summaryrefslogtreecommitdiff
path: root/extractor/data_array.h
diff options
context:
space:
mode:
authorPaul Baltescu <pauldb89@gmail.com>2013-02-22 16:52:25 +0000
committerPaul Baltescu <pauldb89@gmail.com>2013-02-22 16:52:25 +0000
commita660f88aed61a7447ee72b9334a84686121086fe (patch)
tree447eb3d0f1ba167e3cdaeacd025650b75545cdb9 /extractor/data_array.h
parentcfef250df7ef378d6678bb6fc26402407f496184 (diff)
parent120b78445c563c0751997a6e2a5354b86a7cd865 (diff)
Merge branch 'master' into experiment
Diffstat (limited to 'extractor/data_array.h')
-rw-r--r--extractor/data_array.h6
1 files changed, 3 insertions, 3 deletions
diff --git a/extractor/data_array.h b/extractor/data_array.h
index 7c120b3c..96950789 100644
--- a/extractor/data_array.h
+++ b/extractor/data_array.h
@@ -15,6 +15,9 @@ enum Side {
TARGET
};
+// TODO: This class has features for both the source and target data arrays.
+// Maybe we can save some memory by having more specific implementations (e.g.
+// sentence_id is only needed for the source data array).
class DataArray {
public:
static int NULL_WORD;
@@ -48,7 +51,6 @@ class DataArray {
virtual int GetSentenceStart(int position) const;
- //TODO(pauldb): Add unit tests.
virtual int GetSentenceLength(int sentence_id) const;
virtual int GetSentenceId(int position) const;
@@ -67,8 +69,6 @@ class DataArray {
unordered_map<string, int> word2id;
vector<string> id2word;
vector<int> data;
- // TODO(pauldb): We only need sentence_id for the source language. Maybe we
- // can save some memory here.
vector<int> sentence_id;
vector<int> sentence_start;
};