From 6cb1d99332d61edf1377d98a1570bafa4c9a1779 Mon Sep 17 00:00:00 2001
From: Adam Radziszewski <adam.radziszewski@pwr.wroc.pl>
Date: Mon, 1 Oct 2012 13:55:41 +0200
Subject: [PATCH] forgotten anninter sources
---
libwccl/ops/functions/strset/anninter.cpp | 78 +++++++++++++++++++++
libwccl/ops/functions/strset/anninter.h | 83 +++++++++++++++++++++++
2 files changed, 161 insertions(+)
create mode 100644 libwccl/ops/functions/strset/anninter.cpp
create mode 100644 libwccl/ops/functions/strset/anninter.h
diff --git a/libwccl/ops/functions/strset/anninter.cpp b/libwccl/ops/functions/strset/anninter.cpp
new file mode 100644
index 0000000..adcd06f
--- /dev/null
+++ b/libwccl/ops/functions/strset/anninter.cpp
@@ -0,0 +1,78 @@
+/*
+ Copyright (C) 2011 Adam Wardyński, Tomasz Śniatowski, Paweł Kędzia,
+ Adam Radziszewski, Bartosz Broda
+ Part of the WCCL project
+
+ This program is free software; you can redistribute it and/or modify it
+under the terms of the GNU Lesser General Public License as published by the Free
+Software Foundation; either version 3 of the License, or (at your option)
+any later version.
+
+ This program is distributed in the hope that it will be useful, but
+WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+or FITNESS FOR A PARTICULAR PURPOSE.
+
+ See the LICENSE, COPYING.LESSER and COPYING files for more details.
+*/
+
+#include <libwccl/ops/functions/strset/anninter.h>
+#include <libcorpus2/ann/annotatedsentence.h>
+
+#include <libpwrutils/util.h>
+
+namespace Wccl {
+
+AnnInter::BaseRetValPtr AnnInter::apply_internal(const FunExecContext& context) const
+{
+ boost::shared_ptr<Corpus2::AnnotatedSentence> as
+ = boost::dynamic_pointer_cast<Corpus2::AnnotatedSentence>(
+ context.sentence_context().get_sentence_ptr());
+ if (!as) {
+ throw InvalidArgument("context", "Operator needs an annotated sentence.");
+ }
+
+ const boost::shared_ptr<const Position>& pos = pos_expr_->apply(context);
+ const SentenceContext& sc = context.sentence_context();
+ // prepare the output set
+ boost::shared_ptr<StrSet > out_set = boost::make_shared<StrSet>();
+ // return empty set if out of bounds
+ int abs_idx = sc.get_abs_position(*pos);
+
+ if (sc.is_outside(*pos)) {
+ return out_set;
+ }
+
+ // get the value of the set of channel names to consider
+ const boost::shared_ptr<const StrSet >& in_set = strset_expr_->apply(context);
+
+ BOOST_FOREACH (const UnicodeString& u_chan_name, in_set->contents()) {
+ // check if there is any annot at pos
+ const std::string chan_name(PwrNlp::to_utf8(u_chan_name));
+ if (as->has_channel(chan_name)) {
+ const Corpus2::AnnotationChannel& chan = as->get_channel(chan_name);
+ if (chan.get_segment_at(abs_idx) != 0) {
+ // there is an annotation, so we take this channel's name
+ out_set->insert(u_chan_name);
+ }
+ }
+ }
+
+ return out_set;
+}
+
+std::string AnnInter::to_string(const Corpus2::Tagset&) const
+{
+ std::ostringstream ostream;
+ ostream << raw_name() << "(" << *pos_expr_
+ << ", " << *strset_expr_ << ")";
+ return ostream.str();
+}
+
+std::ostream& AnnInter::write_to(std::ostream& ostream) const
+{
+ ostream << raw_name() << "(" << *pos_expr_ << ", "
+ << ", " << *strset_expr_ << ")";
+ return ostream;
+}
+
+} /* end ns Wccl */
diff --git a/libwccl/ops/functions/strset/anninter.h b/libwccl/ops/functions/strset/anninter.h
new file mode 100644
index 0000000..ecbbee4
--- /dev/null
+++ b/libwccl/ops/functions/strset/anninter.h
@@ -0,0 +1,83 @@
+/*
+ Copyright (C) 2011 Adam Wardyński, Tomasz Śniatowski, Paweł Kędzia,
+ Adam Radziszewski, Bartosz Broda
+ Part of the WCCL project
+
+ This program is free software; you can redistribute it and/or modify it
+under the terms of the GNU Lesser General Public License as published by the Free
+Software Foundation; either version 3 of the License, or (at your option)
+any later version.
+
+ This program is distributed in the hope that it will be useful, but
+WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+or FITNESS FOR A PARTICULAR PURPOSE.
+
+ See the LICENSE, COPYING.LESSER and COPYING files for more details.
+*/
+
+#ifndef LIBWCCL_OPS_FUNCTIONS_STRSET_ANNINTER_H
+#define LIBWCCL_OPS_FUNCTIONS_STRSET_ANNINTER_H
+
+#include <libwccl/values/position.h>
+#include <libwccl/values/strset.h>
+#include <libwccl/ops/function.h>
+
+namespace Wccl {
+
+/**
+ * Operator that takes a position and a set of strings designating
+ * channel names and returns a subset of the given string set that
+ * corresponds to those channels with any annotation crossing given
+ * position.
+ */
+class AnnInter : public Function<StrSet>
+{
+public:
+ typedef boost::shared_ptr<Function<Position> > PosFunctionPtr;
+ typedef boost::shared_ptr<Function<StrSet> > StrSetFunctionPtr;
+
+ AnnInter(const PosFunctionPtr& pos_expr,
+ const StrSetFunctionPtr& strset_expr)
+ : pos_expr_(pos_expr),
+ strset_expr_(strset_expr)
+ {
+ BOOST_ASSERT(pos_expr_);
+ BOOST_ASSERT(strset_expr_);
+ }
+
+ /**
+ * String representation of the operator in form of:
+ * "affix(strset_expr_string)"
+ */
+ std::string to_string(const Corpus2::Tagset& tagset) const;
+
+ /**
+ * @returns Name of the function: "affix"
+ */
+ std::string raw_name() const {
+ return "anninter";
+ }
+
+protected:
+ const PosFunctionPtr pos_expr_;
+ const StrSetFunctionPtr strset_expr_;
+
+ /**
+ * Check which the channel names given in the strset_expr. Return those
+ * that have any annotation crossing the given position.
+ */
+ BaseRetValPtr apply_internal(const FunExecContext& context) const;
+
+ /**
+ * Writes raw string representation of the operator in form of:
+ * "anninter(pos_expr_raw_string, strset_expr_raw_string)"
+ * @note This version doesn't require tagset, but may be incomplete and/or
+ * contain internal info.
+ * @returns Stream written to.
+ */
+ std::ostream& write_to(std::ostream& ostream) const;
+};
+
+} /* end ns Wccl */
+
+#endif // LIBWCCL_OPS_FUNCTIONS_STRSET_ANNINTER_H
--
GitLab