helix-query.yaml

version: 1

indices:
  raw_index:
    source: html
    fetch: https://{repo}-{owner}.project-helix.page/{path}
    include:
      - (en|de|fr)/publish/**/*.(md|docx)
    target: https://adobe.sharepoint.com/:x:/r/sites/TheBlog/Shared%20Documents/theblog/en/query-index.xlsx?d=we7bf6b3af3234076968b30a1565f2373&csf=1&web=1&e=q9o8tW
    sitemap: en/query-index.json
    properties:
      author:
        select: main > div:nth-of-type(3) > p:nth-of-type(1)
        value: |
          match(el, 'by (.*)')
      title:
        select: main > div > h1:first-of-type
        value: |
          textContent(el)
      date:
        select: main > div:nth-of-type(3)
        value: |
          parseTimestamp(el, '[POSTED ON] MM-DD-YYYY')
      topics:
        select: main > div:last-of-type > p:first-of-type
        values: |
          match(el, '(Topics: )? ([^,]+)')
      products:
        select: main > div:last-of-type > p:nth-of-type(2)
        values: |
          match(el, '(Products: )? ([^,]+)')
      hero:
        select: main > div:nth-of-type(2) img:first-of-type
        value: |
          attribute(el, 'src')
      teaser:
        select: main > div:nth-child(n+4):not(:last-child) p
        value: |
          words(textContent(el), 0, 20)
      sourceHash:
        select: head > meta[name=x-source-hash
        value: |
          attribute(el, 'content')
      external-path:
        select: none
        value: |
          replace(path, '/publish/', '/')
    queries:
      recent:
        query: "*"
        hitsPerPage: 12
        parameters:
          - language
          - context
          - date
        filters: |
          parents:"${context}${language}" AND date < ${date}
        cache: 600
      recent-by-author:
        query: "*"
        hitsPerPage: 12
        parameters:
          - language
          - context
          - date
          - author
        filters: |
          parents:"${context}${language}" AND date < ${date} AND author:"${author}"
        cache: 600
      recent-by-topic:
        query: "*"
        hitsPerPage: 12
        parameters:
          - language
          - context
          - date
          - topic
        filters: |
          parents:"${context}${language}" AND date < ${date} AND topics:"${topic}"
        cache: 600