✨ Put <description> in fifth field of json output, <title> in seventh

We're now actually parsing <description> from XML :)

Also, allow <description> alongside <label> in pipespec
<option>'s (under prefs)

+ add a schemas.xml for editor

Author: unhammer

src/checkertypes.hpp

@@ -34,7 +34,7 @@ namespace divvun {
  */
 
 typedef std::string Lang;
-typedef std::u16string Msg;
+typedef std::pair<std::u16string, std::u16string> Msg;
 typedef std::u16string ErrId;
 typedef std::basic_regex<char> ErrRe;
 

src/errors.dtd

@@ -1,4 +1,4 @@
-<!-- Copyright (C) 2016-2017, Kevin Brubeck Unhammer <[email protected]> -->
+<!-- Copyright (C) 2016-2019, Kevin Brubeck Unhammer <[email protected]> -->
 
 <!-- This program is free software: you can redistribute it and/or modify -->
 <!-- it under the terms of the GNU General Public License as published by -->

src/errors.rnc

@@ -1,4 +1,5 @@
-# Copyright (C) 2016-2017, Kevin Brubeck Unhammer <[email protected]>
+# THIS IS A GENERATED FILE, DO NOT EDIT!
+# Copyright (C) 2016-2019, Kevin Brubeck Unhammer <[email protected]>
 
 # This program is free software: you can redistribute it and/or modify
 

src/main_checker.cpp

@@ -94,16 +94,16 @@ void printPrefs(const Pipeline& pipeline) {
 		const Prefs& prefs = lp.second;
 		std::cout << "==== Toggles: ====" << std::endl;
 		for(const auto& id : prefs.toggleIds) {
-			std::cout << "- [ ] " << toUtf8(id.first) << " \t" << toUtf8(id.second) << std::endl;
+			std::cout << "- [ ] " << toUtf8(id.first) << " \t" << toUtf8(id.second.first) << std::endl;
 		}
 		for(const auto& re : prefs.toggleRes) {
-			std::cout << "- [ ] [regex] \t" << toUtf8(re.second) << std::endl;
+			std::cout << "- [ ] [regex] \t" << toUtf8(re.second.first) << std::endl;
 		}
 		std::cout << "==== Options: ====" << std::endl;
 		for(const Option& o : prefs.options) {
 			std::cout << "- " << o.name << " (" << o.type << "):" << std::endl;
 			for(const auto& c : o.choices) {
-				std::cout << "- ( ) " << toUtf8(c.first) << " \t" << toUtf8(c.second) << std::endl;
+				std::cout << "- ( ) " << toUtf8(c.first) << " \t" << toUtf8(c.second.first) << std::endl;
 			}
 		}
 	}

src/pipeline.hpp

@@ -217,11 +217,23 @@ inline void parsePrefs(LocalisedPrefs& prefs, const pugi::xml_node& cmd) {
 		unordered_map<Lang, unordered_map<ErrId, Msg>> lems;
 		for (const pugi::xml_node& option: pref.children()) {
 			const auto errId = fromUtf8(option.attribute("err-id").value());
-			for (const pugi::xml_node& label: option.children()) {
+			for (const pugi::xml_node& label: option.children("label")) {
 				const auto lang = label.attribute("xml:lang").value();
 				const auto msg = fromUtf8(label.text().get()); // or xml_raw_cdata(label);
-				lems[lang][errId] = msg;
+				// Let <description> default to <label> first:
+				lems[lang][errId] = std::make_pair(msg, msg);
 			}
+			for (const pugi::xml_node& description: option.children("description")) {
+				const auto lang = description.attribute("xml:lang").value();
+				const auto msg = fromUtf8(description.text().get());
+				if(lems[lang].find(errId) != lems[lang].end()) {
+					lems[lang][errId].second = msg;
+                                }
+				else {
+					// No <label> for this language, fallback to <description>:
+					lems[lang][errId] = std::make_pair(msg, msg);
+				}
+                        }
 		}
 		for(const auto& lem : lems) {
 			const Lang& lang = lem.first;

src/pipespec.dtd

@@ -1,4 +1,4 @@
-<!-- Copyright (C) 2017, Kevin Brubeck Unhammer <[email protected]> -->
+<!-- Copyright (C) 2017–2019, Kevin Brubeck Unhammer <[email protected]> -->
 
 <!-- This program is free software: you can redistribute it and/or modify -->
 <!-- it under the terms of the GNU General Public License as published by -->
@@ -57,7 +57,7 @@
 <!-- type and name are *not* unique here; several pipelines may have a pref for e.g. "Oxford comma" -->
 
 
-<!ELEMENT option (label+)>
+<!ELEMENT option ((label|description)+)>
 <!ATTLIST option
           err-id CDATA #REQUIRED
 >
@@ -69,6 +69,11 @@
           xml:lang CDATA #REQUIRED
 >
 
+<!ELEMENT description %Text;>
+<!ATTLIST description
+          xml:lang CDATA #REQUIRED
+>
+
 <!-- General "system" command – pipelines with this can only be used
      in settings where we can open processes: -->
 <!ELEMENT sh (arg*)> <!-- NOT IMPLEMENTED YET -->

src/pipespec.rnc

@@ -1,5 +1,5 @@
 # THIS IS A GENERATED FILE, DO NOT EDIT!
-# Copyright (C) 2017, Kevin Brubeck Unhammer <[email protected]>
+# Copyright (C) 2017–2019, Kevin Brubeck Unhammer <[email protected]>
 
 # This program is free software: you can redistribute it and/or modify
 
@@ -72,13 +72,16 @@ attlist.pref &=
   attribute type { text },
   attribute name { text }
 # type and name are *not* unique here; several pipelines may have a pref for e.g. "Oxford comma"
-option = element option { attlist.option, label+ }
+option = element option { attlist.option, (label | description)+ }
 attlist.option &= attribute err-id { text }
 
 Text = (text | em)*
 label = element label { attlist.label, Text }
 attlist.label &= attribute xml:lang { text }
 
+description = element description { attlist.description, Text }
+attlist.description &= attribute xml:lang { text }
+
 # General "system" command – pipelines with this can only be used
 # in settings where we can open processes:
 sh = element sh { attlist.sh, arg* }

src/suggest.cpp

@@ -89,36 +89,66 @@ const MsgMap readMessagesXml(pugi::xml_document& doc, pugi::xml_parse_result& re
 	MsgMap msgs;
 
 	if (result) {
+		// <default>'s:
 		for (pugi::xml_node def: doc.child("errors").child("defaults").children("default")) {
-			// std::cerr << "defaults" << std::endl;
+			// For all <title>'s and <description>'s, add all their parent <id>/<re>'s:
 			for (pugi::xml_node child: def.child("header").children("title")) {
 				const auto& msg = fromUtf8(xml_raw_cdata(child));
 				const auto& lang = child.attribute("xml:lang").value();
 				for (pugi::xml_node e: def.child("ids").children("e")) {
 					// e_value assumes we only ever have one PCDATA element here:
 					const auto& errtype = fromUtf8(e.attribute("id").value());
-					// std::cerr << toUtf8(errtype) << std::endl;
 					if(msgs[lang].first.count(errtype) != 0) {
 						std::cerr << "divvun-suggest: WARNING: Duplicate titles for " << e.attribute("id").value() << std::endl;
 					}
-					msgs[lang].first[errtype] = msg;
+					// Default to <title> as <description>, may be overridden below:
+					msgs[lang].first[errtype] = make_pair(msg, msg);
 				}
 				for (pugi::xml_node re: def.child("ids").children("re")) {
 					std::basic_regex<char> r(re.attribute("v").value());
-					msgs[lang].second.push_back(std::make_pair(r, msg));
+					msgs[lang].second.push_back(std::make_pair(r, make_pair(msg, msg)));
 				}
 			}
+			for (pugi::xml_node child: def.child("body").children("description")) {
+				const auto& msg = fromUtf8(xml_raw_cdata(child));
+				const auto& lang = child.attribute("xml:lang").value();
+				for (pugi::xml_node e: def.child("ids").children("e")) {
+					const auto& errtype = fromUtf8(e.attribute("id").value());
+					auto &langmsgs = msgs[lang].first;
+					if (langmsgs.find(errtype) != langmsgs.end()) {
+						langmsgs[errtype].second = msg;
+					} else {
+						// No <title> for this language, fallback to <description>:
+						langmsgs[errtype] = std::make_pair(msg, msg);
+					}
+                                }
+			}
 		}
+		// <error>'s
 		for (pugi::xml_node error: doc.child("errors").children("error")) {
+			const auto& errtype = fromUtf8(error.attribute("id").value());
+			// For all <title>'s and <description>'s, add the <error id> attribute:
 			for (pugi::xml_node child: error.child("header").children("title")) {
 				// child_value assumes we only ever have one PCDATA element here:
-				const auto& errtype = fromUtf8(error.attribute("id").value());
 				const auto& msg = fromUtf8(xml_raw_cdata(child));
 				const auto& lang = child.attribute("xml:lang").value();
-				if(msgs[lang].first.count(errtype) != 0) {
-					std::cerr << "divvun-suggest: WARNING: Duplicate titles for " << error.attribute("id").value() << std::endl;
+				auto& langmsgs = msgs[lang].first;
+				if(langmsgs.count(errtype) != 0) {
+					std::cerr << "divvun-suggest: WARNING: Duplicate <title>'s for " << error.attribute("id").value() << std::endl;
+				}
+				langmsgs[errtype] = make_pair(msg, msg);
+			}
+			for (pugi::xml_node child: error.child("body").children("description")) {
+				const auto& msg = fromUtf8(xml_raw_cdata(child));
+				const auto& lang = child.attribute("xml:lang").value();
+				auto& langmsgs = msgs[lang].first;
+				if(langmsgs.find(errtype) != langmsgs.end()) {
+					langmsgs[errtype].second = msg;
+                                }
+				else {
+					// No <title> for this language, fallback to <description>:
+					langmsgs[errtype] = std::make_pair(msg, msg);
 				}
-				msgs[lang].first[errtype] = msg;
 			}
 		}
 	}
@@ -397,10 +427,10 @@ variant<Nothing, Err> Suggest::cohort_errs(const ErrId& err_id,
 	if(cohort_empty(c) || c.added || ignores.find(err_id) != ignores.end()) {
 		return Nothing();
 	}
-	u16string msg;
+	Msg msg;
 	for(const auto& mlang : sortedmsglangs) {
-		if(msg.empty() && mlang != locale) {
-			std::cerr << "divvun-suggest: WARNING: No message for " << json::str(err_id) << " in xml:lang '" << locale << "', trying '" << mlang << "'" << std::endl;
+		if(msg.second.empty() && mlang != locale) {
+			std::cerr << "divvun-suggest: WARNING: No <description> for " << json::str(err_id) << " in xml:lang '" << locale << "', trying '" << mlang << "'" << std::endl;
 		}
 		const auto& lmsgs = msgs.at(mlang);
 		if(lmsgs.first.count(err_id) != 0) {
@@ -421,23 +451,28 @@ variant<Nothing, Err> Suggest::cohort_errs(const ErrId& err_id,
 				}
 			}
 		}
-		if(!msg.empty()) {
+		if(!msg.second.empty()) {
 			break;
 		}
 	}
-	if(msg.empty()) {
-		std::cerr << "divvun-suggest: WARNING: No message for " << json::str(err_id) << " in any xml:lang" << std::endl;
-		msg = err_id;
+	if(msg.second.empty()) {
+		std::cerr << "divvun-suggest: WARNING: No <description> for " << json::str(err_id) << " in any xml:lang" << std::endl;
+		msg.second = err_id;
+	}
+	if(msg.first.empty()) {
+		msg.first = err_id;
 	}
 	// TODO: Make suitable structure on creating MsgMap instead?
-	replaceAll(msg, u"$1", c.form);
+	replaceAll(msg.first, u"$1", c.form);
+	replaceAll(msg.second, u"$1", c.form);
 	for(const auto& r: c.readings) {
 		if((!r.errtype.empty()) && err_id != r.errtype) {
 			continue;
 		}
 		rel_on_match(r.rels, MSG_TEMPLATE_REL, sentence,
 			     [&] (const string& relname, size_t i_t, const Cohort& trg) {
-				     replaceAll(msg, fromUtf8(relname.c_str()), trg.form);
+				     replaceAll(msg.first, fromUtf8(relname.c_str()), trg.form);
+				     replaceAll(msg.second, fromUtf8(relname.c_str()), trg.form);
 			     });
 	}
 	auto beg = c.pos;
@@ -770,8 +805,9 @@ RunState Suggest::run_json(std::istream& is, std::ostream& os)
 		   << "," << std::to_string(e.beg)
 		   << "," << std::to_string(e.end)
 		   << "," << json::str(e.err)
-		   << "," << json::str(e.msg)
+		   << "," << json::str(e.msg.second)
 		   << "," << json::str_arr(e.rep)
+		   << "," << json::str(e.msg.first)
 		   << "]";
 		wantsep = true;
 	}

test/checker/pipespec.xml

@@ -15,6 +15,7 @@
       <pref type="Punctuation" name="Tusenskilje">
         <option err-id="tusen-mellom">
           <label xml:lang="nn">Eg vil ha mellomrom mellom 000</label>
+          <description xml:lang="nn">Det er lov med anten mellomrom eller punktum som skiljeteikn for tal over tusen.</description>
         </option>
         <option err-id="tusen-punktum">
           <label xml:lang="nn">Eg vil ha punktum mellom 000</label>

test/checker/schemas.xml

@@ -0,0 +1,19 @@
+<?xml version="1.0"?>
+<locatingRules xmlns="http://thaiopensource.com/ns/locating-rules/1.0">
+
+  <!--
+       Having this file in the same dir as your XML's
+       lets Emacs show validation errors.
+       More information at:
+       https://www.gnu.org/software/emacs/manual/html_node/nxml-mode/Schema-locating-files.html#Schema-locating-files
+  -->
+
+  <typeId id="errors" uri="../../src/errors.rnc"/>
+  <documentElement localName="errors" typeId="errors"/>
+  <uri pattern="errors.xml" typeId="errors"/>
+
+  <typeId id="pipespec" uri="../../src/pipespec.rnc"/>
+  <documentElement localName="pipespec" typeId="pipespec"/>
+  <uri pattern="pipespec.xml" typeId="pipespec"/>
+
+</locatingRules>