rust/tantivy
1
0
Fork 0
mirror of https://github.com/quickwit-oss/tantivy.git synced 2026-05-23 19:50:42 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #32 from currymj/pedantry

Browse Source 
Pedantry
This commit is contained in:
Paul Masurel
2016-10-01 00:33:04 +09:00
committed by GitHub
parent 15567e0aa3 ac9aa5cb5e
commit 21a442fc13
17 changed files with 106 additions and 107 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
README.md
View File

@@ -13,11 +13,11 @@ It is strongly inspired by Lucene's design.
# Features
- configurable indexing (optional term frequency and position indexing)
- Tf-Idf scoring
- tf-idf scoring
- Basic query language
- Incremental indexing
- Multithreaded indexing (indexing en wikipedia takes 4mn on my desktop)
- Mmap based
- Multithreaded indexing (indexing English Wikipedia takes 4 minutes on my desktop)
- mmap based
- SIMD integer compression
- u32 fast fields (equivalent of doc values in Lucene)
- LZ4 compressed document store
@@ -35,7 +35,7 @@ It will walk you through getting a wikipedia search engine up and running in a f
Tantivy has a git submodule called `simdcomp`.
After cloning the repository, you will need to initialize and update
the submodules. The project can then be build using `cargo`.
the submodules. The project can then be built using `cargo`.
git clone git@github.com:fulmicoton/tantivy.git
git submodule init

46
examples/simple_search.rs
View File

@@ -25,7 +25,7 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// # Defining the schema
//
// Tantivy index require to have a very strict schema.
// The Tantivy index requires a very strict schema.
// The schema declares which fields are in the index,
// and for each field, its type and "the way it should
// be indexed".
@@ -47,7 +47,7 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// `STORED` means that the field will also be saved
// in a compressed, row-oriented key-value store.
// This store is useful to reconstruct the
// document that were selected during the search phase.
// documents that were selected during the search phase.
schema_builder.add_text_field("title", TEXT | STORED);
// Our first field is body.
@@ -64,29 +64,29 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// Let's create a brand new index.
//
// This will actually just save a meta.json
// with our schema the directory.
// with our schema in the directory.
let index = try!(Index::create(index_path, schema.clone()));
// To insert document we need an index writer.
// There shall be only one writer at a time.
// Besides, this single `IndexWriter` is already
// There must be only one writer at a time.
// This single `IndexWriter` is already
// multithreaded.
//
// Here we used a buffer of 1 GB. Using a bigger
// Here we use a buffer of 1 GB. Using a bigger
// heap for the indexer can increase its throughput.
// This buffer will be split between the indexing
// threads.
let mut index_writer = try!(index.writer(1_000_000_000));
// Let's now index our documents!
// Let's index our documents!
// We first need a handle on the title and the body field.
// ### Create a document "manually".
//
// We can create a document manually, by setting adding the fields
// We can create a document manually, by setting the fields
// one by one in a Document object.
let title = schema.get_field("title").unwrap();
let body = schema.get_field("body").unwrap();
@@ -122,7 +122,7 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// This is an example, so we will only index 3 documents
// here. You can check out tantivy's tutorial to index
// the English wikipedia. Tantivy's indexing is rather fast.
// Indexing 5 millions articles of the English wikipedia takes
// Indexing 5 million articles of the English wikipedia takes
// around 4 minutes on my computer!
@@ -131,56 +131,56 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// At this point our documents are not searchable.
//
//
// We need to call .commit() explicitely to force the
// We need to call .commit() explicitly to force the
// index_writer to finish processing the documents in the queue,
// flush the current index on the disk, and advertise
// flush the current index to the disk, and advertise
// the existence of new documents.
//
// This call is blocking.
try!(index_writer.commit());
// If `.commit()` returns correctly, then all of the
// documents have been added before are guaranteed to be
// documents that have been added are guaranteed to be
// persistently indexed.
//
// In the scenario of a crash or a power failure,
// tantivy behaves as if it rollbacked to its last
// tantivy behaves as if has rolled back to its last
// commit.
// # Searching
//
// Let's search our index. This starts
// Let's search our index. We start
// by creating a searcher. There can be more
// than one searcher at a time.
//
// You are supposed to acquire a search
// You should create a searcher
// every time you start a "search query".
let searcher = index.searcher();
// The query parser can interpret human queries.
// Here, if the user does not specify which
// field he wants to search, tantivy will search
// field they want to search, tantivy will search
// in both title and body.
let query_parser = QueryParser::new(index.schema(), vec!(title, body));
// QueryParser may fail if the query is not in the right
// format. For user facing applications, this can be a problem.
// A ticket has been filled regarding this problem.
// A ticket has been opened regarding this problem.
let query = try!(query_parser.parse_query("sea whale"));
// A query defines a set of documents, as
// well as the way they should be scored.
//
// Query created by the query parser are scoring according
// A query created by the query parser is scored according
// to a metric called Tf-Idf, and will consider
// any document matching at least one of our terms.
// ### Collectors
//
// We are not interested in all of the document but
// only in the top 10. Keep track of our top 10 best documents
// We are not interested in all of the documents but
// only in the top 10. Keeping track of our top 10 best documents
// is the role of the TopCollector.
let mut top_collector = TopCollector::with_limit(10);
@@ -188,14 +188,14 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
// We can now perform our query.
try!(query.search(&searcher, &mut top_collector));
// Our top collector now contains are 10
// Our top collector now contains the 10
// most relevant doc ids...
let doc_addresses = top_collector.docs();
// The actual documents still need to be
// retrieved from Tantivy's store.
//
// Since body was not configured as stored,
// Since the body field was not configured as stored,
// the document returned will only contain
// a title.
@@ -205,4 +205,4 @@ fn run_example(index_path: &Path) -> tantivy::Result<()> {
}
Ok(())
}
}

7
src/collector/count_collector.rs
View File

@@ -5,13 +5,13 @@ use SegmentReader;
use SegmentLocalId;
/// `CountCollector` collector only counts how many
/// document are matching the query.
/// documents match the query.
pub struct CountCollector {
count: usize,
}
impl CountCollector {
/// Returns the count of document that where
/// Returns the count of documents that were
/// collected.
pub fn count(&self,) -> usize {
self.count
@@ -20,8 +20,7 @@ impl CountCollector {
impl Default for CountCollector {
fn default() -> CountCollector {
CountCollector {
count: 0,
CountCollector {count: 0,
}
}
}

24
src/collector/mod.rs
View File

@@ -20,16 +20,16 @@ pub use self::chained_collector::chain;
///
///
/// For instance,
/// - keeping track of the top 10 best documents
/// - computing a break down over a fast field
/// - computing the number of documents matching the query
///
/// - keeping track of the top 10 best documents
/// - computing a breakdown over a fast field
/// - computing the number of documents matching the query
///
/// Queries are in charge of pushing the `DocSet` to the collector.
///
/// As they work on multiple segment, they first inform
/// the collector of a change in segment and then
/// call the collect method to push document to the collector.
/// As they work on multiple segments, they first inform
/// the collector of a change in a segment and then
/// call the `collect` method to push the document to the collector.
///
/// Temporally, our collector will receive calls
/// - `.set_segment(0, segment_reader_0)`
@@ -45,10 +45,10 @@ pub use self::chained_collector::chain;
///
/// Segments are not guaranteed to be visited in any specific order.
pub trait Collector {
/// `set_segment` is called before starting enumerating
/// `set_segment` is called before beginning to enumerate
/// on this segment.
fn set_segment(&mut self, segment_local_id: SegmentLocalId, segment: &SegmentReader) -> io::Result<()>;
/// The query pushes scored document to the collector via this method.
/// The query pushes the scored document to the collector via this method.
fn collect(&mut self, scored_doc: ScoredDoc);
}
@@ -57,7 +57,7 @@ impl<'a, C: Collector> Collector for &'a mut C {
fn set_segment(&mut self, segment_local_id: SegmentLocalId, segment: &SegmentReader) -> io::Result<()> {
(*self).set_segment(segment_local_id, segment)
}
/// The query pushes scored document to the collector via this method.
/// The query pushes the scored document to the collector via this method.
fn collect(&mut self, scored_doc: ScoredDoc) {
(*self).collect(scored_doc);
}
@@ -120,10 +120,10 @@ pub mod tests {
/// Collects in order all of the fast field for all of the
/// doc of the `DocSet`
/// Collects in order all of the fast fields for all of the
/// doc in the `DocSet`
///
/// This collector is essentially useful for tests.
/// This collector is mainly useful for tests.
pub struct FastFieldTestCollector {
vals: Vec<u32>,
field: Field,

4
src/collector/multi_collector.rs
View File

@@ -5,7 +5,7 @@ use SegmentReader;
use SegmentLocalId;
/// Multicollector makes it possible to collect on more than one collector
/// Multicollector makes it possible to collect on more than one collector.
/// It should only be used for use cases where the Collector types is unknown
/// at compile time.
/// If the type of the collectors is known, you should prefer to use `ChainedCollector`.
@@ -60,4 +60,4 @@ mod tests {
assert_eq!(count_collector.count(), 3);
assert!(top_collector.at_capacity());
}
}
}

22
src/collector/top_collector.rs
View File

@@ -50,7 +50,7 @@ pub struct TopCollector {
impl TopCollector {
/// Creates a top collector, with a number of document of "limit"
/// Creates a top collector, with a number of documents equal to "limit".
///
/// # Panics
/// The method panics if limit is 0
@@ -65,9 +65,9 @@ impl TopCollector {
}
}
/// Returns the decreasingly sorted K-best documents.
/// Returns K best documents sorted in decreasing order.
///
/// Calling this method will triggers the sort.
/// Calling this method triggers the sort.
/// The result of the sort is not cached.
pub fn docs(&self) -> Vec<DocAddress> {
self.score_docs()
@@ -76,9 +76,9 @@ impl TopCollector {
.collect()
}
/// Returns the decreasingly sorted K-best ScoredDocument.
/// Returns K best ScoredDocument sorted in decreasing order.
///
/// Calling this method will triggers the sort.
/// Calling this method triggers the sort.
/// The result of the sort is not cached.
pub fn score_docs(&self) -> Vec<(Score, DocAddress)> {
let mut scored_docs: Vec<GlobalScoredDoc> = self.heap
@@ -90,9 +90,9 @@ impl TopCollector {
.map(|GlobalScoredDoc(score, doc_address)| (score, doc_address))
.collect()
}
/// Return true iff at least K document have gone through
/// the collector.
/// Return true iff at least K documents have gone through
/// the collector.
#[inline]
pub fn at_capacity(&self, ) -> bool {
self.heap.len() >= self.limit
@@ -176,8 +176,8 @@ mod tests {
.collect();
assert_eq!(docs, vec!(7, 1, 5, 3));
}
}
#[test]
@@ -185,4 +185,4 @@ mod tests {
fn test_top_0() {
TopCollector::with_limit(0);
}
}
}

16
src/core/index.rs
View File

@@ -90,7 +90,7 @@ impl Index {
/// Creates a new index in a temp directory.
///
/// The index will use the `MMapDirectory` in a newly created directory.
/// The temp directory will be destroyed automatically when the Index object
/// The temp directory will be destroyed automatically when the `Index` object
/// is destroyed.
///
/// The temp directory is only used for testing the `MmapDirectory`.
@@ -100,7 +100,7 @@ impl Index {
Index::from_directory(directory, schema)
}
/// Creates a new index given a directory and an IndexMeta.
/// Creates a new index given a directory and an `IndexMeta`.
fn create_from_metas(directory: Box<Directory>, metas: IndexMeta) -> Result<Index> {
let schema = metas.schema.clone();
let index = Index {
@@ -160,7 +160,7 @@ impl Index {
/// Marks the segment as published.
// TODO find a rusty way to hide that, while keeping
// it visible for IndexWriters.
// it visible for `IndexWriter`s.
pub fn publish_segments(&mut self,
segment_ids: &[SegmentId],
docstamp: u64) -> Result<()> {
@@ -204,7 +204,7 @@ impl Index {
}
/// Return a segment object given a segment_id
/// Return a segment object given a `segment_id`
///
/// The segment may or may not exist.
fn segment(&self, segment_id: SegmentId) -> Segment {
@@ -246,7 +246,7 @@ impl Index {
/// Either
// - it fails, in which case an error is returned,
/// and the `meta.json` remains untouched,
/// - it success, and `meta.json` is written
/// - it succeeds, and `meta.json` is written
/// and flushed.
pub fn save_metas(&mut self,) -> Result<()> {
let mut w = Vec::new();
@@ -286,9 +286,9 @@ impl Index {
///
/// This method should be called every single time a search
/// query is performed.
/// The searcher are taken from a pool of `NUM_SEARCHERS` searchers.
/// The searchers are taken from a pool of `NUM_SEARCHERS` searchers.
/// If no searcher is available
/// it may block.
/// this may block.
///
/// The same searcher must be used for a given query, as it ensures
/// the use of a consistent segment set.
@@ -313,4 +313,4 @@ impl Clone for Index {
searcher_pool: self.searcher_pool.clone(),
}
}
}
}

10
src/core/segment_reader.rs
View File

@@ -25,7 +25,7 @@ use schema::TextIndexingOptions;
use error::Error;
/// Entrypoint to access all of the datastructures of the `Segment`
/// Entry point to access all of the datastructures of the `Segment`
///
/// - term dictionary
/// - postings
@@ -34,8 +34,8 @@ use error::Error;
/// - field norm reader
///
/// The segment reader has a very low memory footprint,
/// as close to all of the memory data is in Mmapped.
///
/// as close to all of the memory data is mmapped.
///
pub struct SegmentReader {
segment_info: SegmentInfo,
segment_id: SegmentId,
@@ -51,7 +51,7 @@ pub struct SegmentReader {
impl SegmentReader {
/// Returns the highest document id ever attributed in
/// this segment + 1.
/// Today, `tantivy` does not handle deletes so, it happens
/// Today, `tantivy` does not handle deletes, so it happens
/// to also be the number of documents in the index.
pub fn max_doc(&self) -> DocId {
self.segment_info.max_doc
@@ -233,7 +233,7 @@ impl SegmentReader {
self.read_postings(term, segment_posting_option)
}
/// Returns the term info of associated with the term.
/// Returns the term info associated with the term.
pub fn get_term_info(&self, term: &Term) -> Option<TermInfo> {
self.term_infos.get(term.as_slice())
}

26
src/directory/directory.rs
View File

@@ -9,7 +9,7 @@ use std::marker::Sync;
/// Write-once read many (WORM) abstraction for where tantivy's index should be stored.
///
/// There is currently two implementations of `Directory`
/// There are currently two implementations of `Directory`
///
/// - The [`MMapDirectory`](struct.MmapDirectory.html), this
/// should be your default choice.
@@ -20,19 +20,19 @@ pub trait Directory: fmt::Debug + Send + Sync + 'static {
/// Opens a virtual file for read.
///
/// Once a virtualfile is open, its data may not
/// Once a virtual file is open, its data may not
/// change.
///
/// Specifically, subsequent write or flush should
/// have no effect the returned `ReadOnlySource` object.
/// Specifically, subsequent writes or flushes should
/// have no effect on the returned `ReadOnlySource` object.
fn open_read(&self, path: &Path) -> result::Result<ReadOnlySource, FileError>;
/// Removes a file
///
/// Removing a file will not affect eventual
/// Removing a file will not affect an eventual
/// existing ReadOnlySource pointing to it.
///
/// Removing a non existing files, yields a
/// Removing a nonexistent file, yields a
/// `FileError::DoesNotExist`.
fn delete(&self, path: &Path) -> result::Result<(), FileError>;
@@ -44,28 +44,28 @@ pub trait Directory: fmt::Debug + Send + Sync + 'static {
/// same path should return a `ReadOnlySource`.
///
/// Write operations may be aggressively buffered.
/// The client of this trait is in charge to call flush
/// The client of this trait is responsible for calling flush
/// to ensure that subsequent `read` operations
/// will take in account preceding `write` operations.
/// will take into account preceding `write` operations.
///
/// Flush operation should also be persistent.
///
/// User shall not rely on `Drop` triggering `flush`.
/// The user shall not rely on `Drop` triggering `flush`.
/// Note that `RAMDirectory` will panic! if `flush`
/// was not called.
///
/// The file may not previously exists.
/// The file may not previously exist.
fn open_write(&mut self, path: &Path) -> Result<WritePtr, OpenWriteError>;
/// Atomically replace the content of a file by data.
/// Atomically replace the content of a file with data.
///
/// This calls ensure that reads can never *observe*
/// a partially written file.
///
/// The file may or may not previously exists.
/// The file may or may not previously exist.
fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()>;
/// Clone the directory and boxes the clone
/// Clones the directory and boxes the clone
fn box_clone(&self) -> Box<Directory>;
}

4
src/directory/mmap_directory.rs
View File

@@ -47,7 +47,7 @@ impl MmapDirectory {
/// Creates a new MmapDirectory in a temporary directory.
///
/// This is mostly useful to test the MmapDirectory itself.
/// For your unit test, prefer the RAMDirectory.
/// For your unit tests, prefer the RAMDirectory.
pub fn create_from_tempdir() -> io::Result<MmapDirectory> {
let tempdir = try!(TempDir::new("index"));
let tempdir_path = PathBuf::from(tempdir.path());
@@ -81,7 +81,7 @@ impl MmapDirectory {
}
/// Joins a relative_path to the directory `root_path`
/// to create proper complete `filepath`.
/// to create a proper complete `filepath`.
fn resolve_path(&self, relative_path: &Path) -> PathBuf {
self.root_path.join(relative_path)
}

8
src/directory/ram_directory.rs
View File

@@ -11,7 +11,7 @@ use directory::error::{OpenWriteError, FileError};
use directory::WritePtr;
use super::shared_vec_slice::SharedVecSlice;
/// Writer associated to the `RAMDirectory`
/// Writer associated with the `RAMDirectory`
///
/// The Writer just writes a buffer.
///
@@ -133,9 +133,9 @@ impl fmt::Debug for RAMDirectory {
}
/// Directory storing everything in anonymous memory.
/// A Directory storing everything in anonymous memory.
///
/// It's main purpose is unit test.
/// It is mainly meant for unit testing.
/// Writes are only made visible upon flushing.
///
#[derive(Clone)]
@@ -161,7 +161,7 @@ impl Directory for RAMDirectory {
fn open_write(&mut self, path: &Path) -> Result<WritePtr, OpenWriteError> {
let path_buf = PathBuf::from(path);
let vec_writer = VecWriter::new(path_buf.clone(), self.fs.clone());
// force the creation of the file to mimick the MMap directory.
// force the creation of the file to mimic the MMap directory.
if try!(self.fs.write(path_buf.clone(), &Vec::new())) {
Err(OpenWriteError::FileAlreadyExists(path_buf))
}

2
src/lib.rs
View File

@@ -96,7 +96,7 @@ pub use postings::SegmentPostingsOption;
/// u32 identifying a document within a segment.
/// Document gets their doc id assigned incrementally,
/// Documents have their doc id assigned incrementally,
/// as they are added in the segment.
pub type DocId = u32;

8
src/postings/docset.rs
View File

@@ -4,7 +4,7 @@ use std::borrow::BorrowMut;
use std::cmp::Ordering;
/// Expressed the outcome of a call to `DocSet`'s `.skip_next(...)`.
/// Expresses the outcome of a call to `DocSet`'s `.skip_next(...)`.
#[derive(PartialEq, Eq, Debug)]
pub enum SkipResult {
/// target was in the docset
@@ -24,8 +24,8 @@ pub trait DocSet {
/// element.
fn advance(&mut self,) -> bool;
/// After skipping position, the iterator in such a way `.doc()`
/// will return a value greater or equal to target.
/// After skipping, position the iterator in such a way that `.doc()`
/// will return a value greater than or equal to target.
///
/// SkipResult expresses whether the `target value` was reached, overstepped,
/// or if the `DocSet` was entirely consumed without finding any value
@@ -97,4 +97,4 @@ impl<'a, TDocSet: DocSet> DocSet for &'a mut TDocSet {
}
}

4
src/postings/postings.rs
View File

@@ -12,8 +12,8 @@ use common::HasLen;
/// as well as the list of term positions.
///
/// Its main implementation is `SegmentPostings`,
/// but some other implementation mocking SegmentPostings exists,
/// in order to help merging segment or for testing.
/// but other implementations mocking SegmentPostings exist,
/// in order to help when merging segments or for testing.
pub trait Postings: DocSet {
/// Returns the term frequency
fn term_freq(&self,) -> u32;

8
src/query/query_parser.rs
View File

@@ -29,7 +29,7 @@ pub enum ParsingError {
/// Tantivy's Query parser
///
/// The language covered by the current is extremely simple.
/// The language covered by the current parser is extremely simple.
///
/// * simple terms: "e.g.: `Barack Obama` are simply analyzed using
/// tantivy's `StandardTokenizer`, hence becoming `["barack", "obama"]`.
@@ -44,7 +44,7 @@ pub enum ParsingError {
///
/// This behavior is slower, but is not a bad idea if the user is sorting
/// by relevance : The user typically just scans through the first few
/// documents in order of decreasing relevance and will stop when the document
/// documents in order of decreasing relevance and will stop when the documents
/// are not relevant anymore.
/// Making it possible to make this behavior customizable is tracked in
/// [issue #27](https://github.com/fulmicoton/tantivy/issues/27).
@@ -135,9 +135,9 @@ impl QueryParser {
/// Parse a query
///
/// Note that `parse_query` returns an error if the input
/// not a valid query.
/// is not a valid query.
///
/// There is currently no lenient mode for the query parse
/// There is currently no lenient mode for the query parser
/// which makes it a bad choice for a public/broad user search engine.
///
/// Implementing a lenient mode for this query parser is tracked

12
src/schema/mod.rs
View File

@@ -4,7 +4,7 @@
# Schema definition
Tantivy has a very strict schema.
The schema defines information about the fields your index contains, that is for each field :
The schema defines information about the fields your index contains, that is, for each field :
* the field name (may only contain letters `[a-zA-Z]`, number `[0-9]`, and `_`)
* the type of the field (currently only `text` and `u32` are supported)
@@ -37,20 +37,20 @@ let schema = schema_builder.build();
We can split the problem of generating a search result page into two phases :
* identifying the list of 10 or so document to be displayed (Conceptually `query -> doc_ids[]`)
* identifying the list of 10 or so documents to be displayed (Conceptually `query -> doc_ids[]`)
* for each of these documents, retrieving the information required to generate the serp page. (`doc_ids[] -> Document[]`)
In the first phase, the hability to search for documents by the given field, is determined by the [`TextIndexingOptions`](enum.TextIndexingOptions.html) of our
In the first phase, the ability to search for documents by the given field is determined by the [`TextIndexingOptions`](enum.TextIndexingOptions.html) of our
[`TextOptions`](struct.TextOptions.html).
The effect of each possible settings is described more in detail [`TextIndexingOptions`](enum.TextIndexingOptions.html).
The effect of each possible setting is described more in detail [`TextIndexingOptions`](enum.TextIndexingOptions.html).
On the other hand setting the field as stored or not determines whether the field should be returned when [`searcher.doc(doc_address)`](../struct.Searcher.html#method.doc)
is called.
### Shortcuts
For convenience, a few special value of `TextOptions` for your convenience.
For convenience, a few special values of `TextOptions`.
They can be composed using the `|` operator.
The example can be rewritten :
@@ -82,7 +82,7 @@ Just like for Text fields (see above),
setting the field as stored defines whether the field will be
returned when [`searcher.doc(doc_address)`](../struct.Searcher.html#method.doc) is called,
and setting the field as indexed means that we will be able perform queries such as `num_stars:10`.
Note that contrary to text fields, u32 can only be indexed in one way for the moment.
Note that unlike text fields, u32 can only be indexed in one way for the moment.
This may change when we will start supporting range queries.
The `fast` option on the other hand is specific to u32 fields, and is only relevant

4
src/schema/schema.rs
View File

@@ -15,7 +15,7 @@ use std::fmt;
/// Tantivy has a very strict schema.
/// You need to specify in advance, whether a field is indexed or not,
/// You need to specify in advance whether a field is indexed or not,
/// stored or not, and RAM-based or not.
///
/// This is done by creating a schema object, and
@@ -483,4 +483,4 @@ mod tests {
}
}
}
}
}