aboutsummaryrefslogtreecommitdiff
path: root/icing/index/string-section-indexing-handler.cc
diff options
context:
space:
mode:
authorJiayu Hu <hujiayu@google.com>2023-11-30 14:10:57 -0800
committerJiayu Hu <hujiayu@google.com>2023-11-30 18:24:06 -0800
commitcb6ac3ede1d2ad050895b588417ea353c75953fe (patch)
tree431aae8a813d3b3e229077175c1363901b870d53 /icing/index/string-section-indexing-handler.cc
parentbe04186537a2e78ef1f27ba646676133d7e83c9a (diff)
downloadicing-cb6ac3ede1d2ad050895b588417ea353c75953fe.tar.gz
Update Icing from upstream.
Descriptions: ======================================================================== [Icing][version 3] Bump kVersion to 3 ======================================================================== Make lite index magic dependent on `IcingSearchEngineOptions::build_property_existence_metadata_hits` ======================================================================== Add a flag in IcingSearchEngineOptions to control whether to build property existence metadata hits ======================================================================== Support `hasProperty(property_path)` in the advanced query language ======================================================================== Add PropertyExistenceIndexingHandler to index property existence metadata hit ======================================================================== [JoinIndex Improvement][11/x] Add IcingSearchEngine initialization unit test for switching join index ======================================================================== [JoinIndex Improvement][10/x] Change/Add IcingSearchEngine unit tests ======================================================================== [JoinIndex Improvement][9/x] Integrate QualifiedIdJoinIndexImplV2 with IcingSearchEngine ======================================================================== [JoinIndex Improvement][8/x] Integrate QualifiedIdJoinIndexImplV2 with JoinProcessor ======================================================================== [JoinIndex Improvement][8/x] Integrate QualifiedIdJoinIndexImplV2 with QualifiedIdJoinIndexingHandler ======================================================================== [JoinIndex Improvement][7/x] Create QualifiedIdJoinIndex interface ======================================================================== [JoinIndex Improvement][6.1/x] Unit test (Optimize) ======================================================================== [JoinIndex Improvement][6.0/x] Unit test (General, Put, GetIterator) ======================================================================== [JoinIndex Improvement][5.3/x] Implement Optimize ======================================================================== Remove accents from Greek letters in normalizer ======================================================================== Make arm emulator tests build-only. ======================================================================== [JoinIndex Improvement][5.2/x] Implement GetIterator ======================================================================== [JoinIndex Improvement][5.1/x] Implement Put ======================================================================== [JoinIndex Improvement][5.0/x] Branch QualifiedIdJoinIndex to QualifiedIdJoinIndexImplV2 ======================================================================== [JoinIndex Improvement][4/x] Implement PostingListJoinDataAccessor ======================================================================== [JoinIndex Improvement][3/x] Implement PostingListJoinDataSerializer and DocumentIdToJoinInfo data type ======================================================================== [JoinIndex Improvement][2/x] Create NamespaceFingerprintIdentifier ======================================================================== [JoinIndex Improvement][1/x] Implement namespace_id_old_to_new in Compaction ======================================================================== Update test to also handle ICU 74 segmentation rules. ======================================================================== [Icing][Expand QueryStats][3/x] Add new fields into QueryStats (1) ======================================================================== [Icing][Expand QueryStats][2/x] Refactor QueryStatsProto ======================================================================== [Icing][Expand QueryStats][1/x] Publish DocHitInfoIterator CallStats ======================================================================== Add additional property filter tests ======================================================================== Deprecate hit_intersect_section_ids_mask in DocHitInfoIterator ======================================================================== Change default requires_full_emulation to False for portable_cc_test (third_party/icing/testing) ======================================================================== Cleanup Set requires_full_emulation to True for selective tests ======================================================================== Fix monkey test failures ======================================================================== Complete monkey test logic to change schema during monkey test runtime ======================================================================== Refactor monkey test to prepare for schema update ======================================================================== Fix the schema bug found by monkey test with seed 2551429844 ======================================================================== Move set query stats to the very top of InternalSearch() ======================================================================== Apply section restriction only on leaf nodes ======================================================================== [6/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (Advanced query parser) ======================================================================== [5/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (PersistentHashMap) ======================================================================== [4/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (PostingListIntegerIndexSerializer) ======================================================================== [3/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (PostingListHitSerializer) ======================================================================== [2/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (Posting list storage) ======================================================================== [1/n] Fix callsites in Icing that forgot to check libtextclassifier3::Status (Non-functional changes) ======================================================================== Decouple section restriction data from iterators ======================================================================== Fix the crash when a schema type gets more indexable properties than allowed ======================================================================== Add a checker to verify the property data type matches the schema. ======================================================================== Change global std::string in i18n-utils to constexpr std::string_view. ======================================================================== Adjust LiteIndex sort at indexing check conditions. ======================================================================== Bug: 305098009 Bug: 307508735 Bug: 291130542 Bug: 275121148 Bug: 303239901 Bug: 301116242 Bug: 299321977 Bug: 300135897 Bug: 297549761 Bug: 309826655 Bug: 296349369 Bug: 302192690 Bug: 302609704 Bug: 301566713 NO_IFTTT="False Alarm: The path is only valid in G3. kVersion is changed to 3, and schema is compatible with version 1." Change-Id: I8c4c3cd9b93e5240bd774f0a3d6d812f7a9ec198
Diffstat (limited to 'icing/index/string-section-indexing-handler.cc')
-rw-r--r--icing/index/string-section-indexing-handler.cc68
1 files changed, 3 insertions, 65 deletions
diff --git a/icing/index/string-section-indexing-handler.cc b/icing/index/string-section-indexing-handler.cc
index f5e06ad..8b20d04 100644
--- a/icing/index/string-section-indexing-handler.cc
+++ b/icing/index/string-section-indexing-handler.cc
@@ -21,15 +21,12 @@
#include "icing/text_classifier/lib3/utils/base/status.h"
#include "icing/text_classifier/lib3/utils/base/statusor.h"
-#include "icing/absl_ports/canonical_errors.h"
#include "icing/index/index.h"
-#include "icing/legacy/core/icing-string-util.h"
#include "icing/proto/logging.pb.h"
#include "icing/proto/schema.pb.h"
#include "icing/schema/section.h"
#include "icing/store/document-id.h"
#include "icing/transform/normalizer.h"
-#include "icing/util/clock.h"
#include "icing/util/logging.h"
#include "icing/util/status-macros.h"
#include "icing/util/tokenized-document.h"
@@ -39,35 +36,18 @@ namespace lib {
/* static */ libtextclassifier3::StatusOr<
std::unique_ptr<StringSectionIndexingHandler>>
-StringSectionIndexingHandler::Create(const Clock* clock,
- const Normalizer* normalizer,
+StringSectionIndexingHandler::Create(const Normalizer* normalizer,
Index* index) {
- ICING_RETURN_ERROR_IF_NULL(clock);
ICING_RETURN_ERROR_IF_NULL(normalizer);
ICING_RETURN_ERROR_IF_NULL(index);
return std::unique_ptr<StringSectionIndexingHandler>(
- new StringSectionIndexingHandler(clock, normalizer, index));
+ new StringSectionIndexingHandler(normalizer, index));
}
libtextclassifier3::Status StringSectionIndexingHandler::Handle(
const TokenizedDocument& tokenized_document, DocumentId document_id,
- bool recovery_mode, PutDocumentStatsProto* put_document_stats) {
- std::unique_ptr<Timer> index_timer = clock_.GetNewTimer();
-
- if (index_.last_added_document_id() != kInvalidDocumentId &&
- document_id <= index_.last_added_document_id()) {
- if (recovery_mode) {
- // Skip the document if document_id <= last_added_document_id in recovery
- // mode without returning an error.
- return libtextclassifier3::Status::OK;
- }
- return absl_ports::InvalidArgumentError(IcingStringUtil::StringPrintf(
- "DocumentId %d must be greater than last added document_id %d",
- document_id, index_.last_added_document_id()));
- }
- index_.set_last_added_document_id(document_id);
-
+ PutDocumentStatsProto* put_document_stats) {
uint32_t num_tokens = 0;
libtextclassifier3::Status status;
for (const TokenizedSection& section :
@@ -122,53 +102,11 @@ libtextclassifier3::Status StringSectionIndexingHandler::Handle(
}
}
- // Check and sort the LiteIndex HitBuffer if we're successful.
- if (status.ok() && index_.LiteIndexNeedSort()) {
- std::unique_ptr<Timer> sort_timer = clock_.GetNewTimer();
- index_.SortLiteIndex();
-
- if (put_document_stats != nullptr) {
- put_document_stats->set_lite_index_sort_latency_ms(
- sort_timer->GetElapsedMilliseconds());
- }
- }
-
if (put_document_stats != nullptr) {
- put_document_stats->set_term_index_latency_ms(
- index_timer->GetElapsedMilliseconds());
put_document_stats->mutable_tokenization_stats()->set_num_tokens_indexed(
num_tokens);
}
- // If we're either successful or we've hit resource exhausted, then attempt a
- // merge.
- if ((status.ok() || absl_ports::IsResourceExhausted(status)) &&
- index_.WantsMerge()) {
- ICING_LOG(INFO) << "Merging the index at docid " << document_id << ".";
-
- std::unique_ptr<Timer> merge_timer = clock_.GetNewTimer();
- libtextclassifier3::Status merge_status = index_.Merge();
-
- if (!merge_status.ok()) {
- ICING_LOG(ERROR) << "Index merging failed. Clearing index.";
- if (!index_.Reset().ok()) {
- return absl_ports::InternalError(IcingStringUtil::StringPrintf(
- "Unable to reset to clear index after merge failure. Merge "
- "failure=%d:%s",
- merge_status.error_code(), merge_status.error_message().c_str()));
- } else {
- return absl_ports::DataLossError(IcingStringUtil::StringPrintf(
- "Forced to reset index after merge failure. Merge failure=%d:%s",
- merge_status.error_code(), merge_status.error_message().c_str()));
- }
- }
-
- if (put_document_stats != nullptr) {
- put_document_stats->set_index_merge_latency_ms(
- merge_timer->GetElapsedMilliseconds());
- }
- }
-
return status;
}