diff options
| author | Patrick Simianer <p@simianer.de> | 2013-12-04 20:13:07 +0100 |
|---|---|---|
| committer | Patrick Simianer <p@simianer.de> | 2013-12-04 20:13:07 +0100 |
| commit | 02647059daa297e7b2b3ca3a2c03d848ae3ad9f2 (patch) | |
| tree | a5d2b5d66a8cff38a9422378c66861a3fb493e80 /extractor/rule_factory.cc | |
| parent | 7b2cd4e93114baaa329c483a98d6f7999aad1ba0 (diff) | |
| parent | 7d391d8eb88d27c5637042fefe2d27e7b12f5587 (diff) | |
fix merge conflict
Diffstat (limited to 'extractor/rule_factory.cc')
| -rw-r--r-- | extractor/rule_factory.cc | 11 |
1 files changed, 8 insertions, 3 deletions
diff --git a/extractor/rule_factory.cc b/extractor/rule_factory.cc index 6ae2d792..18a60695 100644 --- a/extractor/rule_factory.cc +++ b/extractor/rule_factory.cc @@ -12,6 +12,7 @@ #include "phrase_builder.h" #include "rule.h" #include "rule_extractor.h" +#include "phrase_location_sampler.h" #include "sampler.h" #include "scorer.h" #include "suffix_array.h" @@ -68,7 +69,8 @@ HieroCachingRuleFactory::HieroCachingRuleFactory( target_data_array, alignment, phrase_builder, scorer, vocabulary, max_rule_span, min_gap_size, max_nonterminals, max_rule_symbols, true, false, require_tight_phrases); - sampler = make_shared<Sampler>(source_suffix_array, max_samples); + sampler = make_shared<PhraseLocationSampler>( + source_suffix_array, max_samples); } HieroCachingRuleFactory::HieroCachingRuleFactory( @@ -101,7 +103,9 @@ HieroCachingRuleFactory::HieroCachingRuleFactory() {} HieroCachingRuleFactory::~HieroCachingRuleFactory() {} -Grammar HieroCachingRuleFactory::GetGrammar(const vector<int>& word_ids, const unordered_set<int>& blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array) { +Grammar HieroCachingRuleFactory::GetGrammar( + const vector<int>& word_ids, + const unordered_set<int>& blacklisted_sentence_ids) { Clock::time_point start_time = Clock::now(); double total_extract_time = 0; double total_intersect_time = 0; @@ -193,7 +197,8 @@ Grammar HieroCachingRuleFactory::GetGrammar(const vector<int>& word_ids, const u Clock::time_point extract_start = Clock::now(); if (!state.starts_with_x) { // Extract rules for the sampled set of occurrences. - PhraseLocation sample = sampler->Sample(next_node->matchings, blacklisted_sentence_ids, source_data_array); + PhraseLocation sample = sampler->Sample( + next_node->matchings, blacklisted_sentence_ids); vector<Rule> new_rules = rule_extractor->ExtractRules(next_phrase, sample); rules.insert(rules.end(), new_rules.begin(), new_rules.end()); |