Removing the dependency to tantivy in quickwit-common.

quickwit/quickwit-common/Cargo.toml

@@ -13,7 +13,6 @@ documentation = "https://quickwit.io/docs/"
 anyhow = { workspace = true }
 colored = { workspace = true }
 env_logger = { workspace = true }
-fastfield_codecs = { workspace = true }
 home = { workspace = true }
 itertools = { workspace = true }
 num_cpus = { workspace = true }
@@ -23,7 +22,6 @@ prometheus = { workspace = true }
 rand = { workspace = true }
 regex = { workspace = true }
 serde = { workspace = true }
-tantivy = { workspace = true }
 thiserror = { workspace = true }
 tokio = { workspace = true }
 tracing = { workspace = true }

quickwit/quickwit-common/src/lib.rs

@@ -20,7 +20,6 @@
 mod checklist;
 mod coolid;
 
-pub mod fast_field_reader;
 pub mod fs;
 pub mod metrics;
 pub mod net;

quickwit/quickwit-common/src/fast_field_reader.rs → quickwit/quickwit-doc-mapper/src/fast_field_reader.rs

@@ -19,9 +19,9 @@
 
 use std::sync::Arc;
 
-use tantivy::fastfield::{Column, FastFieldReaders};
-use tantivy::schema::{Field, FieldEntry, Type};
-use tantivy::{DateTime, DocId, TantivyError};
+use tantivy::fastfield::Column;
+use tantivy::schema::{Field, Type};
+use tantivy::{DateTime, DocId, SegmentReader, TantivyError};
 
 #[derive(Clone)]
 pub enum GenericFastFieldReader {
@@ -60,9 +60,11 @@ impl GenericFastFieldReader {
 
 pub fn timestamp_field_reader(
     timestamp_field: Field,
-    timestamp_field_entry: &FieldEntry,
-    fast_field_readers: &FastFieldReaders,
+    segment_reader: &SegmentReader,
 ) -> tantivy::Result<GenericFastFieldReader> {
+    let schema = segment_reader.schema();
+    let timestamp_field_entry = schema.get_field_entry(timestamp_field);
+    let fast_field_readers = segment_reader.fast_fields();
     let field_schema_type = timestamp_field_entry.field_type().value_type();
     let timestamp_field_reader = match field_schema_type {
         Type::I64 => GenericFastFieldReader::I64(fast_field_readers.i64(timestamp_field)?),

quickwit/quickwit-doc-mapper/src/lib.rs

@@ -27,6 +27,9 @@
 mod default_doc_mapper;
 mod doc_mapper;
 mod error;
+
+#[allow(missing_docs)]
+pub mod fast_field_reader;
 mod query_builder;
 mod routing_expression;
 mod sort_by;

quickwit/quickwit-indexing/src/actors/merge_executor.rs

@@ -28,9 +28,9 @@ use async_trait::async_trait;
 use fail::fail_point;
 use itertools::Itertools;
 use quickwit_actors::{Actor, ActorContext, ActorExitStatus, Handler, Mailbox, QueueCapacity};
-use quickwit_common::fast_field_reader::timestamp_field_reader;
 use quickwit_common::runtimes::RuntimeType;
 use quickwit_directories::UnionDirectory;
+use quickwit_doc_mapper::fast_field_reader::timestamp_field_reader;
 use quickwit_doc_mapper::{DocMapper, QUICKWIT_TOKENIZER_MANAGER};
 use quickwit_metastore::{Metastore, SplitMetadata};
 use quickwit_proto::metastore_api::DeleteTask;
@@ -452,15 +452,8 @@ impl MergeExecutor {
                             timestamp_field_name
                         ))
                     })?;
-                let timestamp_field_entry = merged_segment_reader
-                    .schema()
-                    .get_field_entry(timestamp_field);
-                let reader = timestamp_field_reader(
-                    timestamp_field,
-                    timestamp_field_entry,
-                    merged_segment_reader.fast_fields(),
-                )?;
-                Some(RangeInclusive::new(reader.min_value(), reader.max_value()))
+                let reader = timestamp_field_reader(timestamp_field, &merged_segment_reader)?;
+                Some(reader.min_value()..=reader.max_value())
             } else {
                 None
             };

quickwit/quickwit-search/src/filters.rs

@@ -19,7 +19,7 @@
 
 use std::ops::{Bound, RangeBounds};
 
-use quickwit_common::fast_field_reader::{timestamp_field_reader, GenericFastFieldReader};
+use quickwit_doc_mapper::fast_field_reader::{timestamp_field_reader, GenericFastFieldReader};
 use tantivy::schema::Field;
 use tantivy::{DocId, SegmentReader};
 
@@ -71,14 +71,7 @@ impl TimestampFilterBuilder {
         &self,
         segment_reader: &SegmentReader,
     ) -> tantivy::Result<Option<TimestampFilter>> {
-        let timestamp_field_entry = segment_reader
-            .schema()
-            .get_field_entry(self.timestamp_field);
-        let timestamp_field_reader = timestamp_field_reader(
-            self.timestamp_field,
-            timestamp_field_entry,
-            segment_reader.fast_fields(),
-        )?;
+        let timestamp_field_reader = timestamp_field_reader(self.timestamp_field, segment_reader)?;
         let segment_range = (
             timestamp_field_reader.min_value(),
             timestamp_field_reader.max_value(),