From ef381e17bbd4647ae9894f6d56f5db87fce2f861 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Cl=C3=A9ment=20Renault?= <clement@meilisearch.com>
Date: Fri, 19 Feb 2021 11:20:42 +0100
Subject: [PATCH] Compute the candidates for each sub query tree

---
 milli/src/search/criteria/words.rs | 77 +++++++++++++++++++++++++++++-
 1 file changed, 75 insertions(+), 2 deletions(-)

diff --git a/milli/src/search/criteria/words.rs b/milli/src/search/criteria/words.rs
index bf3aa8b12..93298b64e 100644
--- a/milli/src/search/criteria/words.rs
+++ b/milli/src/search/criteria/words.rs
@@ -1,10 +1,11 @@
 use std::collections::HashMap;
 use std::mem::take;
 
+use anyhow::bail;
 use roaring::RoaringBitmap;
 
 use crate::search::query_tree::Operation;
-use super::{Candidates, Criterion, CriterionResult, Context};
+use super::{Candidates, Criterion, CriterionResult, Context, query_docids, query_pair_proximity_docids};
 
 pub struct Words<'t> {
     ctx: &'t dyn Context,
@@ -64,9 +65,13 @@ impl<'t> Criterion for Words<'t> {
                         None => candidates.clone(),
                     };
 
+                    let mut found_candidates = resolve_candidates(self.ctx, &qt, &mut self.candidates_cache)?;
+                    found_candidates.intersect_with(&candidates);
+                    candidates.difference_with(&found_candidates);
+
                     return Ok(Some(CriterionResult {
                         query_tree: Some(qt),
-                        candidates: candidates.clone(),
+                        candidates: found_candidates,
                         bucket_candidates,
                     }));
                 },
@@ -107,3 +112,71 @@ fn explode_query_tree(query_tree: Operation) -> Vec<Operation> {
         otherwise => vec![otherwise],
     }
 }
+
+fn resolve_candidates<'t>(
+    ctx: &'t dyn Context,
+    query_tree: &Operation,
+    cache: &mut HashMap<(Operation, u8), RoaringBitmap>,
+) -> anyhow::Result<RoaringBitmap>
+{
+    fn resolve_operation<'t>(
+        ctx: &'t dyn Context,
+        query_tree: &Operation,
+        cache: &mut HashMap<(Operation, u8), RoaringBitmap>,
+    ) -> anyhow::Result<RoaringBitmap>
+    {
+        use Operation::{And, Consecutive, Or, Query};
+
+        match query_tree {
+            And(ops) => {
+                let mut candidates = RoaringBitmap::new();
+                let mut first_loop = true;
+                for op in ops {
+                    let docids = resolve_operation(ctx, op, cache)?;
+                    if first_loop {
+                        candidates = docids;
+                        first_loop = false;
+                    } else {
+                        candidates.intersect_with(&docids);
+                    }
+                }
+                Ok(candidates)
+            },
+            Consecutive(ops) => {
+                let mut candidates = RoaringBitmap::new();
+                let mut first_loop = true;
+                for slice in ops.windows(2) {
+                    match (&slice[0], &slice[1]) {
+                        (Operation::Query(left), Operation::Query(right)) => {
+                            match query_pair_proximity_docids(ctx, left, right, 1)? {
+                                pair_docids if pair_docids.is_empty() => {
+                                    return Ok(RoaringBitmap::new())
+                                },
+                                pair_docids if first_loop => {
+                                    candidates = pair_docids;
+                                    first_loop = false;
+                                },
+                                pair_docids => {
+                                    candidates.intersect_with(&pair_docids);
+                                },
+                            }
+                        },
+                        _ => bail!("invalid consecutive query type"),
+                    }
+                }
+                Ok(candidates)
+            },
+            Or(_, ops) => {
+                let mut candidates = RoaringBitmap::new();
+                for op in ops {
+                    let docids = resolve_operation(ctx, op, cache)?;
+                    candidates.union_with(&docids);
+                }
+                Ok(candidates)
+            },
+            Query(q) => Ok(query_docids(ctx, q)?),
+        }
+    }
+
+    resolve_operation(ctx, query_tree, cache)
+}