@hasna/skills 0.1.59 → 0.1.61

package/skills/extract-invoice/src/cli.ts

@@ -1,97 +0,0 @@
-import { existsSync, readdirSync } from "fs";
-import { extname, join, resolve } from "path";
-import { parseArgs } from "util";
-import { SUPPORTED_EXTENSIONS } from "./file-utils";
-import type { Options } from "./types";
-
-export function parseArguments(): Options {
-  const { values, positionals } = parseArgs({
-    args: process.argv.slice(2),
-    options: {
-      format: { type: "string", default: "json" },
-      confidence: { type: "boolean", default: false },
-      currency: { type: "string" },
-      language: { type: "string" },
-      pages: { type: "string" },
-      batch: { type: "boolean", default: false },
-      output: { type: "string" },
-      verbose: { type: "boolean", default: false },
-      help: { type: "boolean", short: "h" },
-    },
-    allowPositionals: true,
-  });
-
-  if (values.help) {
-    printHelp();
-    process.exit(0);
-  }
-
-  const files: string[] = [];
-  for (const arg of positionals) {
-    if (arg.includes("*")) {
-      files.push(...expandGlob(arg));
-    } else {
-      files.push(resolve(process.cwd(), arg));
-    }
-  }
-
-  return {
-    files,
-    format: values.format as Options["format"],
-    confidence: values.confidence as boolean,
-    currency: values.currency as string,
-    language: values.language as string,
-    pages: values.pages as string,
-    batch: values.batch as boolean,
-    output: values.output as string,
-    verbose: values.verbose as boolean,
-  };
-}
-
-function expandGlob(pattern: string): string[] {
-  const dir = pattern.includes("/") ? pattern.substring(0, pattern.lastIndexOf("/")) : ".";
-  const filePattern = pattern.includes("/") ? pattern.substring(pattern.lastIndexOf("/") + 1) : pattern;
-  const regex = new RegExp("^" + filePattern.replace(/\*/g, ".*") + "$");
-
-  const resolvedDir = resolve(process.cwd(), dir);
-  if (!existsSync(resolvedDir)) {
-    return [];
-  }
-
-  return readdirSync(resolvedDir)
-    .filter((file) => regex.test(file))
-    .map((file) => join(resolvedDir, file))
-    .filter((file) => SUPPORTED_EXTENSIONS.includes(extname(file).toLowerCase()));
-}
-
-function printHelp(): void {
-  console.log(`
-Extract Invoice - AI-powered invoice data extraction
-
-Usage:
-  skills run extract-invoice -- <file(s)> [options]
-
-Arguments:
-  file(s)               Invoice file(s) to process (PDF, PNG, JPG, etc.)
-
-Options:
-  --format <format>     Output format: json, csv, excel, markdown (default: json)
-  --confidence          Include confidence scores for extracted fields
-  --currency <code>     Convert totals to specified currency (e.g., USD, EUR)
-  --language <lang>     Document language hint (e.g., en, es, fr)
-  --pages <range>       Pages to process (e.g., "1-3,5")
-  --batch               Process multiple files and combine output
-  --output <file>       Output file path
-  --verbose             Show detailed progress
-  --help, -h            Show this help
-
-Examples:
-  skills run extract-invoice -- invoice.pdf
-  skills run extract-invoice -- receipt.jpg --format csv
-  skills run extract-invoice -- ./invoices/*.pdf --batch
-  skills run extract-invoice -- invoice.pdf --confidence --verbose
-
-Supported file types:
-  PDF, PNG, JPG, JPEG, TIFF, HEIC, WebP
-`);
-}

package/skills/extract-invoice/src/extractor.ts

@@ -1,170 +0,0 @@
-import { basename } from "path";
-import { fileToBase64, getMimeType } from "./file-utils";
-import { OPENAI_API_URL, getOpenAIApiKey, log } from "./runtime";
-import type { InvoiceData, Options } from "./types";
-
-export async function extractInvoiceData(filePath: string, options: Options): Promise<InvoiceData> {
-  const openAIApiKey = getOpenAIApiKey();
-  if (!openAIApiKey) {
-    throw new Error("OPENAI_API_KEY environment variable is not set");
-  }
-
-  const base64Content = fileToBase64(filePath);
-  const mimeType = getMimeType(filePath);
-
-  if (options.verbose) {
-    log(`Processing ${basename(filePath)} (${mimeType})...`);
-  }
-
-  const response = await fetch(OPENAI_API_URL, {
-    method: "POST",
-    headers: {
-      "Content-Type": "application/json",
-      Authorization: `Bearer ${openAIApiKey}`,
-    },
-    body: JSON.stringify(buildRequestBody(base64Content, mimeType, options)),
-  });
-
-  if (!response.ok) {
-    const error = await response.text();
-    throw new Error(`OpenAI API error (${response.status}): ${error}`);
-  }
-
-  const result = await response.json();
-  const content = result.choices[0]?.message?.content;
-
-  if (!content) {
-    throw new Error("No response from OpenAI API");
-  }
-
-  try {
-    return JSON.parse(stripJsonFence(content));
-  } catch {
-    log(`Failed to parse API response: ${content}`, "error");
-    throw new Error("Failed to parse invoice data from API response");
-  }
-}
-
-function buildRequestBody(base64Content: string, mimeType: string, options: Options): Record<string, unknown> {
-  const systemPrompt = `You are an expert invoice data extraction system. Extract all relevant information from the provided invoice image/document and return it as structured JSON.
-
-Be thorough and extract:
-- Invoice details (number, dates, PO number, payment terms)
-- Vendor information (name, address, contact, tax ID)
-- Customer/Bill-to information
-- All line items with quantities, prices, and totals
-- Tax calculations
-- Discounts if any
-- Grand total and currency
-- Payment information (bank details, references)
-- Any notes or special instructions
-
-For amounts, extract numeric values without currency symbols.
-For dates, use ISO format (YYYY-MM-DD) when possible.
-Estimate your confidence in the extraction (0-100).
-
-${options.language ? `The document is in ${options.language}.` : "Detect the document language automatically."}
-${options.currency ? `Convert all amounts to ${options.currency} if possible.` : "Detect and preserve the original currency."}`;
-
-  const userPrompt = `Extract all invoice data from this document and return it as JSON matching this structure:
-{
-  "invoice_number": "string or null",
-  "invoice_date": "YYYY-MM-DD or null",
-  "due_date": "YYYY-MM-DD or null",
-  "po_number": "string or null",
-  "payment_terms": "string or null",
-  "vendor": {
-    "name": "string or null",
-    "address": "string or null",
-    "phone": "string or null",
-    "email": "string or null",
-    "tax_id": "string or null",
-    "website": "string or null"
-  },
-  "customer": {
-    "name": "string or null",
-    "address": "string or null",
-    "phone": "string or null",
-    "email": "string or null"
-  },
-  "line_items": [
-    {
-      "description": "string",
-      "quantity": number or null,
-      "unit": "string or null",
-      "unit_price": number or null,
-      "tax_rate": number or null,
-      "subtotal": number or null,
-      "discount": number or null
-    }
-  ],
-  "subtotal": number or null,
-  "tax": {
-    "rate": number or null,
-    "amount": number,
-    "type": "string or null"
-  },
-  "discount": {
-    "type": "percentage or fixed or null",
-    "value": number or null,
-    "amount": number
-  },
-  "shipping": number or null,
-  "total": number or null,
-  "currency": "USD, EUR, etc.",
-  "payment_info": {
-    "method": "string or null",
-    "bank_name": "string or null",
-    "account_number": "string or null (masked if sensitive)",
-    "routing_number": "string or null",
-    "iban": "string or null (masked)",
-    "swift": "string or null",
-    "reference": "string or null"
-  },
-  "notes": "string or null",
-  "confidence": number (0-100)
-}
-
-Return ONLY the JSON object, no markdown formatting or explanation.`;
-
-  return {
-    model: "gpt-4o",
-    messages: [
-      {
-        role: "system",
-        content: systemPrompt,
-      },
-      {
-        role: "user",
-        content: [
-          {
-            type: "text",
-            text: userPrompt,
-          },
-          {
-            type: "image_url",
-            image_url: {
-              url: `data:${mimeType};base64,${base64Content}`,
-              detail: "high",
-            },
-          },
-        ],
-      },
-    ],
-    max_tokens: 4096,
-    temperature: 0.1,
-  };
-}
-
-function stripJsonFence(content: string): string {
-  let jsonString = content.trim();
-  if (jsonString.startsWith("```json")) {
-    jsonString = jsonString.slice(7);
-  } else if (jsonString.startsWith("```")) {
-    jsonString = jsonString.slice(3);
-  }
-  if (jsonString.endsWith("```")) {
-    jsonString = jsonString.slice(0, -3);
-  }
-  return jsonString.trim();
-}

package/skills/extract-invoice/src/file-utils.ts

@@ -1,46 +0,0 @@
-import { existsSync, readFileSync } from "fs";
-import { extname } from "path";
-import { log } from "./runtime";
-
-export const SUPPORTED_EXTENSIONS = [".pdf", ".png", ".jpg", ".jpeg", ".tiff", ".tif", ".heic", ".webp"];
-
-export function validateFiles(files: string[]): string[] {
-  const validFiles: string[] = [];
-
-  for (const file of files) {
-    if (!existsSync(file)) {
-      log(`File not found: ${file}`, "warn");
-      continue;
-    }
-
-    const ext = extname(file).toLowerCase();
-    if (!SUPPORTED_EXTENSIONS.includes(ext)) {
-      log(`Unsupported file type: ${file}`, "warn");
-      continue;
-    }
-
-    validFiles.push(file);
-  }
-
-  return validFiles;
-}
-
-export function fileToBase64(filePath: string): string {
-  const buffer = readFileSync(filePath);
-  return buffer.toString("base64");
-}
-
-export function getMimeType(filePath: string): string {
-  const ext = extname(filePath).toLowerCase();
-  const mimeTypes: Record<string, string> = {
-    ".pdf": "application/pdf",
-    ".png": "image/png",
-    ".jpg": "image/jpeg",
-    ".jpeg": "image/jpeg",
-    ".tiff": "image/tiff",
-    ".tif": "image/tiff",
-    ".heic": "image/heic",
-    ".webp": "image/webp",
-  };
-  return mimeTypes[ext] || "application/octet-stream";
-}

package/skills/extract-invoice/src/formatters.ts

@@ -1,144 +0,0 @@
-import { basename } from "path";
-import type { ExtractionResult, Options } from "./types";
-
-export function formatAsJSON(results: ExtractionResult[], options: Options): string {
-  if (results.length === 1 && !options.batch) {
-    const result = results[0];
-    if (result.success && result.data) {
-      if (!options.confidence) {
-        const { confidence, ...dataWithoutConfidence } = result.data;
-        return JSON.stringify(dataWithoutConfidence, null, 2);
-      }
-      return JSON.stringify(result.data, null, 2);
-    }
-    return JSON.stringify({ error: result.error }, null, 2);
-  }
-
-  const output = {
-    processed_at: new Date().toISOString(),
-    total_files: results.length,
-    successful: results.filter((result) => result.success).length,
-    failed: results.filter((result) => !result.success).length,
-    results: results.map((result) => ({
-      file: basename(result.file),
-      success: result.success,
-      data: result.data,
-      error: result.error,
-      processing_time_ms: result.processing_time_ms,
-    })),
-  };
-
-  return JSON.stringify(output, null, 2);
-}
-
-export function formatAsCSV(results: ExtractionResult[]): string {
-  const headers = [
-    "file",
-    "invoice_number",
-    "invoice_date",
-    "due_date",
-    "vendor_name",
-    "vendor_address",
-    "customer_name",
-    "subtotal",
-    "tax_amount",
-    "total",
-    "currency",
-    "status",
-  ];
-
-  const rows: string[][] = [headers];
-
-  for (const result of results) {
-    if (result.success && result.data) {
-      const data = result.data;
-      rows.push([
-        basename(result.file),
-        data.invoice_number || "",
-        data.invoice_date || "",
-        data.due_date || "",
-        data.vendor.name || "",
-        data.vendor.address || "",
-        data.customer.name || "",
-        data.subtotal?.toString() || "",
-        data.tax?.amount?.toString() || "",
-        data.total?.toString() || "",
-        data.currency || "",
-        "success",
-      ]);
-    } else {
-      rows.push([basename(result.file), "", "", "", "", "", "", "", "", "", "", `error: ${result.error}`]);
-    }
-  }
-
-  return rows.map((row) => row.map((cell) => `"${cell.replace(/"/g, '""')}"`).join(",")).join("\n");
-}
-
-export function formatAsMarkdown(results: ExtractionResult[], options: Options): string {
-  let output = "# Invoice Extraction Results\n\n";
-  output += `Processed: ${new Date().toISOString()}\n\n`;
-
-  for (const result of results) {
-    output += "---\n\n";
-    output += `## ${basename(result.file)}\n\n`;
-
-    if (!result.success) {
-      output += `**Error:** ${result.error}\n\n`;
-      continue;
-    }
-
-    const data = result.data!;
-
-    output += "### Invoice Details\n\n";
-    output += "| Field | Value |\n";
-    output += "|-------|-------|\n";
-    output += `| Invoice Number | ${data.invoice_number || "N/A"} |\n`;
-    output += `| Date | ${data.invoice_date || "N/A"} |\n`;
-    output += `| Due Date | ${data.due_date || "N/A"} |\n`;
-    output += `| PO Number | ${data.po_number || "N/A"} |\n`;
-    output += `| Payment Terms | ${data.payment_terms || "N/A"} |\n\n`;
-
-    output += "### Vendor\n\n";
-    output += `**${data.vendor.name || "Unknown Vendor"}**\n`;
-    if (data.vendor.address) output += `${data.vendor.address}\n`;
-    if (data.vendor.email) output += `Email: ${data.vendor.email}\n`;
-    if (data.vendor.phone) output += `Phone: ${data.vendor.phone}\n`;
-    if (data.vendor.tax_id) output += `Tax ID: ${data.vendor.tax_id}\n`;
-    output += "\n";
-
-    if (data.customer.name) {
-      output += "### Customer\n\n";
-      output += `**${data.customer.name}**\n`;
-      if (data.customer.address) output += `${data.customer.address}\n`;
-      if (data.customer.email) output += `Email: ${data.customer.email}\n`;
-      output += "\n";
-    }
-
-    if (data.line_items.length > 0) {
-      output += "### Line Items\n\n";
-      output += "| Description | Qty | Unit Price | Subtotal |\n";
-      output += "|-------------|-----|------------|----------|\n";
-      for (const item of data.line_items) {
-        output += `| ${item.description} | ${item.quantity || "-"} ${item.unit || ""} | ${item.unit_price?.toFixed(2) || "-"} | ${item.subtotal?.toFixed(2) || "-"} |\n`;
-      }
-      output += "\n";
-    }
-
-    output += "### Totals\n\n";
-    if (data.subtotal) output += `- **Subtotal:** ${data.currency} ${data.subtotal.toFixed(2)}\n`;
-    if (data.tax) output += `- **Tax${data.tax.rate ? ` (${data.tax.rate}%)` : ""}:** ${data.currency} ${data.tax.amount.toFixed(2)}\n`;
-    if (data.discount) output += `- **Discount:** -${data.currency} ${data.discount.amount.toFixed(2)}\n`;
-    if (data.shipping) output += `- **Shipping:** ${data.currency} ${data.shipping.toFixed(2)}\n`;
-    output += `- **Total:** ${data.currency} ${data.total?.toFixed(2) || "N/A"}\n\n`;
-
-    if (data.notes) {
-      output += `### Notes\n\n${data.notes}\n\n`;
-    }
-
-    if (options.confidence) {
-      output += `*Extraction confidence: ${data.confidence}%*\n\n`;
-    }
-  }
-
-  return output;
-}

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

@@ -1,129 +0,0 @@
-#!/usr/bin/env bun
-
-import { writeFileSync } from "fs";
-import { basename, join, resolve } from "path";
-import { parseArguments } from "./cli";
-import { EXPORTS_DIR, SESSION_ID, SKILL_NAME, ensureDir, getOpenAIApiKey, log } from "./runtime";
-import { extractInvoiceData } from "./extractor";
-import { formatAsCSV, formatAsJSON, formatAsMarkdown } from "./formatters";
-import { validateFiles } from "./file-utils";
-import type { ExtractionResult } from "./types";
-
-async function main(): Promise<void> {
-  const startTime = Date.now();
-
-  log(`Starting ${SKILL_NAME} session: ${SESSION_ID}`);
-
-  if (!getOpenAIApiKey()) {
-    console.error("Error: OPENAI_API_KEY environment variable is required.");
-    console.error("Get your API key from https://platform.openai.com/api-keys");
-    process.exit(1);
-  }
-
-  const options = parseArguments();
-
-  if (options.verbose) {
-    log(`Options: ${JSON.stringify({ ...options, files: `${options.files.length} file(s)` })}`);
-  }
-
-  if (options.files.length === 0) {
-    console.error("Error: Please specify at least one invoice file.");
-    console.error("\nExample: skills run extract-invoice -- invoice.pdf");
-    process.exit(1);
-  }
-
-  const validFiles = validateFiles(options.files);
-  if (validFiles.length === 0) {
-    console.error("Error: No valid files to process.");
-    process.exit(1);
-  }
-
-  log(`Processing ${validFiles.length} file(s)...`);
-
-  const results: ExtractionResult[] = [];
-
-  for (const file of validFiles) {
-    const fileStartTime = Date.now();
-    log(`Extracting data from ${basename(file)}...`);
-
-    try {
-      const data = await extractInvoiceData(file, options);
-      results.push({
-        file,
-        success: true,
-        data,
-        processing_time_ms: Date.now() - fileStartTime,
-      });
-      log(`Successfully extracted data from ${basename(file)}`, "success");
-    } catch (error) {
-      const errorMessage = error instanceof Error ? error.message : String(error);
-      results.push({
-        file,
-        success: false,
-        data: null,
-        error: errorMessage,
-        processing_time_ms: Date.now() - fileStartTime,
-      });
-      log(`Failed to extract data from ${basename(file)}: ${errorMessage}`, "error");
-    }
-  }
-
-  const { output, extension } = formatResults(results, options);
-
-  ensureDir(EXPORTS_DIR);
-
-  const outputFilename = options.output
-    ? resolve(process.cwd(), options.output)
-    : join(EXPORTS_DIR, `invoice-data.${extension}`);
-
-  writeFileSync(outputFilename, output, "utf-8");
-  log(`Output saved to: ${outputFilename}`, "success");
-
-  printSummary(results, output, outputFilename, options.batch);
-
-  const duration = Date.now() - startTime;
-  log(`Completed in ${duration}ms`, "success");
-}
-
-function formatResults(
-  results: ExtractionResult[],
-  options: ReturnType<typeof parseArguments>,
-): { output: string; extension: string } {
-  switch (options.format) {
-    case "csv":
-      return { output: formatAsCSV(results), extension: "csv" };
-    case "excel":
-      log("Excel format outputs as CSV (compatible with Excel)", "info");
-      return { output: formatAsCSV(results), extension: "csv" };
-    case "markdown":
-      return { output: formatAsMarkdown(results, options), extension: "md" };
-    default:
-      return { output: formatAsJSON(results, options), extension: "json" };
-  }
-}
-
-function printSummary(
-  results: ExtractionResult[],
-  output: string,
-  outputFilename: string,
-  isBatch: boolean,
-): void {
-  console.log("\n" + "=".repeat(60) + "\n");
-
-  const successful = results.filter((result) => result.success).length;
-  const failed = results.filter((result) => !result.success).length;
-
-  console.log("Extraction Summary:");
-  console.log(`  Total files: ${results.length}`);
-  console.log(`  Successful: ${successful}`);
-  console.log(`  Failed: ${failed}`);
-  console.log("");
-
-  if (results.length === 1 && !isBatch) {
-    console.log(output);
-  } else {
-    console.log(`Results saved to: ${outputFilename}`);
-  }
-}
-
-main();

package/skills/extract-invoice/src/runtime.ts

@@ -1,46 +0,0 @@
-import { appendFileSync, existsSync, mkdirSync } from "fs";
-import { randomUUID } from "crypto";
-import { join } from "path";
-
-export const SKILL_NAME = "extract-invoice";
-export const SESSION_ID = randomUUID().slice(0, 8);
-export const SESSION_TIMESTAMP = new Date()
-  .toISOString()
-  .replace(/[:.]/g, "_")
-  .replace(/-/g, "_")
-  .slice(0, 19)
-  .toLowerCase();
-
-export const SKILLS_OUTPUT_DIR = process.env.SKILLS_OUTPUT_DIR || join(process.cwd(), ".skills");
-export const EXPORTS_DIR = join(SKILLS_OUTPUT_DIR, "exports", SKILL_NAME, SESSION_ID);
-export const LOGS_DIR = join(SKILLS_OUTPUT_DIR, "logs", SKILL_NAME);
-export const OPENAI_API_URL = "https://api.openai.com/v1/chat/completions";
-
-export function getOpenAIApiKey(): string | undefined {
-  return process.env.OPENAI_API_KEY;
-}
-
-export function ensureDir(dir: string): void {
-  if (!existsSync(dir)) {
-    mkdirSync(dir, { recursive: true });
-  }
-}
-
-export function log(message: string, level: "info" | "error" | "success" | "warn" = "info"): void {
-  const timestamp = new Date().toISOString();
-  const logFile = join(LOGS_DIR, `log_${SESSION_TIMESTAMP}_${SESSION_ID}.log`);
-
-  ensureDir(LOGS_DIR);
-
-  const logEntry = `[${timestamp}] [${level.toUpperCase()}] ${message}\n`;
-  appendFileSync(logFile, logEntry);
-
-  const prefixes: Record<string, string> = {
-    info: "[INFO]",
-    error: "[ERROR]",
-    success: "[SUCCESS]",
-    warn: "[WARN]",
-  };
-
-  console.log(`${prefixes[level]} ${message}`);
-}