catalogs/wikipedia.json

{
  "_meta": {
    "generated": "2026-04-08",
    "source": "hardcoded — kiwix-serve binary serves a static ZIM file; no application source to analyze",
    "zim_file": "/data/wikipedia_en_all_maxi_2022-05.zim",
    "search_response": "HTML only — GET /search returns HTML page; agent must parse <a href> links for article URLs",
    "article_page": "GET /wikipedia_en_all_maxi_2022-05/A/{title} — returns HTML article",
    "websockets": "none"
  },
  "endpoints": [
    {
      "api_type": "rest",
      "endpoint": "GET /search",
      "auth": "none",
      "query_params": {
        "pattern": {
          "type": "string",
          "source": "TASK_SPEC",
          "notes": "the search query, URL-encoded"
        },
        "books.name": {
          "type": "string",
          "source": "STATIC",
          "value": "wikipedia_en_all_maxi_2022-05",
          "notes": "selects which ZIM book to search"
        }
      },
      "response_key_fields": [],
      "notes": "IMPORTANT: response is HTML, not JSON. Parse <a href> anchor links matching /wikipedia_en_all_maxi_2022-05/A/... to extract article slugs. The .results[0].url jq path does NOT apply — use HTML parsing."
    },
    {
      "api_type": "rest",
      "endpoint": "GET /wikipedia_en_all_maxi_2022-05/A/{article_title}",
      "auth": "none",
      "path_params": {
        "article_title": {
          "type": "string",
          "source": "PREV_CALL",
          "from_endpoint": "GET /search",
          "from_field": "href attribute of first search result <a> tag",
          "notes": "URL-encoded article slug, e.g. Albert_Einstein. Extract from the href on the search results HTML page. Verified live: HTTP 200 for valid titles."
        }
      },
      "response_key_fields": [],
      "notes": "Returns full HTML article page. HTTP 200 when article exists, 404 when not found."
    }
  ]
}