Search

One request that looks through everything the public site publishes: the register of legal acts (title, direction and document number), financing programmes (title, summary), the FAQ base (question and answer) and news (title, intro). Backs the header search box and the /search page.

The same matching also powers every list endpoint's filter[search] (/legal-documents, /programs, /faqs, /posts), so a term behaves the same wherever it is typed.

#How a term is matched

• Case-insensitive substring match (ILIKE).
• The term is tried in every supported locale of the row, not only the negotiated one — a row that carries only uz text is still found by a uz_cyrl reader.
• The term is also tried in the other Uzbek script: «пахта» finds "paxta" and "paxta" finds «пахта».
• Apostrophes (', ‘, ’, ʻ, ʼ) are ignored on both sides.
• Rows are still subject to the usual visibility rules (active catalogue rows, published and live posts in the current locale).

#GET /api/v1/search

Public, throttled 60/min.

{
  "data": {
    "query": "paxta",
    "total": 7,
    "groups": [
      { "kind": "legal_documents", "count": 2, "items": [ /* LegalDocument */ ] },
      { "kind": "programs",        "count": 1, "items": [ /* Program, no terms */ ] },
      { "kind": "faqs",            "count": 3, "items": [ /* Faq with category */ ] },
      { "kind": "posts",           "count": 1, "items": [ /* Post list item */ ] }
    ]
  }
}

Items use the same shapes as the corresponding list endpoints in Catalog and News. Groups always come in this order; a group with no hits has count: 0 and an empty items.

When the term contains digits and a legal document's number equals those digits exactly, that document is placed first in its group.

count is the number of items returned for the group (capped by limit), not the total number of matches — link to the section's own list with filter[search] for the full result.