@hasna/skills 0.1.59 → 0.1.61

package/skills/extract/src/index-local.ts

@@ -1,258 +0,0 @@
-#!/usr/bin/env bun
-
-/**
- * Extraction CLI
- * Extract text and structured data from images and PDFs using OpenAI Vision
- */
-import { handleInstallCommand } from '../../_common';
-
-
-// Handle install/uninstall commands
-const SKILL_META = {
-  name: 'extract',
-  description: 'Extract text and structured data from images and PDFs using OpenAI Vision',
-  version: '1.0.0',
-  commands: `Use: extract --help`,
-  requiredEnvVars: [],
-};
-
-if (await handleInstallCommand(SKILL_META, process.argv.slice(2))) {
-  process.exit(0);
-}
-
-import { extractFromImage, isImageFile, getSupportedImageExtensions } from './extractors/image';
-import { extractFromPDF, isPDFFile } from './extractors/pdf';
-import type { ExtractOptions, OutputFormat } from './types';
-import { basename, extname, dirname, join } from 'path';
-
-// Parse command line arguments
-function parseArgs(): {
-  command: string;
-  input?: string;
-  output?: string;
-  format?: OutputFormat;
-  prompt?: string;
-  model?: string;
-  detail?: 'low' | 'high' | 'auto';
-} {
-  const args = process.argv.slice(2);
-  if (args[0] === '--help' || args[0] === '-h') args[0] = 'help';
-  const parsed: any = { command: args[0] || 'help' };
-
-  for (let i = 1; i < args.length; i++) {
-    const arg = args[i];
-    if (arg.startsWith('--')) {
-      const key = arg.replace(/^--/, '');
-      parsed[key] = args[++i];
-    } else if (arg.startsWith('-')) {
-      const key = arg.replace(/^-/, '');
-      // Short flags
-      switch (key) {
-        case 'i':
-          parsed['input'] = args[++i];
-          break;
-        case 'o':
-          parsed['output'] = args[++i];
-          break;
-        case 'f':
-          parsed['format'] = args[++i];
-          break;
-        case 'p':
-          parsed['prompt'] = args[++i];
-          break;
-        case 'm':
-          parsed['model'] = args[++i];
-          break;
-        case 'd':
-          parsed['detail'] = args[++i];
-          break;
-        case 'h':
-          parsed['command'] = 'help';
-          break;
-      }
-    }
-  }
-
-  return parsed;
-}
-
-// Display help information
-function showHelp(): void {
-  const imageExts = getSupportedImageExtensions().join(', ');
-
-  console.log(`
-Extraction CLI - Extract text and data from images and PDFs
-
-USAGE:
-  bun run src/index.ts <command> [options]
-
-COMMANDS:
-  extract       Extract text/data from an image or PDF
-  help          Show this help message
-
-EXTRACT OPTIONS:
-  --input, -i <path>      Input file path (image or PDF) [required]
-  --output, -o <path>     Output file path (optional, defaults to input name + format extension)
-  --format, -f <type>     Output format: text, markdown, json (default: text)
-  --prompt, -p <text>     Custom extraction prompt (optional)
-  --model, -m <model>     OpenAI model to use (default: gpt-4o)
-  --detail, -d <level>    Image detail level: low, high, auto (default: auto, images only)
-
-SUPPORTED INPUT FORMATS:
-  Images: ${imageExts}
-  Documents: .pdf
-
-EXAMPLES:
-  # Extract text from an image
-  bun run src/index.ts extract --input ./receipt.png --output ./receipt.txt
-
-  # Extract as Markdown from a PDF
-  bun run src/index.ts extract -i ./document.pdf -o ./document.md -f markdown
-
-  # Extract with custom prompt
-  bun run src/index.ts extract \\
-    --input ./invoice.png \\
-    --format json \\
-    --prompt "Extract invoice number, date, total amount, and line items"
-
-  # High-detail extraction for small text
-  bun run src/index.ts extract \\
-    --input ./handwriting.jpg \\
-    --detail high \\
-    --format text
-
-OUTPUT FORMATS:
-  text      - Clean, readable plain text
-  markdown  - Structured Markdown with headings, lists, and tables
-  json      - Structured JSON object with sections, tables, and metadata
-
-ENVIRONMENT VARIABLES:
-  OPENAI_API_KEY          API key for OpenAI Vision API (required)
-
-NOTES:
-  - Image extraction uses OpenAI GPT-4 Vision for accurate OCR
-  - PDF extraction uses native text parsing, with optional OpenAI cleanup
-  - Use --detail high for images with small text or fine details
-  - Use --detail low for faster processing of simple images
-`);
-}
-
-// Get output extension based on format
-function getOutputExtension(format: OutputFormat): string {
-  switch (format) {
-    case 'markdown':
-      return '.md';
-    case 'json':
-      return '.json';
-    case 'text':
-    default:
-      return '.txt';
-  }
-}
-
-// Detect input type
-function detectInputType(filePath: string): 'image' | 'pdf' | 'unknown' {
-  if (isImageFile(filePath)) return 'image';
-  if (isPDFFile(filePath)) return 'pdf';
-  return 'unknown';
-}
-
-// Main CLI handler
-async function main(): Promise<void> {
-  const args = parseArgs();
-
-  try {
-    switch (args.command) {
-      case 'extract': {
-        if (!args.input) {
-          console.error('Error: --input is required');
-          console.error('Use: bun run src/index.ts help');
-          process.exit(1);
-        }
-
-        const inputType = detectInputType(args.input);
-
-        if (inputType === 'unknown') {
-          console.error(`Error: Unsupported file type: ${extname(args.input)}`);
-          console.error('Supported: images (.png, .jpg, .jpeg, .gif, .webp) and PDFs (.pdf)');
-          process.exit(1);
-        }
-
-        const format = (args.format as OutputFormat) || 'text';
-
-        console.log('\n=== Extraction ===\n');
-        console.log(`Input: ${args.input}`);
-        console.log(`Type: ${inputType}`);
-        console.log(`Format: ${format}`);
-        if (args.prompt) {
-          console.log(`Prompt: ${args.prompt}`);
-        }
-        console.log('');
-
-        const options: ExtractOptions = {
-          input: args.input,
-          output: args.output,
-          format,
-          prompt: args.prompt,
-          model: args.model,
-          detail: args.detail as 'low' | 'high' | 'auto',
-        };
-
-        let result;
-
-        if (inputType === 'image') {
-          result = await extractFromImage(options);
-        } else {
-          result = await extractFromPDF(options);
-        }
-
-        // Determine output path
-        let outputPath = args.output;
-        if (!outputPath) {
-          const inputDir = dirname(args.input);
-          const inputName = basename(args.input, extname(args.input));
-          outputPath = join(inputDir, `${inputName}${getOutputExtension(format)}`);
-        }
-
-        // Write output
-        await Bun.write(outputPath, result.text);
-
-        console.log(`\nExtraction complete!`);
-        console.log(`Output: ${outputPath}`);
-        console.log(`Format: ${result.format}`);
-
-        if (result.metadata?.pages) {
-          console.log(`Pages: ${result.metadata.pages}`);
-        }
-        if (result.metadata?.model) {
-          console.log(`Model: ${result.metadata.model}`);
-        }
-        if (result.metadata?.tokens) {
-          console.log(`Tokens used: ${result.metadata.tokens}`);
-        }
-
-        // Show content stats
-        const wordCount = result.text.split(/\s+/).filter(Boolean).length;
-        const charCount = result.text.length;
-        console.log(`Characters: ${charCount.toLocaleString()}`);
-        console.log(`Words: ${wordCount.toLocaleString()}`);
-        break;
-      }
-
-      case 'help':
-        showHelp();
-        break;
-
-      default:
-        console.error(`Unknown command: ${args.command}`);
-        console.error('Use: bun run src/index.ts help');
-        process.exit(1);
-    }
-  } catch (error) {
-    console.error('\nError:', error instanceof Error ? error.message : String(error));
-    process.exit(1);
-  }
-}
-
-// Run the CLI
-main();

package/skills/extract/src/index.ts

@@ -1,4 +0,0 @@
-#!/usr/bin/env bun
-
-// Keep the package root on the local implementation; hosted execution goes through the root Skills CLI.
-import "./index-local";

package/skills/extract/src/providers/openai.ts

@@ -1,222 +0,0 @@
-import type { ExtractOptions, ExtractionResult, OpenAIVisionResponse, OutputFormat } from '../types';
-import { readFile } from 'fs/promises';
-import { extname } from 'path';
-
-export class OpenAIProvider {
-  private apiKey: string;
-  private baseUrl = 'https://api.openai.com/v1';
-
-  constructor(apiKey?: string) {
-    this.apiKey = apiKey || process.env.OPENAI_API_KEY || '';
-    if (!this.apiKey) {
-      throw new Error('OPENAI_API_KEY environment variable is required');
-    }
-  }
-
-  async extractFromImage(
-    imagePath: string,
-    options?: {
-      prompt?: string;
-      model?: string;
-      format?: OutputFormat;
-      detail?: 'low' | 'high' | 'auto';
-    }
-  ): Promise<ExtractionResult> {
-    const model = options?.model || 'gpt-4o';
-    const format = options?.format || 'text';
-    const detail = options?.detail || 'auto';
-
-    console.log(`Extracting from image with OpenAI ${model}...`);
-    console.log(`Image: ${imagePath}`);
-    console.log(`Detail: ${detail}`);
-
-    // Read and encode the image
-    const imageBuffer = await readFile(imagePath);
-    const base64Image = imageBuffer.toString('base64');
-    const mimeType = this.getMimeType(imagePath);
-
-    // Build the prompt based on format
-    const systemPrompt = this.buildSystemPrompt(format, options?.prompt);
-
-    const response = await fetch(`${this.baseUrl}/chat/completions`, {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-        Authorization: `Bearer ${this.apiKey}`,
-      },
-      body: JSON.stringify({
-        model,
-        messages: [
-          {
-            role: 'system',
-            content: systemPrompt,
-          },
-          {
-            role: 'user',
-            content: [
-              {
-                type: 'image_url',
-                image_url: {
-                  url: `data:${mimeType};base64,${base64Image}`,
-                  detail,
-                },
-              },
-              {
-                type: 'text',
-                text: options?.prompt || 'Extract all text and content from this image.',
-              },
-            ],
-          },
-        ],
-        max_tokens: 4096,
-      }),
-    });
-
-    if (!response.ok) {
-      const error = await response.text();
-      throw new Error(`OpenAI API error: ${response.status} - ${error}`);
-    }
-
-    const data: OpenAIVisionResponse = await response.json();
-
-    if (!data.choices || data.choices.length === 0) {
-      throw new Error('No response from OpenAI API');
-    }
-
-    const extractedText = data.choices[0].message.content;
-    console.log('Extraction complete!');
-
-    return {
-      text: extractedText,
-      format,
-      inputType: 'image',
-      inputFile: imagePath,
-      metadata: {
-        model,
-        tokens: data.usage?.total_tokens,
-      },
-    };
-  }
-
-  async extractFromPDFText(
-    pdfText: string,
-    options?: {
-      prompt?: string;
-      model?: string;
-      format?: OutputFormat;
-      pages?: number;
-    }
-  ): Promise<string> {
-    const model = options?.model || 'gpt-4o';
-    const format = options?.format || 'text';
-
-    console.log(`Processing PDF text with OpenAI ${model}...`);
-
-    const systemPrompt = this.buildSystemPrompt(format, options?.prompt);
-
-    const userPrompt = options?.prompt
-      ? `${options.prompt}\n\nDocument content:\n${pdfText}`
-      : `Clean up and structure the following extracted PDF text. Fix any OCR errors, formatting issues, and organize the content clearly.\n\nDocument content:\n${pdfText}`;
-
-    const response = await fetch(`${this.baseUrl}/chat/completions`, {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-        Authorization: `Bearer ${this.apiKey}`,
-      },
-      body: JSON.stringify({
-        model,
-        messages: [
-          {
-            role: 'system',
-            content: systemPrompt,
-          },
-          {
-            role: 'user',
-            content: userPrompt,
-          },
-        ],
-        max_tokens: 4096,
-      }),
-    });
-
-    if (!response.ok) {
-      const error = await response.text();
-      throw new Error(`OpenAI API error: ${response.status} - ${error}`);
-    }
-
-    const data: OpenAIVisionResponse = await response.json();
-
-    if (!data.choices || data.choices.length === 0) {
-      throw new Error('No response from OpenAI API');
-    }
-
-    return data.choices[0].message.content;
-  }
-
-  private buildSystemPrompt(format: OutputFormat, customPrompt?: string): string {
-    const basePrompt = customPrompt
-      ? `You are a document extraction assistant. Follow the user's specific instructions.`
-      : `You are a document extraction assistant. Extract all text, data, and content from the provided document or image accurately.`;
-
-    switch (format) {
-      case 'markdown':
-        return `${basePrompt}
-
-Output the extracted content in clean, well-structured Markdown format:
-- Use proper headings (# ## ###) for sections
-- Use lists (- or 1.) for enumerated items
-- Use tables for tabular data
-- Use code blocks for code or formatted text
-- Preserve the logical structure of the document`;
-
-      case 'json':
-        return `${basePrompt}
-
-Output the extracted content as a valid JSON object with the following structure:
-{
-  "title": "document title if found",
-  "sections": [
-    {
-      "heading": "section heading",
-      "content": "section content"
-    }
-  ],
-  "tables": [
-    {
-      "caption": "table caption if any",
-      "headers": ["col1", "col2"],
-      "rows": [["data1", "data2"]]
-    }
-  ],
-  "metadata": {
-    "any": "relevant metadata"
-  }
-}`;
-
-      case 'text':
-      default:
-        return `${basePrompt}
-
-Output the extracted content as clean, readable plain text:
-- Preserve paragraphs and logical structure
-- Use blank lines to separate sections
-- Maintain the reading order of the document`;
-    }
-  }
-
-  private getMimeType(filePath: string): string {
-    const ext = extname(filePath).toLowerCase();
-    const mimeTypes: Record<string, string> = {
-      '.png': 'image/png',
-      '.jpg': 'image/jpeg',
-      '.jpeg': 'image/jpeg',
-      '.gif': 'image/gif',
-      '.webp': 'image/webp',
-      '.bmp': 'image/bmp',
-      '.tiff': 'image/tiff',
-      '.tif': 'image/tiff',
-    };
-    return mimeTypes[ext] || 'image/png';
-  }
-}

package/skills/extract/src/types.ts

@@ -1,61 +0,0 @@
-export type InputType = 'image' | 'pdf';
-export type OutputFormat = 'text' | 'markdown' | 'json';
-
-export interface ExtractOptions {
-  input: string;
-  output?: string;
-  format?: OutputFormat;
-  prompt?: string;
-  model?: string;
-  detail?: 'low' | 'high' | 'auto';
-}
-
-export interface ExtractionResult {
-  text: string;
-  format: OutputFormat;
-  inputType: InputType;
-  inputFile: string;
-  metadata?: {
-    pages?: number;
-    model?: string;
-    tokens?: number;
-  };
-}
-
-export interface ExtractorProvider {
-  name: string;
-  extract(options: ExtractOptions): Promise<ExtractionResult>;
-}
-
-export interface OpenAIVisionResponse {
-  id: string;
-  object: string;
-  created: number;
-  model: string;
-  choices: Array<{
-    index: number;
-    message: {
-      role: string;
-      content: string;
-    };
-    finish_reason: string;
-  }>;
-  usage?: {
-    prompt_tokens: number;
-    completion_tokens: number;
-    total_tokens: number;
-  };
-}
-
-export interface PDFMetadata {
-  pages: number;
-  info?: {
-    Title?: string;
-    Author?: string;
-    Subject?: string;
-    Creator?: string;
-    Producer?: string;
-    CreationDate?: string;
-    ModDate?: string;
-  };
-}

package/skills/extract/tsconfig.json

@@ -1,19 +0,0 @@
-{
-  "extends": "../tsconfig.base.json",
-  "compilerOptions": {
-    "target": "ESNext",
-    "module": "ESNext",
-    "lib": [
-      "ESNext"
-    ],
-    "forceConsistentCasingInFileNames": true,
-    "resolveJsonModule": true,
-    "allowSyntheticDefaultImports": true
-  },
-  "include": [
-    "src/**/*"
-  ],
-  "exclude": [
-    "node_modules"
-  ]
-}

package/skills/extract-invoice/CLAUDE.md

@@ -1,19 +0,0 @@
-# extract-invoice
-
-Extract Invoice skill.
-
-## Tech Stack
-
-- Runtime: Bun
-- Language: TypeScript
-- CLI: Commander.js
-
-## CLI
-
-```bash
-extract-invoice run
-```
-
-## Data Directory
-
-~/.skill/extract-invoice/

package/skills/extract-invoice/package.json

@@ -1,27 +0,0 @@
-{
-  "name": "extract-invoice",
-  "version": "0.1.0",
-  "description": "Extract Invoice skill",
-  "type": "module",
-  "bin": {
-    "extract-invoice": "./src/index.ts"
-  },
-  "scripts": {
-    "dev": "bun run src/index.ts",
-    "build": "bun build src/index.ts --outdir dist --target node",
-    "typecheck": "tsc --noEmit"
-  },
-  "dependencies": {},
-  "devDependencies": {
-    "@types/bun": "latest",
-    "typescript": "^5.7.0"
-  },
-  "publishConfig": {
-    "access": "restricted",
-    "registry": "https://registry.npmjs.org/"
-  },
-  "files": [
-    "src",
-    "tsconfig.json"
-  ]
-}