fix applying 'is_user_defined' and 'is_visible' in createAttributesListQuery

[gibkigonzo]

Short Description and Why It's Useful

Right now when we make search for attributes we build invalid query (createAttributesListQuery). 'is_user_defined' and 'is_visible' can also have '1' as value, not only true.

Which Environment This Relates To

Check your case. In case of any doubts please read about Release Cycle

• Test version (https://test.storefrontcloud.io) - this is a new feature or improvement for Vue Storefront. I've created branch from develop branch and want to merge it back to develop
• RC version (https://next.storefrontcloud.io) - this is a stabilisation fix for Release Candidate of Vue Storefront. I've created branch from release branch and want to merge it back to release
• Stable version (https://demo.storefrontcloud.io) - this is an important fix for current stable version. I've created branch from hotfix or master branch and want to merge it back to hotfix

Upgrade Notes and Changelog

• No upgrade steps required (100% backward compatibility and no breaking changes)
• I've updated the Upgrade notes and Changelog on how to port existing Vue Storefront sites with this new feature

IMPORTANT NOTICE - Remember to update CHANGELOG.md with description of your change

Contribution and Currently Important Rules Acceptance

• I read and followed contribution rules

[flancer64]

I have an error

{
  "type": "illegal_argument_exception",
  "reason": "Can't parse boolean value [1], expected [true] or [false]"
}

for query (Elasticsearch version: "7.6.1"):

{
  "bool": {
    "filter": [
      {
        "bool": {
          "must": [
            {
              "terms": {
                "attribute_code": ["..."],
                "boost": 1.0
              }
            },
            {
              "terms": {
                "is_user_defined": [true, "1"],
                "boost": 1.0
              }
            },
            {
              "terms": {
                "is_visible": [true, "1"],
                "boost": 1.0
              }
            }
          ],
          "adjust_pure_negative": true,
          "boost": 1.0
        }
      }
    ],
    "adjust_pure_negative": true,
    "boost": 1.0
  }
}

This is data types mapping for my index of attributes:

{
  "mappings": {
    "_doc": {
      "properties": {
        "frontend_label": {"type": "text", "fields": {"keyword": {"ignore_above": 256, "type": "keyword"}}},
        "is_visible": {"type": "boolean"},
        "is_visible_on_front": {"type": "boolean"},
        "attribute_id": {"type": "long"},
        "options": {
          "properties": {
            "label": {"type": "text", "fields": {"keyword": {"ignore_above": 256, "type": "keyword"}}},
            "value": {"type": "text"}
          }
        },
        "id": {"type": "long"},
        "frontend_input": {"type": "text", "fields": {"keyword": {"ignore_above": 256, "type": "keyword"}}},
        "is_user_defined": {"type": "boolean"},
        "attribute_code": {"type": "text", "fields": {"keyword": {"ignore_above": 256, "type": "keyword"}}},
        "is_comparable": {"type": "boolean"}
      }
    }
  }
}

I'm nub in Elasticsearch, probably there is some configuration option to ignore data types in queries but I don't know. Just share my case.

[flancer64]

I see that Elasticsearch version 5.6 allows different values for boolean:

False values | false, "false", "off", "no", "0", "" (empty string), 0, 0.0
-- | --
True values | Anything that isn’t false.

but version 6.0 and above don't:

False values | false, "false"
-- | --
True values | true, "true"

I suppose we should revert these changes or make query more smart (add ES version dependency).

[mwalol]

hi,

i'm using elasticsearch 7.3 and i had to change :

  if (onlyDefinedByUser) {
    searchQuery = searchQuery.applyFilter({ key: 'is_user_defined', value: { 'in': [true] } })
  }
  if (onlyVisible) {
    searchQuery = searchQuery.applyFilter({ key: 'is_visible', value: { 'in': [true] } })
  }

instead of [true, '1'] to fix the issue a more generic solution would be welcome

[SDLyu]

Hi @gibkigonzo
This is not working in ElasticSearch 7.
I get error

{
  "type": "illegal_argument_exception",
  "reason": "Can't parse boolean value [1], expected [true] or [false]"
}

[gepdev]

Hi @gibkigonzo
This is not working in ElasticSearch 7.
I get error

{
  "type": "illegal_argument_exception",
  "reason": "Can't parse boolean value [1], expected [true] or [false]"
}

I have the same error with ES 7