@byted-las/contextlake-openclaw 1.0.7 → 1.0.8

package/dist/src/skills/las-data-profiler/SKILL.md

@@ -7,7 +7,7 @@ description: |
   and writes all results to a local LanceDB. It is also compatible with Alibaba Cloud OSS, Tencent Cloud COS, AWS S3, and the local file system.
 
   IMPORTANT RULE: You are STRICTLY FORBIDDEN from writing or executing Python scripts to access S3/TOS or LanceDB.
-  You MUST exclusively use the provided tools (`list-s3-objects`, `read-s3-object`, `write-lance-catalog`) to accomplish the profiling tasks.
+  You MUST exclusively use the provided tools (`list-s3-objects`, `read-s3-object`, `write-lance-catalog`) via formal tool calls to accomplish the profiling tasks. DO NOT execute them as bash/shell commands.
 ---
 
 ## Trigger Scenarios
@@ -38,8 +38,8 @@ This Skill acts as a Dataset Profiling Guide. You should use the `list-s3-object
 - Table names: `files`, `structured_schemas`, `media_metadata`
 
 ## Available Tools for this Skill
-- `list-s3-objects`: To traverse and list files in the bucket/directory.
-- `read-s3-object`: To read specific bytes of a file for schema inference or metadata extraction.
-- `write-lance-catalog`: To write the profiling results to the LanceDB catalog.
+- `list-s3-objects`: To traverse and list files in the bucket/directory. (Call this as an Agent Tool, NOT a bash command).
+- `read-s3-object`: To read specific bytes of a file for schema inference or metadata extraction. (Call this as an Agent Tool, NOT a bash command).
+- `write-lance-catalog`: To write the profiling results to the LanceDB catalog. (Call this as an Agent Tool, NOT a bash command).
 
 Always report the final profiling summary back to the user once the `write-lance-catalog` completes successfully.

package/openclaw.plugin.json

@@ -1,7 +1,7 @@
 {
   "id": "contextlake-openclaw",
   "name": "ContextLake",
-  "version": "1.0.7",
+  "version": "1.0.8",
   "description": "A lightweight knowledge base plugin for OpenClaw using LanceDB and TOS, with data profiling support",
   "skills": ["./src/skills"],
   "configSchema": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@byted-las/contextlake-openclaw",
-  "version": "1.0.7",
+  "version": "1.0.8",
   "description": "ContextLake OpenClaw Plugin for managing knowledge base",
   "main": "index.ts",
   "files": [
@@ -16,7 +16,7 @@
   },
   "scripts": {
     "build": "tsc && npm run copy-assets",
-    "copy-assets": "cp -r src/skills/*/*.md dist/src/skills/ 2>/dev/null || true && for dir in src/skills/*; do if [ -d \"$dir\" ]; then target=\"dist/$dir\"; mkdir -p \"$target\"; cp \"$dir\"/*.md \"$target\"/ 2>/dev/null || true; fi; done && mkdir -p dist/src/lib/scripts && cp src/lib/scripts/*.py dist/src/lib/scripts/ 2>/dev/null || true",
+    "copy-assets": "mkdir -p dist/src/skills && cp -r src/skills/* dist/src/skills/ 2>/dev/null || true && mkdir -p dist/src/lib/scripts && cp src/lib/scripts/*.py dist/src/lib/scripts/ 2>/dev/null || true",
     "test": "vitest --reporter verbose",
     "test:local": "npx ts-node scripts/local-test.ts",
     "test:profiler": "npx ts-node scripts/local-profiler-test.ts",

package/src/skills/las-data-profiler/SKILL.md

@@ -7,7 +7,7 @@ description: |
   and writes all results to a local LanceDB. It is also compatible with Alibaba Cloud OSS, Tencent Cloud COS, AWS S3, and the local file system.
 
   IMPORTANT RULE: You are STRICTLY FORBIDDEN from writing or executing Python scripts to access S3/TOS or LanceDB.
-  You MUST exclusively use the provided tools (`list-s3-objects`, `read-s3-object`, `write-lance-catalog`) to accomplish the profiling tasks.
+  You MUST exclusively use the provided tools (`list-s3-objects`, `read-s3-object`, `write-lance-catalog`) via formal tool calls to accomplish the profiling tasks. DO NOT execute them as bash/shell commands.
 ---
 
 ## Trigger Scenarios
@@ -38,8 +38,8 @@ This Skill acts as a Dataset Profiling Guide. You should use the `list-s3-object
 - Table names: `files`, `structured_schemas`, `media_metadata`
 
 ## Available Tools for this Skill
-- `list-s3-objects`: To traverse and list files in the bucket/directory.
-- `read-s3-object`: To read specific bytes of a file for schema inference or metadata extraction.
-- `write-lance-catalog`: To write the profiling results to the LanceDB catalog.
+- `list-s3-objects`: To traverse and list files in the bucket/directory. (Call this as an Agent Tool, NOT a bash command).
+- `read-s3-object`: To read specific bytes of a file for schema inference or metadata extraction. (Call this as an Agent Tool, NOT a bash command).
+- `write-lance-catalog`: To write the profiling results to the LanceDB catalog. (Call this as an Agent Tool, NOT a bash command).
 
 Always report the final profiling summary back to the user once the `write-lance-catalog` completes successfully.

package/dist/src/lib/actions/ingest-source.d.ts

@@ -1,15 +0,0 @@
-import { ContextLakeConfig } from '../../utils/config';
-export interface IngestSourceParams {
-    datasource_name: string;
-}
-export declare function ingestSource(params: IngestSourceParams, config: ContextLakeConfig, logger?: any): Promise<({
-    file: any;
-    status: string;
-    chunks: number;
-    message?: undefined;
-} | {
-    file: any;
-    status: string;
-    message: any;
-    chunks?: undefined;
-})[]>;

package/dist/src/lib/actions/ingest-source.js

@@ -1,193 +0,0 @@
-"use strict";
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
-    Object.defineProperty(o, "default", { enumerable: true, value: v });
-}) : function(o, v) {
-    o["default"] = v;
-});
-var __importStar = (this && this.__importStar) || (function () {
-    var ownKeys = function(o) {
-        ownKeys = Object.getOwnPropertyNames || function (o) {
-            var ar = [];
-            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
-            return ar;
-        };
-        return ownKeys(o);
-    };
-    return function (mod) {
-        if (mod && mod.__esModule) return mod;
-        var result = {};
-        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
-        __setModuleDefault(result, mod);
-        return result;
-    };
-})();
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.ingestSource = ingestSource;
-const factory_1 = require("../../service/metadata/factory");
-const las_api_1 = require("./las-api");
-const lancedb = __importStar(require("@lancedb/lancedb"));
-const path = __importStar(require("path"));
-const fs = __importStar(require("fs"));
-const os = __importStar(require("os"));
-// @ts-ignore
-const uuid_1 = require("uuid");
-const BASE_DIR = path.join(os.homedir(), '.openclaw', 'contextlake', 'profiler');
-async function ingestSource(params, config, logger) {
-    if (logger) {
-        logger.info(`[ContextLake-Action] Calling ingestSource with params: ${JSON.stringify(params)}`);
-    }
-    else {
-        // eslint-disable-next-line no-console
-        console.log(`[ContextLake-Action] Calling ingestSource with params: ${JSON.stringify(params)}`);
-    }
-    const dsDir = path.join(BASE_DIR, params.datasource_name);
-    const dbPath = path.join(dsDir, 'catalog_db');
-    if (!fs.existsSync(dbPath)) {
-        throw new Error(`Data source database not found at ${dbPath}. Please run profiler connect first.`);
-    }
-    const metaConfig = config.metadata_storage || { type: 'local', lancedb_uri: './data/contextlake' };
-    const metadataProvider = (0, factory_1.createMetadataProvider)(metaConfig);
-    await metadataProvider.connect();
-    const lasClient = new las_api_1.LasApiClient(config, logger);
-    const results = [];
-    // Connect to the profiler LanceDB to read the file catalog
-    const profilerDb = await lancedb.connect(dbPath);
-    const tableNames = await profilerDb.tableNames();
-    if (!tableNames.includes('file_catalog')) {
-        throw new Error(`table 'file_catalog' not found in ${dbPath}`);
-    }
-    const catalogTable = await profilerDb.openTable('file_catalog');
-    const files = await catalogTable.query().toArray();
-    logger?.info(`[ContextLake-Action] Found ${files.length} files in catalog`);
-    // Simple chunking for text
-    const splitText = (text, chunkSize = 500, overlap = 50) => {
-        const chunks = [];
-        if (!text)
-            return chunks;
-        let i = 0;
-        while (i < text.length) {
-            chunks.push(text.slice(i, i + chunkSize));
-            i += chunkSize - overlap;
-        }
-        return chunks;
-    };
-    const processText = async (text, fileInfo) => {
-        const chunks = splitText(text);
-        const docs = [];
-        for (const chunk of chunks) {
-            const vector = await metadataProvider.generateMultimodalEmbedding([{ type: 'text', text: chunk }]);
-            docs.push({
-                id: (0, uuid_1.v4)(),
-                vector,
-                text: chunk,
-                source: fileInfo.key,
-                file_type: fileInfo.category,
-                storage_type: 'source',
-                url: fileInfo.url || `tos://${fileInfo.bucket}/${fileInfo.key}`,
-                metadata: JSON.stringify({ datasource: params.datasource_name }),
-                created_at: Date.now(),
-                binary_data: Buffer.from('')
-            });
-        }
-        return docs;
-    };
-    for (const file of files) {
-        try {
-            logger?.info(`[ContextLake-Action] Processing file: ${file.key}, type: ${file.media_type}`);
-            let docs = [];
-            const fileUrl = file.url || `tos://${file.bucket}/${file.key}`;
-            if (file.media_type === 'pdf') {
-                // PDF Parse
-                const result = await lasClient.submitAndPoll('las_pdf_parse_doubao', {
-                    url: fileUrl
-                });
-                const markdown = result.data?.markdown || '';
-                docs = await processText(markdown, file);
-            }
-            else if (file.media_type === 'image') {
-                // Multimodal Embedding directly
-                const vector = await metadataProvider.generateMultimodalEmbedding([
-                    { type: 'image_url', image_url: { url: fileUrl } },
-                    { type: 'text', text: 'This is an image from the dataset.' }
-                ]);
-                docs.push({
-                    id: (0, uuid_1.v4)(),
-                    vector,
-                    text: 'Image from dataset',
-                    source: file.key,
-                    file_type: 'image',
-                    storage_type: 'source',
-                    url: fileUrl,
-                    metadata: JSON.stringify({ datasource: params.datasource_name }),
-                    created_at: Date.now(),
-                    binary_data: Buffer.from('')
-                });
-            }
-            else if (file.media_type === 'audio') {
-                // ASR
-                const result = await lasClient.submitAndPoll('las_asr_pro', {
-                    audio: { url: fileUrl, format: file.key.split('.').pop() || 'wav' },
-                    request: { model_name: 'bigmodel' }
-                });
-                const text = result.data?.result?.text || '';
-                docs = await processText(text, file);
-            }
-            else if (file.media_type === 'video') {
-                // Video understanding -> text -> embedding
-                const result = await lasClient.submitAndPoll('las_long_video_understand', {
-                    video_url: fileUrl,
-                    query: "详细描述这个视频的内容",
-                    model_name: "doubao-seed-2-0-lite-260215"
-                });
-                // Assuming video output is a text description somewhere in the response. 
-                // Note: the exact structure depends on the API return, adjusting to generic text.
-                const text = JSON.stringify(result.data || '');
-                // Also need audio extract and ASR for video
-                // 1. Extract audio
-                // The output_path_template needs a unique path per video
-                const audioOutputPath = `tos://${file.bucket}/.tmp/audio/${(0, uuid_1.v4)()}.wav`;
-                await lasClient.process('las_audio_extract_and_split', {
-                    input_path: fileUrl,
-                    output_path_template: audioOutputPath,
-                    output_format: 'wav'
-                });
-                // 2. ASR on the extracted audio
-                // Wait briefly for object to be available if needed (often synchronous but tos takes a ms)
-                const asrResult = await lasClient.submitAndPoll('las_asr_pro', {
-                    audio: { url: audioOutputPath.replace('{index}.{output_file_ext}', '0.wav'), format: 'wav' },
-                    request: { model_name: 'bigmodel' }
-                });
-                const audioText = asrResult.data?.result?.text || '';
-                // Combine video text and audio text
-                const combinedText = `Video Description: ${text}\n\nAudio Transcription: ${audioText}`;
-                docs = await processText(combinedText, file);
-            }
-            else if (file.category === 'structured' || file.category === 'non-structured') {
-                // If we had a direct text content, we could process it here. 
-                // Assuming basic local download or similar is available, but for now we skip raw file reading from TOS in this demo script unless implemented.
-                // Fallback just logs
-                logger?.warn(`[ContextLake-Action] Skipping raw text/structured download for ${file.key} - implement TOS download if needed`);
-            }
-            if (docs.length > 0) {
-                await metadataProvider.addAssets(docs);
-                results.push({ file: file.key, status: 'success', chunks: docs.length });
-            }
-        }
-        catch (error) {
-            logger?.error(`[ContextLake-Action] Error processing ${file.key}: ${error.message}`);
-            results.push({ file: file.key, status: 'error', message: error.message });
-        }
-    }
-    return results;
-}

package/dist/src/lib/actions/las.d.ts

@@ -1,64 +0,0 @@
-import { ContextLakeConfig } from '../../utils/config';
-export declare function lasPdfParseDoubao(params: {
-    url: string;
-    start_page?: number;
-    num_pages?: number;
-    parse_mode?: string;
-}, config?: ContextLakeConfig): Promise<any>;
-export declare function lasLongVideoUnderstand(params: {
-    video_url: string;
-    prompt: string;
-    system_prompt?: string;
-    return_chunk_text?: boolean;
-    max_tokens?: number;
-    temperature?: number;
-    top_p?: number;
-}, config?: ContextLakeConfig): Promise<any>;
-export declare function lasBareImageTextEmbedding(params: {
-    input: Array<{
-        type: string;
-        text?: string;
-        image_url?: string;
-    }>;
-    encoding_format?: string;
-}, config?: ContextLakeConfig): Promise<any>;
-export declare function lasSeed20(params: {
-    model: string;
-    messages: Array<any>;
-    stream?: boolean;
-    max_tokens?: number;
-    temperature?: number;
-    top_p?: number;
-    frequency_penalty?: number;
-    presence_penalty?: number;
-    tools?: Array<any>;
-    tool_choice?: any;
-    user?: string;
-    logprobs?: boolean;
-    top_logprobs?: number;
-}, config?: ContextLakeConfig): Promise<any>;
-export declare function lasAsrPro(params: {
-    url?: string;
-    format?: string;
-    language?: string;
-    resource?: string;
-    use_itn?: boolean;
-    use_sn?: boolean;
-    enable_alignment?: boolean;
-    channel_id?: number;
-    use_word_info?: boolean;
-    text_format?: number;
-    enable_semantic_sentence_detection?: boolean;
-    boost_words?: Array<{
-        word: string;
-        weight: number;
-    }>;
-}, config?: ContextLakeConfig): Promise<any>;
-export declare function lasAudioExtractAndSplit(params: {
-    input_path: string;
-    output_path_template: string;
-    split_duration?: number;
-    output_format?: string;
-    timeout?: number;
-    extra_params?: string[];
-}, config?: ContextLakeConfig): Promise<any>;

package/dist/src/lib/actions/las.js

@@ -1,72 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.lasPdfParseDoubao = lasPdfParseDoubao;
-exports.lasLongVideoUnderstand = lasLongVideoUnderstand;
-exports.lasBareImageTextEmbedding = lasBareImageTextEmbedding;
-exports.lasSeed20 = lasSeed20;
-exports.lasAsrPro = lasAsrPro;
-exports.lasAudioExtractAndSplit = lasAudioExtractAndSplit;
-function getLASConfig(config) {
-    // Attempt to get from env vars or config
-    const endpoint = process.env.LAS_ENDPOINT || (config?.las)?.endpoint;
-    const apiKey = process.env.LAS_API_KEY || (config?.las)?.api_key;
-    if (!endpoint || !apiKey) {
-        throw new Error("LAS_ENDPOINT and LAS_API_KEY must be set in environment variables or config");
-    }
-    return { endpoint, apiKey };
-}
-async function lasFetch(path, payload, config) {
-    const { endpoint, apiKey } = getLASConfig(config);
-    const url = `${endpoint.replace(/\/$/, '')}${path}`;
-    const response = await fetch(url, {
-        method: 'POST',
-        headers: {
-            'Content-Type': 'application/json',
-            'Authorization': `Bearer ${apiKey}`
-        },
-        body: JSON.stringify(payload)
-    });
-    if (!response.ok) {
-        let errorText = await response.text().catch(() => '');
-        throw new Error(`LAS API Error: ${response.status} ${response.statusText} - ${errorText}`);
-    }
-    return await response.json();
-}
-async function lasPdfParseDoubao(params, config) {
-    return lasFetch('/api/v1/submit', {
-        operator_id: 'las_pdf_parse_doubao',
-        operator_version: 'v1',
-        data: params
-    }, config);
-}
-async function lasLongVideoUnderstand(params, config) {
-    return lasFetch('/api/v1/submit', {
-        operator_id: 'las_long_video_understand',
-        operator_version: 'v1',
-        data: params
-    }, config);
-}
-async function lasBareImageTextEmbedding(params, config) {
-    return lasFetch('/api/v1/embeddings/multimodal', {
-        model: 'doubao-embedding-vision',
-        input: params.input,
-        encoding_format: params.encoding_format
-    }, config);
-}
-async function lasSeed20(params, config) {
-    return lasFetch('/api/v1/chat/completions', params, config);
-}
-async function lasAsrPro(params, config) {
-    return lasFetch('/api/v1/submit', {
-        operator_id: 'las_asr_pro',
-        operator_version: 'v1',
-        data: params
-    }, config);
-}
-async function lasAudioExtractAndSplit(params, config) {
-    return lasFetch('/api/v1/process', {
-        operator_id: 'las_audio_extract_and_split',
-        operator_version: 'v1',
-        data: params
-    }, config);
-}

package/dist/src/lib/scripts/s3_catalog.py

@@ -1,617 +0,0 @@
-#!/usr/bin/env python3
-"""
-s3_catalog.py  --  Data source profiling tool for LanceDB
-
-Three-pass scanning:
-  Pass 1: Walk all files -> file_catalog
-  Pass 2: Sample structured data -> structured_schemas
-  Pass 3: Read media file headers -> media_metadata
-
-Supported sources: volcengine TOS / alibaba OSS / tencent COS / aws S3 / local
-Output: LanceDB tables (file_catalog, structured_schemas, media_metadata)
-"""
-
-import argparse
-import datetime
-import io
-import json
-import mimetypes
-import os
-import re
-import sys
-import time
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-# ---------------------------------------------------------------------------
-# Lazy imports
-# ---------------------------------------------------------------------------
-
-def _import_boto3():
-    try:
-        import boto3
-        from botocore.config import Config as BotoConfig
-        return boto3, BotoConfig
-    except ImportError:
-        print("ERROR: boto3 not installed. Run: pip install boto3", file=sys.stderr)
-        sys.exit(1)
-
-def _import_lancedb():
-    try:
-        import lancedb
-        return lancedb
-    except ImportError:
-        print("ERROR: lancedb not installed. Run: pip install lancedb", file=sys.stderr)
-        sys.exit(1)
-
-def _import_pandas():
-    try:
-        import pandas as pd
-        return pd
-    except ImportError:
-        print("ERROR: pandas not installed. Run: pip install pandas", file=sys.stderr)
-        sys.exit(1)
-
-def _import_pyarrow():
-    try:
-        import pyarrow as pa
-        import pyarrow.parquet as pq
-        return pa, pq
-    except ImportError:
-        print("ERROR: pyarrow not installed. Run: pip install pyarrow", file=sys.stderr)
-        sys.exit(1)
-
-# ---------------------------------------------------------------------------
-# Constants
-# ---------------------------------------------------------------------------
-
-STRUCTURED_EXTS = {'.json', '.jsonl', '.ndjson', '.csv', '.tsv', '.parquet', '.pq'}
-IMAGE_EXTS = {'.jpg', '.jpeg', '.png', '.gif', '.bmp', '.webp', '.tiff', '.tif', '.svg', '.ico', '.heic', '.heif'}
-AUDIO_EXTS = {'.mp3', '.wav', '.flac', '.aac', '.ogg', '.m4a', '.wma', '.opus'}
-VIDEO_EXTS = {'.mp4', '.avi', '.mov', '.mkv', '.webm', '.wmv', '.flv', '.m4v', '.3gp'}
-PDF_EXTS = {'.pdf'}
-
-IMAGE_HEAD_BYTES = 64 * 1024
-AUDIO_HEAD_BYTES = 512 * 1024
-VIDEO_HEAD_BYTES = 2 * 1024 * 1024
-PDF_HEAD_BYTES = 256 * 1024
-
-
-def classify_file(ext: str) -> Tuple[str, Optional[str]]:
-    ext = ext.lower()
-    if ext in STRUCTURED_EXTS:
-        return ('structured', None)
-    if ext in IMAGE_EXTS:
-        return ('non-structured', 'image')
-    if ext in AUDIO_EXTS:
-        return ('non-structured', 'audio')
-    if ext in VIDEO_EXTS:
-        return ('non-structured', 'video')
-    if ext in PDF_EXTS:
-        return ('non-structured', 'pdf')
-    return ('non-structured', None)
-
-
-# ===================================================================
-# S3 Client Abstraction
-# ===================================================================
-
-class S3Client:
-    """Thin wrapper around boto3 S3 with vendor-specific configuration."""
-
-    def __init__(self, vendor: str, endpoint: str, credential_id: str, credential_secret: str, region: str):
-        boto3, BotoConfig = _import_boto3()
-        self.vendor = vendor
-
-        config_kwargs: Dict[str, Any] = {}
-        if vendor == 'volcengine':
-            config_kwargs['s3'] = {'addressing_style': 'virtual'}
-        elif vendor == 'alibaba':
-            config_kwargs['signature_version'] = 's3'
-            config_kwargs['s3'] = {'addressing_style': 'virtual'}
-        elif vendor == 'tencent':
-            config_kwargs['s3'] = {'addressing_style': 'virtual'}
-
-        self.client = boto3.client(
-            's3',
-            endpoint_url=endpoint or None,
-            aws_access_key_id=credential_id,
-            aws_secret_access_key=credential_secret,
-            region_name=region,
-            config=BotoConfig(**config_kwargs) if config_kwargs else None,
-        )
-
-    def list_objects(self, bucket: str, prefix: str):
-        paginator = self.client.get_paginator('list_objects_v2')
-        for page in paginator.paginate(Bucket=bucket, Prefix=prefix):
-            for obj in page.get('Contents', []):
-                yield obj
-
-    def get_range(self, bucket: str, key: str, start: int, end: int) -> bytes:
-        resp = self.client.get_object(Bucket=bucket, Key=key, Range=f'bytes={start}-{end}')
-        return resp['Body'].read()
-
-    def get_object(self, bucket: str, key: str, max_bytes: Optional[int] = None) -> bytes:
-        kwargs: Dict[str, Any] = {'Bucket': bucket, 'Key': key}
-        if max_bytes:
-            kwargs['Range'] = f'bytes=0-{max_bytes - 1}'
-        resp = self.client.get_object(**kwargs)
-        return resp['Body'].read()
-
-
-# ===================================================================
-# Local FS Abstraction
-# ===================================================================
-
-class LocalClient:
-    def __init__(self, root: str):
-        self.root = root
-
-    def list_objects(self, bucket: str, prefix: str):
-        base = Path(bucket)
-        prefix_path = base / prefix if prefix and prefix != '.' else base
-        for dirpath, _dirs, files in os.walk(prefix_path):
-            for fname in files:
-                full = Path(dirpath) / fname
-                stat = full.stat()
-                key = str(full.relative_to(base))
-                yield {
-                    'Key': key,
-                    'Size': stat.st_size,
-                    'LastModified': datetime.datetime.fromtimestamp(stat.st_mtime),
-                    'ETag': '',
-                    'StorageClass': 'LOCAL',
-                    '_created_time': datetime.datetime.fromtimestamp(stat.st_ctime),
-                }
-
-    def get_range(self, bucket: str, key: str, start: int, end: int) -> bytes:
-        fpath = Path(bucket) / key
-        with open(fpath, 'rb') as f:
-            f.seek(start)
-            return f.read(end - start + 1)
-
-    def get_object(self, bucket: str, key: str, max_bytes: Optional[int] = None) -> bytes:
-        fpath = Path(bucket) / key
-        with open(fpath, 'rb') as f:
-            if max_bytes:
-                return f.read(max_bytes)
-            return f.read()
-
-
-# ===================================================================
-# Pass 1: File Catalog
-# ===================================================================
-
-def build_file_catalog(client, bucket: str, prefix: str, vendor: str) -> List[Dict]:
-    catalog: List[Dict] = []
-    scan_ts = datetime.datetime.utcnow().isoformat() + 'Z'
-
-    for obj in client.list_objects(bucket, prefix):
-        key = obj['Key']
-        if key.endswith('/'):
-            continue
-
-        name = os.path.basename(key)
-        ext = os.path.splitext(name)[1].lower()
-        mime, _ = mimetypes.guess_type(name)
-        category, media_type = classify_file(ext)
-        etag = obj.get('ETag', '').strip('"')
-        depth = key.count('/')
-        parent_dir = os.path.basename(os.path.dirname(key)) if '/' in key else ''
-
-        catalog.append({
-            'file_path': key,
-            'file_name': name,
-            'extension': ext,
-            'mime_type': mime or '',
-            'category': category,
-            'media_type': media_type or '',
-            'size_bytes': obj.get('Size', 0),
-            'last_modified': str(obj.get('LastModified', '')),
-            'created_time': str(obj.get('_created_time', '')),
-            'etag': etag,
-            'storage_class': obj.get('StorageClass', ''),
-            'is_multipart': '-' in etag,
-            'depth': depth,
-            'parent_dir': parent_dir,
-            'vendor': vendor,
-            'bucket': bucket,
-            'has_schema': False,
-            'has_media_meta': False,
-            'scan_timestamp': scan_ts,
-        })
-
-    return catalog
-
-
-# ===================================================================
-# Pass 2: Structured Schema Analysis
-# ===================================================================
-
-def infer_semantic_hint(series) -> Tuple[str, str]:
-    pd = _import_pandas()
-    non_null = series.dropna()
-    n = len(non_null)
-    if n == 0:
-        return ('constant', 'all null')
-
-    unique_count = non_null.nunique()
-    unique_ratio = unique_count / n if n > 0 else 0
-
-    if set(non_null.unique()).issubset({True, False, 0, 1, 'true', 'false', 'True', 'False'}):
-        return ('boolean', f'{unique_count} distinct values')
-
-    if unique_count == 1:
-        return ('constant', f'value: {non_null.iloc[0]}')
-
-    dtype_str = str(series.dtype)
-
-    if 'int' in dtype_str or 'float' in dtype_str:
-        return ('numeric', dtype_str)
-
-    sample_val = non_null.iloc[0]
-    if isinstance(sample_val, (list, dict)):
-        return ('structured', type(sample_val).__name__)
-
-    try:
-        str_vals = non_null.astype(str)
-        avg_len = str_vals.str.len().mean()
-
-        path_pattern = re.compile(r'[\\\\/]|^s3://|^tos://|^gs://|^https?://')
-        path_ratio = str_vals.apply(lambda x: bool(path_pattern.search(x))).mean()
-        if path_ratio > 0.5:
-            return ('file_path', f'{path_ratio:.0%} match path/URI pattern')
-
-        ts_pattern = re.compile(r'\d{4}[-/]\d{2}[-/]\d{2}')
-        ts_ratio = str_vals.apply(lambda x: bool(ts_pattern.search(x))).mean()
-        if ts_ratio > 0.5:
-            return ('timestamp', f'{ts_ratio:.0%} match timestamp pattern')
-
-        if unique_ratio > 0.9 and avg_len < 50:
-            return ('id', f'unique_ratio={unique_ratio:.2f}, avg_len={avg_len:.1f}')
-
-        if unique_count < 50 or unique_ratio < 0.2:
-            return ('categorical', f'{unique_count} categories')
-
-        if avg_len > 50 and unique_ratio > 0.5:
-            return ('text', f'avg_len={avg_len:.1f}')
-
-    except Exception:
-        pass
-
-    return ('text', '')
-
-
-def analyze_structured_file(client, bucket: str, key: str, ext: str, sample_rows: int) -> List[Dict]:
-    pd = _import_pandas()
-    pa, pq = _import_pyarrow()
-
-    schemas: List[Dict] = []
-    max_download = 2 * 1024 * 1024
-
-    try:
-        raw = client.get_object(bucket, key, max_bytes=max_download)
-    except Exception as e:
-        return [{'file_path': key, 'error': str(e)}]
-
-    df = None
-    fmt = ext.lstrip('.')
-
-    try:
-        if ext in ('.parquet', '.pq'):
-            buf = io.BytesIO(raw)
-            table = pq.read_table(buf)
-            df = table.to_pandas().head(sample_rows)
-            fmt = 'parquet'
-        elif ext == '.csv':
-            df = pd.read_csv(io.BytesIO(raw), nrows=sample_rows, on_bad_lines='skip')
-        elif ext == '.tsv':
-            df = pd.read_csv(io.BytesIO(raw), sep='\t', nrows=sample_rows, on_bad_lines='skip')
-        elif ext in ('.jsonl', '.ndjson'):
-            lines = raw.decode('utf-8', errors='replace').strip().split('\n')[:sample_rows]
-            records = [json.loads(line) for line in lines if line.strip()]
-            df = pd.json_normalize(records)
-        elif ext == '.json':
-            data = json.loads(raw.decode('utf-8', errors='replace'))
-            if isinstance(data, list):
-                df = pd.json_normalize(data[:sample_rows])
-            elif isinstance(data, dict):
-                df = pd.json_normalize([data])
-    except Exception as e:
-        return [{'file_path': key, 'error': f'parse error: {e}'}]
-
-    if df is None or df.empty:
-        return []
-
-    for col in df.columns:
-        series = df[col]
-        non_null = series.dropna()
-        unique_count = int(non_null.nunique()) if len(non_null) > 0 else 0
-        non_null_ratio = len(non_null) / len(series) if len(series) > 0 else 0.0
-
-        sample_values = []
-        try:
-            sample_values = [str(v) for v in non_null.unique()[:3]]
-        except Exception:
-            pass
-
-        hint, detail = infer_semantic_hint(series)
-
-        schemas.append({
-            'file_path': key,
-            'vendor': '',
-            'bucket': bucket,
-            'format': fmt,
-            'column_name': str(col),
-            'column_type': str(series.dtype),
-            'non_null_ratio': round(non_null_ratio, 4),
-            'unique_count': unique_count,
-            'sample_values': json.dumps(sample_values, ensure_ascii=False),
-            'semantic_hint': hint,
-            'semantic_detail': detail,
-        })
-
-    return schemas
-
-
-# ===================================================================
-# Pass 3: Media Metadata Extraction
-# ===================================================================
-
-def extract_image_meta(data: bytes) -> Dict:
-    meta: Dict[str, Any] = {
-        'width': 0, 'height': 0, 'image_format': '', 'color_mode': '', 'exif_summary': '{}',
-    }
-    try:
-        from PIL import Image
-        img = Image.open(io.BytesIO(data))
-        meta['width'] = img.width
-        meta['height'] = img.height
-        meta['image_format'] = img.format or ''
-        meta['color_mode'] = img.mode or ''
-
-        exif = {}
-        exif_data = getattr(img, '_getexif', lambda: None)()
-        if exif_data:
-            for tag_id, value in list(exif_data.items())[:10]:
-                try:
-                    from PIL.ExifTags import TAGS
-                    tag_name = TAGS.get(tag_id, str(tag_id))
-                    exif[tag_name] = str(value)[:100]
-                except Exception:
-                    pass
-        meta['exif_summary'] = json.dumps(exif, ensure_ascii=False)
-    except Exception as e:
-        meta['extract_error'] = str(e)
-    return meta
-
-
-def extract_audio_meta(data: bytes) -> Dict:
-    meta: Dict[str, Any] = {
-        'duration_sec': 0.0, 'codec': '', 'sample_rate': 0, 'channels': 0,
-        'bitrate': 0, 'tags_summary': '{}',
-    }
-    try:
-        import mutagen
-        f = mutagen.File(io.BytesIO(data))
-        if f:
-            info = getattr(f, 'info', None)
-            if info:
-                meta['duration_sec'] = round(getattr(info, 'length', 0.0), 2)
-                meta['sample_rate'] = getattr(info, 'sample_rate', 0)
-                meta['channels'] = getattr(info, 'channels', 0)
-                meta['bitrate'] = getattr(info, 'bitrate', 0)
-                meta['codec'] = type(info).__name__
-
-            tags = {}
-            if f.tags:
-                for k in list(f.tags.keys())[:10]:
-                    try:
-                        tags[str(k)] = str(f.tags[k])[:100]
-                    except Exception:
-                        pass
-            meta['tags_summary'] = json.dumps(tags, ensure_ascii=False)
-    except Exception as e:
-        meta['extract_error'] = str(e)
-    return meta
-
-
-def extract_video_meta(data: bytes) -> Dict:
-    meta: Dict[str, Any] = {
-        'width': 0, 'height': 0, 'duration_sec': 0.0, 'container': '',
-    }
-    try:
-        if data[:4] == b'\x1a\x45\xdf\xa3':
-            meta['container'] = 'mkv/webm'
-        elif len(data) > 8 and data[4:8] == b'ftyp':
-            ftyp = data[8:12].decode('ascii', errors='replace').strip()
-            meta['container'] = ftyp
-        elif data[:4] == b'RIFF':
-            meta['container'] = 'avi'
-        elif data[:3] == b'FLV':
-            meta['container'] = 'flv'
-    except Exception as e:
-        meta['extract_error'] = str(e)
-    return meta
-
-
-def extract_pdf_meta(data: bytes) -> Dict:
-    meta: Dict[str, Any] = {
-        'page_count': 0, 'pdf_title': '', 'pdf_author': '',
-        'creation_date': '', 'encrypted': False,
-        'page_width_pt': 0.0, 'page_height_pt': 0.0,
-    }
-    try:
-        import fitz
-        doc = fitz.open(stream=data, filetype='pdf')
-        meta['page_count'] = doc.page_count
-        md = doc.metadata or {}
-        meta['pdf_title'] = md.get('title', '')
-        meta['pdf_author'] = md.get('author', '')
-        meta['creation_date'] = md.get('creationDate', '')
-        meta['encrypted'] = doc.is_encrypted
-
-        if doc.page_count > 0:
-            page = doc[0]
-            rect = page.rect
-            meta['page_width_pt'] = round(rect.width, 2)
-            meta['page_height_pt'] = round(rect.height, 2)
-        doc.close()
-    except Exception as e:
-        meta['extract_error'] = str(e)
-    return meta
-
-
-def extract_media_metadata(client, bucket: str, key: str, media_type: str) -> Dict:
-    head_size = {
-        'image': IMAGE_HEAD_BYTES,
-        'audio': AUDIO_HEAD_BYTES,
-        'video': VIDEO_HEAD_BYTES,
-        'pdf': PDF_HEAD_BYTES,
-    }.get(media_type, IMAGE_HEAD_BYTES)
-
-    try:
-        data = client.get_object(bucket, key, max_bytes=head_size)
-    except Exception as e:
-        return {'extract_error': f'download failed: {e}'}
-
-    if media_type == 'image':
-        return extract_image_meta(data)
-    elif media_type == 'audio':
-        return extract_audio_meta(data)
-    elif media_type == 'video':
-        return extract_video_meta(data)
-    elif media_type == 'pdf':
-        return extract_pdf_meta(data)
-    return {}
-
-
-# ===================================================================
-# LanceDB Writer
-# ===================================================================
-
-def write_to_lancedb(db_path: str, table_name: str, records: List[Dict]):
-    if not records:
-        return
-    lancedb = _import_lancedb()
-    pd = _import_pandas()
-
-    db = lancedb.connect(db_path)
-    df = pd.DataFrame(records)
-
-    table_names = db.table_names()
-    if table_name in table_names:
-        db.drop_table(table_name)
-
-    db.create_table(table_name, data=df)
-    print(f"  [LanceDB] Wrote {len(records)} records to '{table_name}'")
-
-
-# ===================================================================
-# Main
-# ===================================================================
-
-def main():
-    parser = argparse.ArgumentParser(description='S3-compatible data profiler -> LanceDB')
-    parser.add_argument('--vendor', required=True,
-                        choices=['volcengine', 'alibaba', 'tencent', 'aws', 'local'])
-    parser.add_argument('--endpoint', default='')
-    parser.add_argument('--ak', default='', dest='cred_id',
-                        help='Access credential ID')
-    parser.add_argument('--sk', default='', dest='cred_secret',
-                        help='Access credential value')
-    parser.add_argument('--region', default='')
-    parser.add_argument('--bucket', required=True)
-    parser.add_argument('--prefix', required=True)
-    parser.add_argument('--db-path', default=None, help='Path to LanceDB database. Defaults to ~/.openclaw/las-data-profiler/{datasource_name}/catalog_db if datasource_name is provided.')
-    parser.add_argument('--datasource-name', default='', help='Name of the datasource. Used to determine default db-path if not explicitly provided.')
-    parser.add_argument('--sample-rows', type=int, default=100)
-    args = parser.parse_args()
-
-    if not args.db_path:
-        if args.datasource_name:
-            import os
-            home_dir = os.path.expanduser('~')
-            args.db_path = os.path.join(home_dir, '.openclaw', 'las-data-profiler', args.datasource_name, 'catalog_db')
-        else:
-            args.db_path = './catalog_db'
-
-    print(f"[las-data-profiler] vendor={args.vendor}, bucket={args.bucket}, prefix={args.prefix}")
-    print(f"[las-data-profiler] db_path={args.db_path}")
-
-    if args.vendor == 'local':
-        client = LocalClient(args.bucket)
-    else:
-        client = S3Client(
-            vendor=args.vendor,
-            endpoint=args.endpoint,
-            credential_id=args.cred_id,
-            credential_secret=args.cred_secret,
-            region=args.region,
-        )
-
-    # ---- Pass 1: File Catalog ----
-    print("\n[Pass 1] Scanning files...")
-    catalog = build_file_catalog(client, args.bucket, args.prefix, args.vendor)
-    print(f"  Found {len(catalog)} files")
-
-    # ---- Pass 2: Structured Schemas ----
-    print("\n[Pass 2] Analyzing structured data...")
-    structured_files = [f for f in catalog if f['category'] == 'structured']
-    all_schemas: List[Dict] = []
-    for i, entry in enumerate(structured_files):
-        key = entry['file_path']
-        ext = entry['extension']
-        print(f"  [{i+1}/{len(structured_files)}] {key}")
-        schemas = analyze_structured_file(client, args.bucket, key, ext, args.sample_rows)
-        for s in schemas:
-            s['vendor'] = args.vendor
-        all_schemas.extend(schemas)
-        entry['has_schema'] = True
-
-    print(f"  Analyzed {len(structured_files)} files, {len(all_schemas)} column records")
-
-    # ---- Pass 3: Media Metadata ----
-    print("\n[Pass 3] Extracting media metadata...")
-    media_files = [f for f in catalog if f['media_type'] in ('image', 'audio', 'video', 'pdf')]
-    all_media_meta: List[Dict] = []
-    for i, entry in enumerate(media_files):
-        key = entry['file_path']
-        media_type = entry['media_type']
-        print(f"  [{i+1}/{len(media_files)}] {key} ({media_type})")
-        meta = extract_media_metadata(client, args.bucket, key, media_type)
-        meta['file_path'] = key
-        meta['vendor'] = args.vendor
-        meta['bucket'] = args.bucket
-        meta['media_type'] = media_type
-        for col in ['width', 'height', 'image_format', 'color_mode', 'exif_summary',
-                     'duration_sec', 'codec', 'sample_rate', 'channels', 'bitrate',
-                     'tags_summary', 'container',
-                     'page_count', 'pdf_title', 'pdf_author', 'creation_date',
-                     'encrypted', 'page_width_pt', 'page_height_pt', 'extract_error']:
-            meta.setdefault(col, '' if isinstance(meta.get(col), str) else 0)
-        all_media_meta.append(meta)
-        entry['has_media_meta'] = True
-
-    print(f"  Extracted metadata for {len(media_files)} media files")
-
-    # ---- Write to LanceDB ----
-    print(f"\n[LanceDB] Writing to {args.db_path}")
-    write_to_lancedb(args.db_path, 'file_catalog', catalog)
-    write_to_lancedb(args.db_path, 'structured_schemas', all_schemas)
-    write_to_lancedb(args.db_path, 'media_metadata', all_media_meta)
-
-    # ---- Summary JSON (stdout, for Node.js to parse) ----
-    summary = {
-        'summary': {
-            'total_files': len(catalog),
-            'structured_files': len(structured_files),
-            'media_files': len(media_files),
-        }
-    }
-    print(f"\n{json.dumps(summary)}")
-    print("\n[las-data-profiler] Done!")
-
-
-if __name__ == '__main__':
-    main()

package/dist/src/service/embedding/local.d.ts

@@ -1,14 +0,0 @@
-import { EmbeddingProvider, EmbeddingConfig } from './interface';
-export declare const setNodeLlamaCppImporter: (importer: () => Promise<any>) => void;
-export declare class LocalEmbeddingProvider implements EmbeddingProvider {
-    private llama;
-    private model;
-    private context;
-    private initPromise;
-    private modelPath;
-    constructor(config: EmbeddingConfig);
-    private ensureInitialized;
-    private doInitialize;
-    generateEmbedding(text: string): Promise<number[]>;
-    generateEmbeddings(texts: string[]): Promise<number[][]>;
-}

package/dist/src/service/embedding/local.js

@@ -1,107 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.LocalEmbeddingProvider = exports.setNodeLlamaCppImporter = void 0;
-// import type { Llama, LlamaEmbeddingContext, LlamaModel } from 'node-llama-cpp';
-const DEFAULT_LOCAL_MODEL = 'hf:CompendiumLabs/bge-small-zh-v1.5-gguf/bge-small-zh-v1.5-f16.gguf';
-let nodeLlamaImportPromise = null;
-const setNodeLlamaCppImporter = (importer) => {
-    nodeLlamaImportPromise = importer();
-};
-exports.setNodeLlamaCppImporter = setNodeLlamaCppImporter;
-const importNodeLlamaCpp = async () => {
-    if (!nodeLlamaImportPromise) {
-        nodeLlamaImportPromise = import('node-llama-cpp');
-    }
-    return nodeLlamaImportPromise;
-};
-class LocalEmbeddingProvider {
-    llama = null;
-    model = null;
-    context = null;
-    initPromise = null;
-    modelPath;
-    constructor(config) {
-        // Override transformers.js default with node-llama-cpp default
-        this.modelPath = config.model_name === 'Xenova/all-MiniLM-L6-v2'
-            ? DEFAULT_LOCAL_MODEL
-            : (config.model_name || DEFAULT_LOCAL_MODEL);
-    }
-    async ensureInitialized() {
-        if (this.context) {
-            return;
-        }
-        if (this.initPromise) {
-            return this.initPromise;
-        }
-        this.initPromise = this.doInitialize().catch((err) => {
-            this.initPromise = null;
-            this.context = null;
-            this.model = null;
-            this.llama = null;
-            throw err;
-        });
-        return this.initPromise;
-    }
-    async doInitialize() {
-        try {
-            const { getLlama, resolveModelFile, LlamaLogLevel } = await importNodeLlamaCpp();
-            if (!this.llama) {
-                this.llama = await getLlama({ logLevel: LlamaLogLevel.error });
-            }
-            if (!this.model) {
-                const resolved = await resolveModelFile(this.modelPath);
-                this.model = await this.llama.loadModel({ modelPath: resolved });
-            }
-            if (!this.context) {
-                this.context = await this.model.createEmbeddingContext();
-            }
-        }
-        catch (err) {
-            const detail = err instanceof Error ? err.message : String(err);
-            throw new Error(`Local embeddings unavailable. Reason: ${detail}`, {
-                cause: err,
-            });
-        }
-    }
-    async generateEmbedding(text) {
-        if (!text || !text.trim()) {
-            throw new Error('Embedding input text must be a non-empty string');
-        }
-        await this.ensureInitialized();
-        const embedding = await this.context.getEmbeddingFor(text);
-        const vector = embedding.vector; // TypedArray
-        // Optimized normalization loop
-        let sumSq = 0;
-        const len = vector.length;
-        for (let i = 0; i < len; i++) {
-            const val = vector[i];
-            if (Number.isFinite(val)) {
-                sumSq += val * val;
-            }
-            else {
-                vector[i] = 0;
-            }
-        }
-        const magnitude = Math.sqrt(sumSq);
-        const result = new Array(len);
-        if (magnitude > 0) {
-            const scale = 1.0 / magnitude;
-            for (let i = 0; i < len; i++) {
-                result[i] = vector[i] * scale;
-            }
-        }
-        else {
-            for (let i = 0; i < len; i++) {
-                result[i] = vector[i];
-            }
-        }
-        return result;
-    }
-    async generateEmbeddings(texts) {
-        if (!Array.isArray(texts)) {
-            throw new Error('Embedding input must be an array of strings');
-        }
-        return Promise.all(texts.map(text => this.generateEmbedding(text)));
-    }
-}
-exports.LocalEmbeddingProvider = LocalEmbeddingProvider;

package/dist/src/skills/SKILL.md

@@ -1,45 +0,0 @@
----
-name: byted-las-data-profiler
-description: |
-  Volcengine TOS Dataset Profiling Tool. Based on the S3-compatible protocol, it scans the file structure in TOS buckets and catalogs them,
-  performs schema inference and column semantic analysis on structured data (JSONL/CSV/Parquet/JSON),
-  extracts key meta-information for media files (Image/Audio/Video/PDF) by reading only header bytes,
-  and writes all results to a local LanceDB. It is also compatible with Alibaba Cloud OSS, Tencent Cloud COS, AWS S3, and the local file system.
-
-  IMPORTANT RULE: You are STRICTLY FORBIDDEN from writing or executing Python scripts to access S3/TOS or LanceDB.
-  You MUST exclusively use the provided tools (`list-s3-objects`, `read-s3-object`, `write-lance-catalog`) to accomplish the profiling tasks.
----
-
-## Trigger Scenarios
-Be sure to use this Skill when the user mentions the following scenarios:
-- Need to scan the file structure in a TOS bucket or understand the dataset composition
-- Need to connect to object storage (TOS/OSS/COS/S3) using the S3 protocol
-- Need to scan, traverse, or catalog the file structure of a specific bucket or local directory
-- Need to understand what a batch of data files contains and what their schema looks like
-- Need to extract meta-information such as image resolution, audio/video duration, PDF page count, etc.
-- Need to write the meta-information of object storage or local files into LanceDB
-- Mentions TOS, boto3, or object storage data profiling
-- Mentions keywords like "dataset scanning", "file cataloging", "data catalog", "data profiling", etc.
-- Need to batch identify the type and size of remote/local files and build an index
-- Need to quickly understand the structure of an unfamiliar dataset (what files are there, how the schema is, field meanings)
-- Need to connect/dock a data source for profiling
-- Mentions "connect" data source, docking data source
-
-## Overview
-This Skill acts as a Dataset Profiling Guide. You should use the `list-s3-objects` tool to traverse the S3 bucket or local directory, use `read-s3-object` to read file contents or headers, parse the schema or media metadata, and finally use `write-lance-catalog` to save the catalog into a local LanceDB.
-
-1. **Cataloging**: Use `list-s3-objects` to record the meta-information (path, size, etc.) of files.
-2. **Understanding Structured Data**: Use `read-s3-object` to sample JSONL / CSV / TSV / Parquet / JSON.
-3. **Extracting Media Meta-information**: Use `read-s3-object` with `maxBytes` to read only the file header (without downloading the full file) for images, audio, video, and PDFs to extract key attributes.
-4. **Writing to LanceDB**: Use `write-lance-catalog` to save the results.
-
-## Output Location
-- LanceDB table storage path: `~/.openclaw/contextlake/profiler/{datasource_name}/catalog_db`
-- Table names: `files`, `structured_schemas`, `media_metadata`
-
-## Available Tools for this Skill
-- `list-s3-objects`: To traverse and list files in the bucket/directory.
-- `read-s3-object`: To read specific bytes of a file for schema inference or metadata extraction.
-- `write-lance-catalog`: To write the profiling results to the LanceDB catalog.
-
-Always report the final profiling summary back to the user once the `write-lance-catalog` completes successfully.