1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366
/*! # Collectors Collectors define the information you want to extract from the documents matching the queries. In tantivy jargon, we call this information your search "fruit". Your fruit could for instance be : - [the count of matching documents](./struct.Count.html) - [the top 10 documents, by relevancy or by a fast field](./struct.TopDocs.html) - [facet counts](./struct.FacetCollector.html) At one point in your code, you will trigger the actual search operation by calling [the `search(...)` method of your `Searcher` object](../struct.Searcher.html#method.search). This call will look like this. ```verbatim let fruit = searcher.search(&query, &collector)?; ``` Here the type of fruit is actually determined as an associated type of the collector (`Collector::Fruit`). # Combining several collectors A rich search experience often requires to run several collectors on your search query. For instance, - selecting the top-K products matching your query - counting the matching documents - computing several facets - computing statistics about the matching product prices A simple and efficient way to do that is to pass your collectors as one tuple. The resulting `Fruit` will then be a typed tuple with each collector's original fruits in their respective position. ```rust # use tantivy::schema::*; # use tantivy::*; # use tantivy::query::*; use tantivy::collector::{Count, TopDocs}; # # fn main() -> tantivy::Result<()> { # let mut schema_builder = Schema::builder(); # let title = schema_builder.add_text_field("title", TEXT); # let schema = schema_builder.build(); # let index = Index::create_in_ram(schema); # let mut index_writer = index.writer(3_000_000)?; # index_writer.add_document(doc!( # title => "The Name of the Wind", # )); # index_writer.add_document(doc!( # title => "The Diary of Muadib", # )); # index_writer.commit()?; # let reader = index.reader()?; # let searcher = reader.searcher(); # let query_parser = QueryParser::for_index(&index, vec![title]); # let query = query_parser.parse_query("diary")?; let (doc_count, top_docs): (usize, Vec<(Score, DocAddress)>) = searcher.search(&query, &(Count, TopDocs::with_limit(2)))?; # Ok(()) # } ``` The `Collector` trait is implemented for up to 4 collectors. If you have more than 4 collectors, you can either group them into tuples of tuples `(a,(b,(c,d)))`, or rely on [`MultiCollector`](./struct.MultiCollector.html). # Combining several collectors dynamically Combining collectors into a tuple is a zero-cost abstraction: everything happens as if you had manually implemented a single collector combining all of our features. Unfortunately it requires you to know at compile time your collector types. If on the other hand, the collectors depend on some query parameter, you can rely on `MultiCollector`'s. # Implementing your own collectors. See the `custom_collector` example. */ use crate::DocId; use crate::Result; use crate::Score; use crate::SegmentLocalId; use crate::SegmentReader; use downcast_rs::impl_downcast; mod count_collector; pub use self::count_collector::Count; mod multi_collector; pub use self::multi_collector::MultiCollector; mod top_collector; mod top_score_collector; pub use self::top_score_collector::TopDocs; mod custom_score_top_collector; pub use self::custom_score_top_collector::{CustomScorer, CustomSegmentScorer}; mod tweak_score_top_collector; pub use self::tweak_score_top_collector::{ScoreSegmentTweaker, ScoreTweaker}; mod facet_collector; pub use self::facet_collector::FacetCollector; /// `Fruit` is the type for the result of our collection. /// e.g. `usize` for the `Count` collector. pub trait Fruit: Send + downcast_rs::Downcast {} impl<T> Fruit for T where T: Send + downcast_rs::Downcast {} /// Collectors are in charge of collecting and retaining relevant /// information from the document found and scored by the query. /// /// For instance, /// /// - keeping track of the top 10 best documents /// - computing a breakdown over a fast field /// - computing the number of documents matching the query /// /// Our search index is in fact a collection of segments, so /// a `Collector` trait is actually more of a factory to instance /// `SegmentCollector`s for each segments. /// /// The collection logic itself is in the `SegmentCollector`. /// /// Segments are not guaranteed to be visited in any specific order. pub trait Collector: Sync { /// `Fruit` is the type for the result of our collection. /// e.g. `usize` for the `Count` collector. type Fruit: Fruit; /// Type of the `SegmentCollector` associated to this collector. type Child: SegmentCollector<Fruit = Self::Fruit>; /// `set_segment` is called before beginning to enumerate /// on this segment. fn for_segment( &self, segment_local_id: SegmentLocalId, segment: &SegmentReader, ) -> Result<Self::Child>; /// Returns true iff the collector requires to compute scores for documents. fn requires_scoring(&self) -> bool; /// Combines the fruit associated to the collection of each segments /// into one fruit. fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit>; } /// The `SegmentCollector` is the trait in charge of defining the /// collect operation at the scale of the segment. /// /// `.collect(doc, score)` will be called for every documents /// matching the query. pub trait SegmentCollector: 'static { /// `Fruit` is the type for the result of our collection. /// e.g. `usize` for the `Count` collector. type Fruit: Fruit; /// The query pushes the scored document to the collector via this method. fn collect(&mut self, doc: DocId, score: Score); /// Extract the fruit of the collection from the `SegmentCollector`. fn harvest(self) -> Self::Fruit; } // ----------------------------------------------- // Tuple implementations. impl<Left, Right> Collector for (Left, Right) where Left: Collector, Right: Collector, { type Fruit = (Left::Fruit, Right::Fruit); type Child = (Left::Child, Right::Child); fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> { let left = self.0.for_segment(segment_local_id, segment)?; let right = self.1.for_segment(segment_local_id, segment)?; Ok((left, right)) } fn requires_scoring(&self) -> bool { self.0.requires_scoring() || self.1.requires_scoring() } fn merge_fruits( &self, children: Vec<(Left::Fruit, Right::Fruit)>, ) -> Result<(Left::Fruit, Right::Fruit)> { let mut left_fruits = vec![]; let mut right_fruits = vec![]; for (left_fruit, right_fruit) in children { left_fruits.push(left_fruit); right_fruits.push(right_fruit); } Ok(( self.0.merge_fruits(left_fruits)?, self.1.merge_fruits(right_fruits)?, )) } } impl<Left, Right> SegmentCollector for (Left, Right) where Left: SegmentCollector, Right: SegmentCollector, { type Fruit = (Left::Fruit, Right::Fruit); fn collect(&mut self, doc: DocId, score: Score) { self.0.collect(doc, score); self.1.collect(doc, score); } fn harvest(self) -> <Self as SegmentCollector>::Fruit { (self.0.harvest(), self.1.harvest()) } } // 3-Tuple impl<One, Two, Three> Collector for (One, Two, Three) where One: Collector, Two: Collector, Three: Collector, { type Fruit = (One::Fruit, Two::Fruit, Three::Fruit); type Child = (One::Child, Two::Child, Three::Child); fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> { let one = self.0.for_segment(segment_local_id, segment)?; let two = self.1.for_segment(segment_local_id, segment)?; let three = self.2.for_segment(segment_local_id, segment)?; Ok((one, two, three)) } fn requires_scoring(&self) -> bool { self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring() } fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> { let mut one_fruits = vec![]; let mut two_fruits = vec![]; let mut three_fruits = vec![]; for (one_fruit, two_fruit, three_fruit) in children { one_fruits.push(one_fruit); two_fruits.push(two_fruit); three_fruits.push(three_fruit); } Ok(( self.0.merge_fruits(one_fruits)?, self.1.merge_fruits(two_fruits)?, self.2.merge_fruits(three_fruits)?, )) } } impl<One, Two, Three> SegmentCollector for (One, Two, Three) where One: SegmentCollector, Two: SegmentCollector, Three: SegmentCollector, { type Fruit = (One::Fruit, Two::Fruit, Three::Fruit); fn collect(&mut self, doc: DocId, score: Score) { self.0.collect(doc, score); self.1.collect(doc, score); self.2.collect(doc, score); } fn harvest(self) -> <Self as SegmentCollector>::Fruit { (self.0.harvest(), self.1.harvest(), self.2.harvest()) } } // 4-Tuple impl<One, Two, Three, Four> Collector for (One, Two, Three, Four) where One: Collector, Two: Collector, Three: Collector, Four: Collector, { type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit); type Child = (One::Child, Two::Child, Three::Child, Four::Child); fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> { let one = self.0.for_segment(segment_local_id, segment)?; let two = self.1.for_segment(segment_local_id, segment)?; let three = self.2.for_segment(segment_local_id, segment)?; let four = self.3.for_segment(segment_local_id, segment)?; Ok((one, two, three, four)) } fn requires_scoring(&self) -> bool { self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring() || self.3.requires_scoring() } fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> { let mut one_fruits = vec![]; let mut two_fruits = vec![]; let mut three_fruits = vec![]; let mut four_fruits = vec![]; for (one_fruit, two_fruit, three_fruit, four_fruit) in children { one_fruits.push(one_fruit); two_fruits.push(two_fruit); three_fruits.push(three_fruit); four_fruits.push(four_fruit); } Ok(( self.0.merge_fruits(one_fruits)?, self.1.merge_fruits(two_fruits)?, self.2.merge_fruits(three_fruits)?, self.3.merge_fruits(four_fruits)?, )) } } impl<One, Two, Three, Four> SegmentCollector for (One, Two, Three, Four) where One: SegmentCollector, Two: SegmentCollector, Three: SegmentCollector, Four: SegmentCollector, { type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit); fn collect(&mut self, doc: DocId, score: Score) { self.0.collect(doc, score); self.1.collect(doc, score); self.2.collect(doc, score); self.3.collect(doc, score); } fn harvest(self) -> <Self as SegmentCollector>::Fruit { ( self.0.harvest(), self.1.harvest(), self.2.harvest(), self.3.harvest(), ) } } impl_downcast!(Fruit); #[cfg(test)] pub mod tests;