From fcea786ea5ee59203eb09155359cfd8a6d62ecfa Mon Sep 17 00:00:00 2001
From: Bartosz Broda <bartosz.broda@gmail.com>
Date: Fri, 10 Jun 2011 10:42:18 +0200
Subject: [PATCH] add more mwe for ad hoc testing, printing of partial parses
---
libmwereader/mweparser.cpp | 25 +++++++++++++++++--------
libmwereader/mweparser.h | 4 +++-
libmwereader/mwereader.cpp | 2 +-
libmwereader/test_mwe.xml | 23 +++++++++++++++++++++++
4 files changed, 44 insertions(+), 10 deletions(-)
diff --git a/libmwereader/mweparser.cpp b/libmwereader/mweparser.cpp
index ab85753..60ccdb0 100644
--- a/libmwereader/mweparser.cpp
+++ b/libmwereader/mweparser.cpp
@@ -12,7 +12,6 @@ namespace Corpus2 {
MWEParser::MWEParser()
: BasicSaxParser(), state_(NONE)
{
-
}
std::string MWEParser::get_attribute(const AttributeList& attributes,
@@ -82,7 +81,6 @@ namespace Corpus2 {
state_ = NONE;
} else if(state_ == CONDITION && name == "condition"){
wccl_operator_ = finish_get_text();
- std::cout << wccl_operator_ << std::endl;
state_ = MWEGROUP;
} else if(state_ == MWEGROUP && name == "mwegroup"){
state_ = UNITSDESC;
@@ -91,12 +89,8 @@ namespace Corpus2 {
} else if(state_ == MWE && name == "MWE"){
state_ = INSTANCES;
// TODO: tworzenie jednostki
- std::cout << "Tworzenie jednostki: " << mwe_base_ << " dla ";
- foreach(str_map::value_type &i, variables_)
- std::cout << i.first << ": " << i.second << ", ";
- std::cout << "\nhead: " << head_cond_ << "\nop: "
- << wccl_operator_ << std::endl;
- std::cout << "MWE Group name: " << group_name_ << std::endl;
+ print_current_mwe(true);
+
} else if(state_ == VAR && name == "var"){
state_ = MWE;
variables_[var_name_] = finish_get_text();
@@ -109,6 +103,21 @@ namespace Corpus2 {
}
}
+ void MWEParser::print_current_mwe(bool with_condition)
+ {
+ std::cout << "Forma podstawowa: " << mwe_base_ << "\nZmienne: ";
+
+ foreach(str_map::value_type &i, variables_)
+ std::cout << i.first << ": " << i.second << ", ";
+ std::cout << "\nWarunek głowy: " << head_cond_ << "\n";
+ if(with_condition){
+ std::cout << "Grupa jednostek: " << group_name_ << std::endl;
+ std::cout << "Operator: " << wccl_operator_ << std::endl;
+
+ }
+
+ }
+
std::string MWEParser::finish_get_text()
{
std::string str = get_buf();
diff --git a/libmwereader/mweparser.h b/libmwereader/mweparser.h
index 442e1a8..4ccdaf3 100644
--- a/libmwereader/mweparser.h
+++ b/libmwereader/mweparser.h
@@ -6,7 +6,6 @@
namespace Corpus2 {
-typedef std::map<std::string, std::string> str_map;
class MWEParser : public BasicSaxParser
{
@@ -14,12 +13,15 @@ public:
MWEParser();
protected:
+ typedef std::map<std::string, std::string> str_map;
void on_start_element(const Glib::ustring &name,
const AttributeList& attributes);
void on_end_element(const Glib::ustring &name);
std::string finish_get_text();
+ void print_current_mwe(bool with_condition = false);
+
/// retrives tagset= attribute
std::string get_attribute(const AttributeList& attributes,
diff --git a/libmwereader/mwereader.cpp b/libmwereader/mwereader.cpp
index f6bb63b..6fe0d1e 100644
--- a/libmwereader/mwereader.cpp
+++ b/libmwereader/mwereader.cpp
@@ -9,7 +9,7 @@ bool MWEReader::registered = TokenReader::register_path_reader<MWEReader>(
MWEReader::MWEReader(const Tagset &tagset, const std::string &filename)
: TokenReader(tagset), inner_filename_(filename)
{
- // TODO implementataion
+ // TODO implementataion?
}
MWEReader::~MWEReader()
diff --git a/libmwereader/test_mwe.xml b/libmwereader/test_mwe.xml
index 78c3d1a..07bb808 100644
--- a/libmwereader/test_mwe.xml
+++ b/libmwereader/test_mwe.xml
@@ -16,6 +16,29 @@
<var name="Subst2">paweł</var>
<head>inter(cas[0], {nom})</head>
</MWE>
+ <MWE base="waga netto">
+ <var name="Subst1">waga</var>
+ <var name="Subst2">netto</var>
+ <head>inter(base[0], "waga")</head>
+ </MWE>
+ </instances>
+ </mwegroup>
+ <mwegroup name="AdjSubstFix" type="fix" class="subst">
+ <condition>
+ and(
+ inter(base[0],{$Adj}),
+ inter(class[0],{adj}),
+ inter(base[1],$Subst),
+ inter(class[1],{subst,ger,depr}),
+ agrpp(0,1,{nmb,gnd,cas})
+ )
+ </condition>
+ <instances>
+ <MWE name="dobre imiÄ™">
+ <var name="Adj">dobry</var>
+ <var name="Subst">imiÄ™</var>
+ <head>inter(class[0],{subst,ger,depr})</head>
+ </MWE>
</instances>
</mwegroup>
--
GitLab