searchengines-web/MeiliSearch
searchengines-web/MeiliSearch
1
0
Fork 0
mirror of https://github.com/meilisearch/MeiliSearch synced 2025-07-04 12:27:13 +02:00
Code Issues Releases Wiki Activity

Add indices field to _matchesPosition to specify where in an array a match comes from (#5005)

Browse source 
* Remove unreachable code

* Add `indices` field to `MatchBounds`

For matches inside arrays, this field holds the indices of the array
elements that matched. For example, searching for `cat` inside
`{ "a": ["dog", "cat", "fox"] }` would return `indices: [1]`. For nested
arrays, this contains multiple indices, starting with the one for the
top-most array. For matches in fields without arrays, `indices` is not
serialized (does not exist) to save space.
This commit is contained in:
Lukas Kalbertodt 2024-11-20 01:00:43 +01:00 • committed by GitHub
parent c1d8ee2a8d
commit 057fcb3993
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 139 additions and 97 deletions
Download patch file Download diff file Expand all files Collapse all files

124
crates/meilisearch/src/search/mod.rs
View file

@ -1733,46 +1733,51 @@ fn format_fields(
// select the attributes to retrieve
let displayable_names =
displayable_ids.iter().map(|&fid| field_ids_map.name(fid).expect("Missing field name"));
permissive_json_pointer::map_leaf_values(&mut document, displayable_names, |key, value| {
// To get the formatting option of each key we need to see all the rules that applies
// to the value and merge them together. eg. If a user said he wanted to highlight `doggo`
// and crop `doggo.name`. `doggo.name` needs to be highlighted + cropped while `doggo.age` is only
// highlighted.
// Warn: The time to compute the format list scales with the number of fields to format;
// cumulated with map_leaf_values that iterates over all the nested fields, it gives a quadratic complexity:
// d*f where d is the total number of fields to display and f is the total number of fields to format.
let format = formatting_fields_options
.iter()
.filter(|(name, _option)| {
milli::is_faceted_by(name, key) || milli::is_faceted_by(key, name)
})
.map(|(_, option)| **option)
.reduce(|acc, option| acc.merge(option));
let mut infos = Vec::new();
// if no locales has been provided, we try to find the locales in the localized_attributes.
let locales = locales.or_else(|| {
localized_attributes
permissive_json_pointer::map_leaf_values(
&mut document,
displayable_names,
|key, array_indices, value| {
// To get the formatting option of each key we need to see all the rules that applies
// to the value and merge them together. eg. If a user said he wanted to highlight `doggo`
// and crop `doggo.name`. `doggo.name` needs to be highlighted + cropped while `doggo.age` is only
// highlighted.
// Warn: The time to compute the format list scales with the number of fields to format;
// cumulated with map_leaf_values that iterates over all the nested fields, it gives a quadratic complexity:
// d*f where d is the total number of fields to display and f is the total number of fields to format.
let format = formatting_fields_options
.iter()
.find(|rule| rule.match_str(key))
.map(LocalizedAttributesRule::locales)
});
.filter(|(name, _option)| {
milli::is_faceted_by(name, key) || milli::is_faceted_by(key, name)
})
.map(|(_, option)| **option)
.reduce(|acc, option| acc.merge(option));
let mut infos = Vec::new();
*value = format_value(
std::mem::take(value),
builder,
format,
&mut infos,
compute_matches,
locales,
);
// if no locales has been provided, we try to find the locales in the localized_attributes.
let locales = locales.or_else(|| {
localized_attributes
.iter()
.find(|rule| rule.match_str(key))
.map(LocalizedAttributesRule::locales)
});
if let Some(matches) = matches_position.as_mut() {
if !infos.is_empty() {
matches.insert(key.to_owned(), infos);
*value = format_value(
std::mem::take(value),
builder,
format,
&mut infos,
compute_matches,
array_indices,
locales,
);
if let Some(matches) = matches_position.as_mut() {
if !infos.is_empty() {
matches.insert(key.to_owned(), infos);
}
}
}
});
},
);
let selectors = formatted_options
.keys()
@ -1790,13 +1795,14 @@ fn format_value(
format_options: Option<FormatOptions>,
infos: &mut Vec<MatchBounds>,
compute_matches: bool,
array_indices: &[usize],
locales: Option<&[Language]>,
) -> Value {
match value {
Value::String(old_string) => {
let mut matcher = builder.build(&old_string, locales);
if compute_matches {
let matches = matcher.matches();
let matches = matcher.matches(array_indices);
infos.extend_from_slice(&matches[..]);
}
@ -1808,51 +1814,15 @@ fn format_value(
None => Value::String(old_string),
}
}
Value::Array(values) => Value::Array(
values
.into_iter()
.map(|v| {
format_value(
v,
builder,
format_options.map(|format_options| FormatOptions {
highlight: format_options.highlight,
crop: None,
}),
infos,
compute_matches,
locales,
)
})
.collect(),
),
Value::Object(object) => Value::Object(
object
.into_iter()
.map(|(k, v)| {
(
k,
format_value(
v,
builder,
format_options.map(|format_options| FormatOptions {
highlight: format_options.highlight,
crop: None,
}),
infos,
compute_matches,
locales,
),
)
})
.collect(),
),
// `map_leaf_values` makes sure this is only called for leaf fields
Value::Array(_) => unreachable!(),
Value::Object(_) => unreachable!(),
Value::Number(number) => {
let s = number.to_string();
let mut matcher = builder.build(&s, locales);
if compute_matches {
let matches = matcher.matches();
let matches = matcher.matches(array_indices);
infos.extend_from_slice(&matches[..]);
}