No results once implementing an analyzer in Elasticsearch - php

Question

I am needing to ignore the apostrophe with indexed results so that searching for "Johns potato" will show results for "John's potato"
I was able to get the analyzer accepted but now I return no search results. Does anyone see something obvious that I am missing?
$params = [
'index' => $index,
'body' => [
'settings' => [
'number_of_shards' => 5,
'number_of_replicas' => 2,
'analysis' => [
"analyzer" => [
"my_analyzer" => [
"tokenizer" => "keyword",
"char_filter" => [
"my_char_filter"
]
]
],
"char_filter" => [
"my_char_filter" => [
"type" => "mapping",
"mappings" => [
"' => "
]
]
]
]
],
'mappings' => [
$type => [
'_source' => [
'enabled' => true
],
'properties' => [
'title' => [
'type' => 'text',
'analyzer' => 'my_analyzer'
],
'content' => [
'type' => 'text',
'analyzer' => 'my_analyzer'
]
]
]
]
]
];
I did find out that removing the analyzer from my field mappings allowed results to reappear, but I get no results the second I add the analyzer.
Here's an example query that I make.
{
"body": {
"query": {
"bool": {
"must": {
"multi_match": {
"query": "apples",
"fields": [
"title",
"content"
]
}
},
"filter": {
"terms": {
"site_id": [
"1351",
"1349"
]
}
},
"must_not": [
{
"match": {
"visible": "false"
}
},
{
"match": {
"locked": "true"
}
}
]
}
}
}
}

Answer 1

Probably, what you really want, is to use the english analyzer that is provided. The standard analyzer which is the default will tokenize on whitespace and some punctuation, but will leave apostrophes alone. The english analyzer can stem and remove stop words since the language is known.
Here is the standard analyzer's output, where you can see "john's":
POST _analyze
{
"analyzer": "standard",
"text": "John's potato"
}
{
"tokens": [
{
"token": "john's",
"start_offset": 0,
"end_offset": 6,
"type": "<ALPHANUM>",
"position": 0
},
{
"token": "potato",
"start_offset": 7,
"end_offset": 13,
"type": "<ALPHANUM>",
"position": 1
}
]
}
And here is the english analyzer where you can see the 's is removed. The stemming will allow "John's", "Johns", and "John" to all match the document.
POST _analyze
{
"analyzer": "english",
"text": "John's potato"
}
{
"tokens": [
{
"token": "john",
"start_offset": 0,
"end_offset": 6,
"type": "<ALPHANUM>",
"position": 0
},
{
"token": "potato",
"start_offset": 7,
"end_offset": 13,
"type": "<ALPHANUM>",
"position": 1
}
]
}