RU Procurement Toolkit

Framework-neutral TypeScript utilities for Russian procurement and tender products.

This repository contains safe, generic helpers for EIS/Zakupki links, purchase numbers, Russian money/date parsing and document prioritization. It intentionally does not include AI prompts, commercial scoring logic, customer data or TenderCheck/TenderCRM proprietary report logic.

Why

Russian procurement products repeatedly need the same low-level building blocks:

• parse EIS/Zakupki URLs;
• extract purchase numbers from links and text;
• normalize Russian ruble amounts;
• detect document types from filenames;
• rank tender documents before parsing;
• keep public tender tooling typed and reusable.

This toolkit provides those small pieces as open-source infrastructure.

Install

bun add ru-procurement-toolkit
# or
npm install ru-procurement-toolkit

Package publishing is planned. For now, use directly from GitHub or as a workspace dependency.

Usage

import {
  extractPurchaseNumber,
  parseEisUrl,
  parseRubAmount,
  prioritizeTenderDocuments
} from 'ru-procurement-toolkit';

const url =
  'https://zakupki.gov.ru/epz/order/notice/ea20/view/common-info.html?regNumber=0373100001026000001';

const parsed = parseEisUrl(url);
// { source: 'zakupki.gov.ru', purchaseNumber: '0373100001026000001', ... }

const amount = parseRubAmount('1 250 000,50 руб.');
// 1250000.5

const docs = prioritizeTenderDocuments([
  { name: 'Проект контракта.docx', url: 'https://example.test/contract.docx' },
  { name: 'Техническое задание.pdf', url: 'https://example.test/tz.pdf' },
  { name: 'Архив.zip', url: 'https://example.test/archive.zip' }
]);

Package Boundary

flowchart LR
    A[Raw public procurement input] --> B[URL and number parsing]
    A --> C[Money/date normalization]
    A --> D[Document type detection]
    D --> E[Document priority ranking]
    B --> F[Your product]
    C --> F
    E --> F

    X[AI prompts] -. not included .-> F
    Y[Commercial scoring] -. not included .-> F
    Z[Customer reports] -. not included .-> F

Design Principles

• No secrets.
• No network calls.
• No AI prompts.
• No customer data.
• No TenderCRM/TenderCheck proprietary logic.
• Small, typed, testable functions.
• Works in Node.js, Bun and browser-compatible bundles where possible.

API

parseEisUrl(input)

Parses supported EIS/Zakupki URLs and extracts source, purchase number and known query parameters.

extractPurchaseNumber(input)

Extracts a Russian purchase number from URL or free text.

parseRubAmount(input)

Parses common Russian ruble amount formats into a number.

detectTenderDocumentType(name)

Detects generic document type from a public filename: contract, technical specification, notice, protocol, estimate, attachment, archive or unknown.

prioritizeTenderDocuments(documents, options?)

Sorts public tender documents by generic usefulness before parsing.

Architecture

classDiagram
    class EisUrlParser {
      +parseEisUrl(input) ParsedEisUrl
      +extractPurchaseNumber(input) string | null
    }

    class Money {
      +parseRubAmount(input) number | null
      +formatRubAmount(value) string
    }

    class Documents {
      +detectTenderDocumentType(name) TenderDocumentType
      +prioritizeTenderDocuments(docs, options) RankedTenderDocument[]
    }

    EisUrlParser --> Money : independent
    EisUrlParser --> Documents : independent

What This Is Not

This is not:

• a tender monitoring service;
• an AI tender auditor;
• a parser for private user documents;
• a legal-risk engine;
• a TenderCRM or TenderCheck AI code dump.

Commercial AI analysis, prompts, fit scoring and report generation remain private.

Author

Прянишников Артём Алексеевич Saratov, Russia GitHub: @FrankFMY Telegram: @FrankFMY Email: [email protected] LinkedIn: frankfmy X/Twitter: @FrankFMY

License

Apache-2.0. See LICENSE.