From ad9740b4627d1025d2d12a5363b42bd8243be756 Mon Sep 17 00:00:00 2001
From: ilor <kailoran@gmail.com>
Date: Fri, 25 Feb 2011 10:22:05 +0100
Subject: [PATCH] loose tag parsing option in xmlreader
---
libcorpus2/io/xmlreader.cpp | 4 +++-
libcorpus2/io/xmlreader.h | 7 +++++++
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/libcorpus2/io/xmlreader.cpp b/libcorpus2/io/xmlreader.cpp
index 7a51708..a9b8e0c 100644
--- a/libcorpus2/io/xmlreader.cpp
+++ b/libcorpus2/io/xmlreader.cpp
@@ -33,6 +33,7 @@ XmlReader::XmlReader(const Tagset& tagset,
, sbuf_(), tok_(NULL), sent_(), chunk_(), obuf_(obuf)
, disamb_only_(false), disamb_sh_(false)
, warn_on_inconsistent_(true), warn_on_unexpected_(true)
+ , loose_tag_parsing_(false)
{
}
@@ -221,7 +222,8 @@ void XmlReader::on_end_element(const Glib::ustring &name)
grab_characters_ = false;
state_ = STATE_LEX;
} else if (state_ == STATE_TAG && name == "ctag") {
- Tag tag = tagset_.parse_simple_tag(get_buf(), true);
+ Tag tag = tagset_.parse_simple_tag(get_buf(),
+ loose_tag_parsing_ ? Tagset::ParseLoose : Tagset::ParseDefault);
tok_->lexemes().back().set_tag(tag);
grab_characters_ = false;
state_ = STATE_LEX;
diff --git a/libcorpus2/io/xmlreader.h b/libcorpus2/io/xmlreader.h
index fe88533..f2b9b14 100644
--- a/libcorpus2/io/xmlreader.h
+++ b/libcorpus2/io/xmlreader.h
@@ -51,6 +51,10 @@ public:
warn_on_unexpected_ = v;
}
+ void set_loose_tag_parsing(bool v) {
+ loose_tag_parsing_ = v;
+ }
+
protected:
std::string get_type_from_attributes(const AttributeList& attributes) const;
@@ -130,6 +134,9 @@ protected:
/// Tag name for sentence objects, customized in child class ctors
std::string sentence_tag_name_;
+
+ /// Flag to disable strict tag correctness checking
+ bool loose_tag_parsing_;
};
} /* end ns Corpus2 */
--
GitLab