' },
+                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'], description: 'Required if url is not provided' },
+                    bucket: { type: 'string', description: 'Required if url is not provided' },
+                    key: { type: 'string', description: 'Required if url is not provided' },
                     endpoint: { type: 'string' },
-                    access_key: { type: 'string' },
-                    secret_key: { type: 'string' },
-                    region: { type: 'string' },
                     maxBytes: { type: 'integer' }
                 },
-                required: ['vendor', 'bucket', 'key'],
+                required: [],
                 additionalProperties: false
             },
             async execute(toolCallId: string, params: any) {
                 let actualParams = params.params || params;
                 try {
-                    const buf = await readS3Object(actualParams, actualParams.key, actualParams.maxBytes);
+                    // Extract key from url if provided
+                    let key = actualParams.key;
+                    if (actualParams.url && !key) {
+                        try {
+                            if (actualParams.url.startsWith('file://')) {
+                                // Key is not strictly needed for file://, bucket contains the path in parseS3Url
+                                key = ''; 
+                            } else {
+                                const parsedUrl = new URL(actualParams.url);
+                                key = parsedUrl.pathname.replace(/^\//, '');
+                            }
+                        } catch (e) {
+                            // let it fail in readS3Object
+                        }
+                    }
+                    if (!key && !actualParams.url?.startsWith('file://')) {
+                        throw new Error('key is required or must be part of the url');
+                    }
+                    const buf = await readS3Object(actualParams, key, actualParams.maxBytes);
                     // Return as base64 string
                     return { content: [{ type: "text", text: buf.toString('base64') }], details: { length: buf.length } } as any;
                 } catch (e: any) {
@@ -370,6 +380,49 @@ Example User Queries:
                 }
             }
         },
+        generatePresignedUrlTool: {
+            name: 'generate-presigned-url',
+            label: 'Generate Presigned URL',
+            description: 'Generate a presigned HTTP URL for an S3/TOS object, allowing temporary public access',
+            parameters: {
+                type: 'object',
+                properties: {
+                    url: { type: 'string', description: 'Full URL to the object (e.g. tos://bucket/path/to/key.txt)' },
+                    vendor: { type: 'string', enum: ['volcengine', 'alibaba', 'tencent', 'aws', 'local'], description: 'Required if url is not provided' },
+                    bucket: { type: 'string', description: 'Required if url is not provided' },
+                    key: { type: 'string', description: 'Required if url is not provided' },
+                    endpoint: { type: 'string' },
+                    expiresIn: { type: 'integer', description: 'Expiration time in seconds (default 3600)' }
+                },
+                required: [],
+                additionalProperties: false
+            },
+            async execute(toolCallId: string, params: any) {
+                let actualParams = params.params || params;
+                try {
+                    let key = actualParams.key;
+                    if (actualParams.url && !key) {
+                        try {
+                            if (actualParams.url.startsWith('file://')) {
+                                key = ''; 
+                            } else {
+                                const parsedUrl = new URL(actualParams.url);
+                                key = parsedUrl.pathname.replace(/^\//, '');
+                            }
+                        } catch (e) {
+                            // let it fail in getPresignedUrl
+                        }
+                    }
+                    if (!key && !actualParams.url?.startsWith('file://')) {
+                        throw new Error('key is required or must be part of the url');
+                    }
+                    const url = await getPresignedUrl(actualParams, key, actualParams.expiresIn);
+                    return { content: [{ type: "text", text: url }], details: { url } } as any;
+                } catch (e: any) {
+                    return { content: [{ type: "text", text: String(e.message) }], details: { error: e.message } } as any;
+                }
+            }
+        },
         writeLanceCatalogTool: {
             name: 'write-lance-catalog',
             label: 'Write LanceDB Catalog',
@@ -393,6 +446,31 @@ Example User Queries:
                     return { content: [{ type: "text", text: String(e.message) }], details: { error: e.message } } as any;
                 }
             }
+        },
+        readLanceCatalogTool: {
+            name: 'read-lance-catalog',
+            label: 'Read LanceDB Catalog',
+            description: 'Read records from a local LanceDB table for validation or ingestion processes',
+            parameters: {
+                type: 'object',
+                properties: {
+                    db_path: { type: 'string', description: 'Path to the local LanceDB database' },
+                    table_name: { type: 'string', description: 'Name of the table to read' },
+                    limit: { type: 'integer', description: 'Maximum number of records to return' },
+                    filter: { type: 'string', description: 'SQL-like filter string (e.g., "category = \'structured\'")' }
+                },
+                required: ['db_path', 'table_name'],
+                additionalProperties: false
+            },
+            async execute(toolCallId: string, params: any) {
+                let actualParams = params.params || params;
+                try {
+                    const results = await readLanceCatalog(actualParams);
+                    return { content: [{ type: "text", text: JSON.stringify(results) }], details: { count: results.length, data: results } } as any;
+                } catch (e: any) {
+                    return { content: [{ type: "text", text: String(e.message) }], details: { error: e.message } } as any;
+                }
+            }
         }
     };
 }

package/src/lib/actions/lance-tools.ts

@@ -1,4 +1,5 @@
 import * as lancedb from '@lancedb/lancedb';
+import * as fs from 'fs';
 
 export interface LanceWriteParams {
     db_path: string;
@@ -6,6 +7,13 @@ export interface LanceWriteParams {
     records: any[];
 }
 
+export interface LanceReadParams {
+    db_path: string;
+    table_name: string;
+    limit?: number;
+    filter?: string;
+}
+
 export async function writeLanceCatalog(params: LanceWriteParams) {
     if (!params.records || params.records.length === 0) {
         return;
@@ -21,3 +29,30 @@ export async function writeLanceCatalog(params: LanceWriteParams) {
         await db.createTable(params.table_name, params.records);
     }
 }
+
+export async function readLanceCatalog(params: LanceReadParams) {
+    if (!fs.existsSync(params.db_path)) {
+        throw new Error(`Database not found at ${params.db_path}`);
+    }
+
+    const db = await lancedb.connect(params.db_path);
+    const tableNames = await db.tableNames();
+    
+    if (!tableNames.includes(params.table_name)) {
+        throw new Error(`Table '${params.table_name}' not found in database`);
+    }
+
+    const table = await db.openTable(params.table_name);
+    let query = table.query();
+    
+    if (params.filter) {
+        query = query.where(params.filter);
+    }
+    
+    if (params.limit && params.limit > 0) {
+        query = query.limit(params.limit);
+    }
+    
+    const results = await query.toArray();
+    return results;
+}

package/src/lib/actions/las-tools.ts

@@ -1,11 +1,36 @@
 import { LasApiClient } from './las-api';
 import { ContextLakeConfig } from '../../utils/config';
+import { getPresignedUrl } from './s3-tools';
 // @ts-ignore
 import type { AnyAgentTool } from 'openclaw/plugin-sdk';
 
 export function getLasTools(pluginConfig: ContextLakeConfig, logger: any): AnyAgentTool[] {
     const apiClient = new LasApiClient(pluginConfig, logger);
 
+    const processUrl = async (url: string): Promise<string> => {
+        if (!url) return url;
+        if (url.startsWith('tos://')) {
+            // LAS operators prefer native tos:// paths when supported, leave as is
+            return url;
+        } else if (url.startsWith('oss://') || url.startsWith('s3://') || url.startsWith('cos://') || url.startsWith('file://')) {
+            // Need presigned HTTP url for other vendors
+            logger.info(`[LasTools] Presigning URL for vendor: ${url}`);
+            try {
+                // If it's a file:// we also presign it to file:// which might not be supported by remote LAS, 
+                // but local files typically need to be uploaded to TOS first. We'll leave file:// to fail or be handled elsewhere.
+                if (url.startsWith('file://')) return url;
+                
+                const urlParts = new URL(url);
+                const key = urlParts.pathname.replace(/^\//, '');
+                return await getPresignedUrl({ url }, key, 3600);
+            } catch (e: any) {
+                logger.warn(`[LasTools] Failed to presign URL: ${url}`, { error: e.message });
+                return url; // fallback to original
+            }
+        }
+        return url;
+    };
+
     const callApi = async (method: string, args: any[]) => {
         try {
             // @ts-ignore
@@ -35,6 +60,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.image) {
+                    params.data.image = await processUrl(params.data.image);
+                }
                 return await callApi('process', ['las_image_resample', params.data]);
             }
         },
@@ -55,6 +83,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.input_path) {
+                    params.data.input_path = await processUrl(params.data.input_path);
+                }
                 return await callApi('process', ['las_audio_extract_and_split', params.data]);
             }
         },
@@ -73,6 +104,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.input_path) {
+                    params.data.input_path = await processUrl(params.data.input_path);
+                }
                 return await callApi('process', ['las_audio_convert', params.data]);
             }
         },
@@ -91,6 +125,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.audio?.url) {
+                    params.data.audio.url = await processUrl(params.data.audio.url);
+                }
                 return await callApi('submitAndPoll', ['las_asr_pro', params.data]);
             }
         },
@@ -107,6 +144,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.audio?.url) {
+                    params.data.audio.url = await processUrl(params.data.audio.url);
+                }
                 return await callApi('submitAndPoll', ['las_seed_2_0', params.data]);
             }
         },
@@ -130,6 +170,13 @@ Parameters:
                 required: ['model', 'input']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.input && Array.isArray(params.input)) {
+                    for (const item of params.input) {
+                        if (item.type === 'image_url' && item.image_url?.url) {
+                            item.image_url.url = await processUrl(item.image_url.url);
+                        }
+                    }
+                }
                 return await callApi('multimodalEmbedding', [
                     params.model,
                     params.input,
@@ -153,6 +200,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.video_url) {
+                    params.data.video_url = await processUrl(params.data.video_url);
+                }
                 return await callApi('submitAndPoll', ['las_long_video_understand', params.data]);
             }
         },
@@ -171,6 +221,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.url) {
+                    params.data.url = await processUrl(params.data.url);
+                }
                 return await callApi('submitAndPoll', ['las_pdf_parse_doubao', params.data]);
             }
         },
@@ -189,6 +242,9 @@ Parameters in data:
                 required: ['data']
             },
             async execute(toolCallId: string, params: any) {
+                if (params.data?.video_url) {
+                    params.data.video_url = await processUrl(params.data.video_url);
+                }
                 return await callApi('submitAndPoll', ['las_video_resize', params.data]);
             }
         }

package/src/lib/actions/profiler.ts

@@ -7,13 +7,14 @@ import * as mime from 'mime-types';
 
 export interface ConnectParams {
   datasource_name: string;
-  vendor: 'volcengine' | 'alibaba' | 'tencent' | 'aws' | 'local';
+  url: string;
+  vendor?: 'volcengine' | 'alibaba' | 'tencent' | 'aws' | 'local';
   endpoint?: string;
   access_key?: string;
   secret_key?: string;
   region?: string;
-  bucket: string;
-  prefix: string;
+  bucket?: string;
+  prefix?: string;
   sample_rows?: number;
 }
 
@@ -90,9 +91,36 @@ export async function connectDataSource(
   _ctx?: any
 ): Promise<ConnectResult> {
   if (!params.datasource_name) throw new Error('datasource_name is required');
-  if (!params.vendor) throw new Error('vendor is required');
-  if (!params.bucket) throw new Error('bucket is required');
-  if (params.prefix === undefined || params.prefix === null) throw new Error('prefix is required');
+  if (!params.url) throw new Error('url is required (e.g. tos://bucket/prefix)');
+  
+  // Parse URL: tos://bucket/prefix
+  try {
+      if (params.url.startsWith('file://') || params.url.startsWith('/')) {
+          params.vendor = 'local';
+          const localPath = params.url.startsWith('file://') ? params.url.slice(7) : params.url;
+          params.bucket = localPath;
+          params.prefix = '.';
+      } else {
+          const parsedUrl = new URL(params.url);
+          const protocol = parsedUrl.protocol.replace(':', '');
+          
+          if (['tos', 'oss', 'cos', 's3'].includes(protocol)) {
+              if (protocol === 'tos') params.vendor = 'volcengine';
+              else if (protocol === 'oss') params.vendor = 'alibaba';
+              else if (protocol === 'cos') params.vendor = 'tencent';
+              else if (protocol === 's3') params.vendor = 'aws';
+              
+              params.bucket = parsedUrl.hostname;
+              params.prefix = parsedUrl.pathname.replace(/^\//, ''); // Remove leading slash
+          } else {
+              throw new Error(`Unsupported protocol: ${protocol}`);
+          }
+      }
+  } catch (e: any) {
+      if (!params.vendor || !params.bucket || params.prefix === undefined) {
+          throw new Error(`Invalid url format: ${e.message}`);
+      }
+  }
 
   if (params.vendor !== 'local') {
     if (!params.endpoint && params.vendor !== 'aws') throw new Error(`endpoint is required for vendor "${params.vendor}"`);
@@ -134,7 +162,14 @@ export async function connectDataSource(
       const scan_ts = new Date().toISOString() + 'Z';
 
       while (isTruncated) {
-          const response = await listS3Objects(params, params.prefix, 1000, continuationToken);
+          const response = await listS3Objects({
+              vendor: params.vendor as any,
+              bucket: params.bucket as string,
+              endpoint: params.endpoint,
+              access_key: params.access_key,
+              secret_key: params.secret_key,
+              region: params.region
+          }, params.prefix || '', 1000, continuationToken);
           
           for (const obj of response.Contents) {
               const key = obj.Key || '';

package/src/lib/actions/s3-tools.ts

@@ -1,17 +1,51 @@
 import { S3Client, ListObjectsV2Command, GetObjectCommand } from '@aws-sdk/client-s3';
+import { getSignedUrl } from '@aws-sdk/s3-request-presigner';
 import * as fs from 'fs';
 import * as path from 'path';
 
 export interface S3Params {
-    vendor: 'volcengine' | 'alibaba' | 'tencent' | 'aws' | 'local';
+    url?: string;
+    vendor?: 'volcengine' | 'alibaba' | 'tencent' | 'aws' | 'local';
     endpoint?: string;
     access_key?: string;
     secret_key?: string;
     region?: string;
-    bucket: string;
+    bucket?: string;
+}
+
+export function parseS3Url(params: S3Params): S3Params {
+    if (params.url) {
+        if (params.url.startsWith('file://') || params.url.startsWith('/')) {
+            params.vendor = 'local';
+            const localPath = params.url.startsWith('file://') ? params.url.slice(7) : params.url;
+            params.bucket = localPath;
+        } else {
+            const parsedUrl = new URL(params.url);
+            const protocol = parsedUrl.protocol.replace(':', '');
+            
+            if (['tos', 'oss', 'cos', 's3'].includes(protocol)) {
+                if (protocol === 'tos') params.vendor = 'volcengine';
+                else if (protocol === 'oss') params.vendor = 'alibaba';
+                else if (protocol === 'cos') params.vendor = 'tencent';
+                else if (protocol === 's3') params.vendor = 'aws';
+                
+                params.bucket = parsedUrl.hostname;
+                // Prefix is usually parsed separately or passed explicitly for listing
+            } else {
+                throw new Error(`Unsupported protocol: ${protocol}`);
+            }
+        }
+    }
+    
+    if (!params.vendor || !params.bucket) {
+         throw new Error('Could not determine vendor or bucket. Please provide a valid url or vendor/bucket directly.');
+    }
+    
+    return params;
 }
 
 function createS3Client(params: S3Params): S3Client | null {
+    params = parseS3Url(params);
     if (params.vendor === 'local') return null;
 
     let endpoint = params.endpoint;
@@ -56,8 +90,9 @@ function createS3Client(params: S3Params): S3Client | null {
 }
 
 export async function listS3Objects(params: S3Params, prefix: string, maxKeys: number = 1000, continuationToken?: string) {
+    params = parseS3Url(params);
     if (params.vendor === 'local') {
-        const root = params.bucket;
+        const root = params.bucket as string;
         const prefixPath = prefix && prefix !== '.' ? path.join(root, prefix) : root;
         const files: any[] = [];
         
@@ -108,8 +143,9 @@ export async function listS3Objects(params: S3Params, prefix: string, maxKeys: n
 }
 
 export async function readS3Object(params: S3Params, key: string, maxBytes?: number): Promise<Buffer> {
+    params = parseS3Url(params);
     if (params.vendor === 'local') {
-        const fullPath = path.join(params.bucket, key);
+        const fullPath = path.join(params.bucket as string, key);
         if (maxBytes) {
             const fd = fs.openSync(fullPath, 'r');
             const buffer = Buffer.alloc(maxBytes);
@@ -146,3 +182,22 @@ export async function readS3Object(params: S3Params, key: string, maxBytes?: num
     }
     return Buffer.alloc(0);
 }
+
+export async function getPresignedUrl(params: S3Params, key: string, expiresIn: number = 3600): Promise<string> {
+    params = parseS3Url(params);
+    if (params.vendor === 'local') {
+        const fullPath = path.join(params.bucket as string, key);
+        return `file://${fullPath}`;
+    }
+
+    const client = createS3Client(params);
+    if (!client) throw new Error('Failed to create S3 client');
+
+    const command = new GetObjectCommand({
+        Bucket: params.bucket,
+        Key: key
+    });
+
+    const signedUrl = await getSignedUrl(client, command, { expiresIn });
+    return signedUrl;
+}