All notable changes to the EU Public Tenders (TED) Scraper are listed here. The version number is
the version field of .actor/actor.json.
First release.
- Fetches procurement notices from the official TED Search API
(
POST https://api.ted.europa.eu/v3/notices/search, no API key) in iteration mode, so any
number of notices can be retrieved.
- Structured filters (
countries, cpvCodes, keywords, noticeTypes, publishedAfter,
publishedBefore, deadlineAfter, minValue, onlyOpenForSubmission) are compiled into a TED
expert query; a raw query can be given instead. The query is validated with TED's syntax
check before any data is fetched, so typos produce a clear error. Without any filter the
newest notices of the last 7 days are returned (TED rejects a query that matches everything).
- Output per notice: publication number, notice and form type, title (preferred language with
fallback and
titleLanguage), buyer (name, country, city, website, legal type), CPV codes,
place of performance, estimated and awarded values, publication date, earliest submission
deadline, procedure type, contract nature, lot count, HTML/PDF/XML/TED links, document URL,
language, scrape timestamp. fields: "full" adds the description, identifiers, OJ S number,
dispatch date, submission URL and the raw TED record.
- No personal data: only organisation-level fields are requested; contact, e-mail and phone
fields are never requested and are stripped from
raw.
- Options:
maxNotices (0 = all), sortBy, scope, preferredLanguage, includeDocumentLinks.
- Politeness and reliability: 2 concurrent requests at most, 200-500 ms jitter, exponential
backoff on 429/5xx (5 retries), automatic retry without a field TED rejects, de-duplication by
publication number, state persisted across migrations.
- Pay-per-event pricing: dataset items charged automatically, optional
actor-start event.
