Compare commits

...

4 Commits

Author SHA1 Message Date
Paul Masurel
98d17c9c77 rem broken seg
brotli by default
2021-01-09 09:30:44 +09:00
Paul Masurel
5583e5d0ad test for seg3 2021-01-09 09:11:57 +09:00
Paul Masurel
1fcf656ed7 added test with other broken segment 2021-01-09 08:55:40 +09:00
Paul Masurel
68fa58d74b DO NOT MERGE 2021-01-09 08:55:40 +09:00
5 changed files with 69 additions and 2 deletions

View File

@@ -73,7 +73,7 @@ debug-assertions = true
overflow-checks = true
[features]
default = ["mmap"]
default = ["mmap", "brotli"]
mmap = ["fs2", "tempfile", "memmap"]
brotli-compression = ["brotli"]
lz4-compression = ["lz4"]

View File

@@ -36,6 +36,8 @@ and should rely on either
mod index;
mod reader;
mod writer;
mod tests_store;
pub use self::reader::StoreReader;
pub use self::writer::StoreWriter;

View File

@@ -50,6 +50,22 @@ impl StoreReader {
self.skip_index.checkpoints()
}
pub fn documents(&self) -> Vec<Document> {
let mut documents = Vec::new();
for checkpoint in self.skip_index.checkpoints() {
println!("{:?}", checkpoint);
let block = self.read_block(&checkpoint).unwrap();
let mut cursor = &block[..];
while cursor.len() > 0 {
let doc_length = VInt::deserialize(&mut cursor).unwrap().val() as usize;
let doc = Document::deserialize(&mut &cursor[..doc_length]).unwrap();
documents.push(doc);
cursor = &cursor[doc_length..];
}
}
documents
}
fn block_checkpoint(&self, doc_id: DocId) -> Option<Checkpoint> {
self.skip_index.seek(doc_id)
}
@@ -104,7 +120,6 @@ impl StoreReader {
let doc_length = VInt::deserialize(&mut cursor)?.val() as usize;
cursor = &cursor[doc_length..];
}
let doc_length = VInt::deserialize(&mut cursor)?.val() as usize;
cursor = &cursor[..doc_length];
Ok(Document::deserialize(&mut cursor)?)

50
src/store/tests_store.rs Normal file
View File

@@ -0,0 +1,50 @@
use std::path::Path;
use crate::HasLen;
use crate::directory::{Directory, ManagedDirectory, MmapDirectory, RAMDirectory};
use crate::fastfield::DeleteBitSet;
use super::{StoreReader, StoreWriter};
#[test]
fn test_toto2() -> crate::Result<()> {
let directory = ManagedDirectory::wrap(MmapDirectory::open("src/store/broken_seg")?)?;
let path = Path::new("b6029ade1b954ea1acad15b432eaacb9.store");
assert!(directory.validate_checksum(path)?);
let store_file = directory.open_read(path)?;
let store = StoreReader::open(store_file)?;
let documents = store.documents();
// for doc in documents {
// println!("{:?}", doc);
// }
let doc= store.get(15_086)?;
Ok(())
}
#[test]
fn test_toto() -> crate::Result<()> {
let directory = ManagedDirectory::wrap(MmapDirectory::open("src/store/broken_seg")?)?;
assert!(directory.validate_checksum(Path::new("e6ece22e5bca4e0dbe7ce3e4dcbd5bbf.store"))?);
let store_file = directory.open_read(Path::new("e6ece22e5bca4e0dbe7ce3e4dcbd5bbf.store.patched"))?;
let store = StoreReader::open(store_file)?;
let doc= store.get(53)?;
println!("{:?}", doc);
// let documents = store.documents();
// let ram_directory = RAMDirectory::create();
// let path = Path::new("store");
// let store_wrt = ram_directory.open_write(path)?;
// let mut store_writer = StoreWriter::new(store_wrt);
// for doc in &documents {
// store_writer.store(doc)?;
// }
// store_writer.close()?;
// let store_data = ram_directory.open_read(path)?;
// let new_store = StoreReader::open(store_data)?;
// for doc in 0..59 {
// println!("{}", doc);
// let doc = new_store.get(doc)?;
// println!("{:?}", doc);
// }
Ok(())
}