sync: minja (google/minja@418a236) (ggml-org#11574)

Author: ochafik

common/minja.hpp

@@ -693,7 +693,7 @@ enum SpaceHandling { Keep, Strip, StripSpaces, StripNewline };
 
 class TemplateToken {
 public:
-    enum class Type { Text, Expression, If, Else, Elif, EndIf, For, EndFor, Generation, EndGeneration, Set, EndSet, Comment, Macro, EndMacro, Filter, EndFilter };
+    enum class Type { Text, Expression, If, Else, Elif, EndIf, For, EndFor, Generation, EndGeneration, Set, EndSet, Comment, Macro, EndMacro, Filter, EndFilter, Break, Continue };
 
     static std::string typeToString(Type t) {
         switch (t) {
@@ -714,6 +714,8 @@ class TemplateToken {
             case Type::EndFilter: return "endfilter";
             case Type::Generation: return "generation";
             case Type::EndGeneration: return "endgeneration";
+            case Type::Break: return "break";
+            case Type::Continue: return "continue";
         }
         return "Unknown";
     }
@@ -815,6 +817,22 @@ struct CommentTemplateToken : public TemplateToken {
     CommentTemplateToken(const Location & location, SpaceHandling pre, SpaceHandling post, const std::string& t) : TemplateToken(Type::Comment, location, pre, post), text(t) {}
 };
 
+enum class LoopControlType { Break, Continue };
+
+class LoopControlException : public std::runtime_error {
+public:
+    LoopControlType control_type;
+    LoopControlException(const std::string & message, LoopControlType control_type) : std::runtime_error(message), control_type(control_type) {}
+    LoopControlException(LoopControlType control_type)
+      : std::runtime_error((std::ostringstream() << (control_type == LoopControlType::Continue ? "continue" : "break") << " outside of a loop").str()),
+        control_type(control_type) {}
+};
+
+struct LoopControlTemplateToken : public TemplateToken {
+    LoopControlType control_type;
+    LoopControlTemplateToken(const Location & location, SpaceHandling pre, SpaceHandling post, LoopControlType control_type) : TemplateToken(Type::Break, location, pre, post), control_type(control_type) {}
+};
+
 class TemplateNode {
     Location location_;
 protected:
@@ -825,6 +843,12 @@ class TemplateNode {
     void render(std::ostringstream & out, const std::shared_ptr<Context> & context) const {
         try {
             do_render(out, context);
+        } catch (const LoopControlException & e) {
+            // TODO: make stack creation lazy. Only needed if it was thrown outside of a loop.
+            std::ostringstream err;
+            err << e.what();
+            if (location_.source) err << error_location_suffix(*location_.source, location_.pos);
+            throw LoopControlException(err.str(), e.control_type);
         } catch (const std::exception & e) {
             std::ostringstream err;
             err << e.what();
@@ -897,6 +921,15 @@ class IfNode : public TemplateNode {
     }
 };
 
+class LoopControlNode : public TemplateNode {
+    LoopControlType control_type_;
+  public:
+    LoopControlNode(const Location & location, LoopControlType control_type) : TemplateNode(location), control_type_(control_type) {}
+    void do_render(std::ostringstream &, const std::shared_ptr<Context> &) const override {
+      throw LoopControlException(control_type_);
+    }
+};
+
 class ForNode : public TemplateNode {
     std::vector<std::string> var_names;
     std::shared_ptr<Expression> iterable;
@@ -961,7 +994,12 @@ class ForNode : public TemplateNode {
                   loop.set("last", i == (n - 1));
                   loop.set("previtem", i > 0 ? filtered_items.at(i - 1) : Value());
                   loop.set("nextitem", i < n - 1 ? filtered_items.at(i + 1) : Value());
-                  body->render(out, loop_context);
+                  try {
+                      body->render(out, loop_context);
+                  } catch (const LoopControlException & e) {
+                      if (e.control_type == LoopControlType::Break) break;
+                      if (e.control_type == LoopControlType::Continue) continue;
+                  }
               }
           }
       };
@@ -2159,7 +2197,7 @@ class Parser {
       static std::regex comment_tok(R"(\{#([-~]?)(.*?)([-~]?)#\})");
       static std::regex expr_open_regex(R"(\{\{([-~])?)");
       static std::regex block_open_regex(R"(^\{%([-~])?[\s\n\r]*)");
-      static std::regex block_keyword_tok(R"((if|else|elif|endif|for|endfor|generation|endgeneration|set|endset|block|endblock|macro|endmacro|filter|endfilter)\b)");
+      static std::regex block_keyword_tok(R"((if|else|elif|endif|for|endfor|generation|endgeneration|set|endset|block|endblock|macro|endmacro|filter|endfilter|break|continue)\b)");
       static std::regex non_text_open_regex(R"(\{\{|\{%|\{#)");
       static std::regex expr_close_regex(R"([\s\n\r]*([-~])?\}\})");
       static std::regex block_close_regex(R"([\s\n\r]*([-~])?%\})");
@@ -2291,6 +2329,9 @@ class Parser {
             } else if (keyword == "endfilter") {
               auto post_space = parseBlockClose();
               tokens.push_back(std::make_unique<EndFilterTemplateToken>(location, pre_space, post_space));
+            } else if (keyword == "break" || keyword == "continue") {
+              auto post_space = parseBlockClose();
+              tokens.push_back(std::make_unique<LoopControlTemplateToken>(location, pre_space, post_space, keyword == "break" ? LoopControlType::Break : LoopControlType::Continue));
             } else {
               throw std::runtime_error("Unexpected block: " + keyword);
             }
@@ -2414,6 +2455,8 @@ class Parser {
               children.emplace_back(std::make_shared<FilterNode>(token->location, std::move(filter_token->filter), std::move(body)));
           } else if (dynamic_cast<CommentTemplateToken*>(token.get())) {
               // Ignore comments
+          } else if (auto ctrl_token = dynamic_cast<LoopControlTemplateToken*>(token.get())) {
+              children.emplace_back(std::make_shared<LoopControlNode>(token->location, ctrl_token->control_type));
           } else if (dynamic_cast<EndForTemplateToken*>(token.get())
                   || dynamic_cast<EndSetTemplateToken*>(token.get())
                   || dynamic_cast<EndMacroTemplateToken*>(token.get())