Fix Kimi-K2 tool-call parsing issues (#17376)

* Fix kimi-k2 parsing

* fix template & add more tests for kimi-k2

* Another fix for Kimi-K2 chat template.

* enable allow_toolcall_in_think for Kimi-K2

* Refine key-value separator and value end format

* Enable tool call in think for kimi-k2

* allow_toolcall_in_think is now tested with Kimi-K2

* Remove outdated TODO comment in XML tool call parser

Removed TODO comment about untested tool call feature.

* Rename function from "utf8_truncate_safe" to "utf8_truncate_safe_len"

Author: hksdpc255

common/chat-parser-xml-toolcall.cpp

@@ -724,16 +724,10 @@ inline void parse_msg_with_xml_tool_calls(common_chat_msg_parser & builder, cons
         if (reasoning_unclosed) {
             if (auto pos = content.find(end_think); pos == std::string::npos && builder.pos() != builder.input().size()) {
                 unclosed_reasoning_content += content;
-                if (form.allow_toolcall_in_think) {
-                    builder.move_to(tc->groups[0].begin);
-                    if (!builder.try_consume_xml_tool_calls(form)) {
-                        unclosed_reasoning_content += tool_call_start;
-                        builder.move_to(tc->groups[0].end);
-                    }
-                } else {
+                if (!(form.allow_toolcall_in_think && tc)) {
                     unclosed_reasoning_content += tool_call_start;
+                    continue;
                 }
-                continue;
             } else {
                 reasoning_unclosed = false;
                 std::string reasoning_content;
@@ -781,8 +775,12 @@ inline void parse_msg_with_xml_tool_calls(common_chat_msg_parser & builder, cons
                 }
             } else {
                 // This <tool_call> start is in thinking block, skip this tool call
-                auto pos = think_start + start_think.size();
-                unclosed_reasoning_content = content.substr(pos) + tool_call_start;
+                // This <tool_call> start is in thinking block
+                if (form.allow_toolcall_in_think) {
+                    unclosed_reasoning_content = content.substr(think_start + start_think.size());
+                } else {
+                    unclosed_reasoning_content = content.substr(think_start + start_think.size()) + tool_call_start;
+                }
                 reasoning_unclosed = true;
                 content.resize(think_start);
                 toolcall_in_think = true;
@@ -805,22 +803,43 @@ inline void parse_msg_with_xml_tool_calls(common_chat_msg_parser & builder, cons
         }
 
         // remove potential partial suffix
-        if (content.size() > 0 && builder.pos() == builder.input().size() && unclosed_reasoning_content.empty()) {
-            rstrip(content);
-            trim_potential_partial_word(content);
-            rstrip(content);
+        if (builder.pos() == builder.input().size()) {
+            if (unclosed_reasoning_content.empty()) {
+                rstrip(content);
+                trim_potential_partial_word(content);
+                rstrip(content);
+            } else {
+                rstrip(unclosed_reasoning_content);
+                trim_potential_partial_word(unclosed_reasoning_content);
+                rstrip(unclosed_reasoning_content);
+            }
+        }
+
+        // consume unclosed_reasoning_content if allow_toolcall_in_think is set
+        if (form.allow_toolcall_in_think && !unclosed_reasoning_content.empty()) {
+            if (builder.syntax().reasoning_format != COMMON_REASONING_FORMAT_NONE && !builder.syntax().reasoning_in_content) {
+                builder.add_reasoning_content(unclosed_reasoning_content);
+            } else {
+                if (content.empty()) {
+                    content = start_think + unclosed_reasoning_content;
+                } else {
+                    content += "\n\n" + start_think;
+                    content += unclosed_reasoning_content;
+                }
+            }
+            unclosed_reasoning_content.clear();
         }
 
         // Add content
-        if (content.size() != 0) {
+        if (!content.empty()) {
             // If there are multiple content blocks
             if (builder.syntax().reasoning_format != COMMON_REASONING_FORMAT_NONE && !builder.syntax().reasoning_in_content && builder.result().content.size() != 0) {
                 builder.add_content("\n\n");
             }
             builder.add_content(content);
         }
 
-        // This <tool_call> start is in thinking block, skip this tool call
+        // This <tool_call> start is in thinking block and toolcall_in_think not set, skip this tool call
         if (toolcall_in_think && !form.allow_toolcall_in_think) {
             continue;
         }
@@ -829,7 +848,7 @@ inline void parse_msg_with_xml_tool_calls(common_chat_msg_parser & builder, cons
         if (!tc) {
             GGML_ASSERT(builder.pos() == builder.input().size());
             GGML_ASSERT(unclosed_reasoning_content.empty());
-            GGML_ASSERT(!reasoning_unclosed);
+            if (!form.allow_toolcall_in_think) GGML_ASSERT(!reasoning_unclosed);
             break;
         }
 
@@ -854,7 +873,6 @@ inline void parse_msg_with_xml_tool_calls(common_chat_msg_parser & builder, cons
 
 /**
  * Parse content uses reasoning and XML-Style tool call
- * TODO: Note that form.allow_toolcall_in_think is not tested yet. If anyone confirms it works, this comment can be removed.
  */
 void common_chat_msg_parser::consume_reasoning_with_xml_tool_calls(const struct xml_tool_call_format & form, const std::string & start_think, const std::string & end_think) {
     parse_msg_with_xml_tool_calls(*this, form, start_think, end_think);

common/chat-parser-xml-toolcall.h

@@ -31,7 +31,7 @@ struct xml_tool_call_format {
     std::optional<std::string> last_val_end = std::nullopt;
     std::optional<std::string> last_tool_end = std::nullopt;
     bool trim_raw_argval = false;
-    bool allow_toolcall_in_think = false; // TODO: UNTESTED!!!
+    bool allow_toolcall_in_think = false;
 };
 
 // make a GBNF that accept any strings except those containing any of the forbidden strings.

common/chat-parser.cpp

@@ -917,12 +917,13 @@ static void common_chat_parse_kimi_k2(common_chat_msg_parser & builder) {
         form.tool_start  = "<|tool_call_begin|>";
         form.tool_sep    = "<|tool_call_argument_begin|>{";
         form.key_start   = "\"";
-        form.key_val_sep = "\": ";
-        form.val_end     = ", ";
+        form.key_val_sep = "\":";
+        form.val_end     = ",";
         form.tool_end    = "}<|tool_call_end|>";
         form.scope_end   = "<|tool_calls_section_end|>";
         form.raw_argval  = false;
         form.last_val_end = "";
+        form.allow_toolcall_in_think = true;
         return form;
     })();
     builder.consume_reasoning_with_xml_tool_calls(form, "<think>", "</think>");

models/templates/Kimi-K2-Instruct.jinja

@@ -14,7 +14,7 @@
 {%- endmacro %}
 
 {%- set tool_response_queue = namespace(ids=[]) -%}
-{%- set tool_call_counter = namespace(value=1) -%}
+{%- set tool_call_counter = namespace(value=0) -%}
 
 {%- if tools -%}
   <|im_system|>tool_declare<|im_middle|>{{ tools | tojson }}<|im_end|>
@@ -36,12 +36,8 @@
   {%- if message['role'] == 'assistant' and message.get('tool_calls') -%}
     {{render_content(message)}}<|tool_calls_section_begin|>
     {%- for tool_call in message['tool_calls'] -%}
-      {%- if tool_call['id'] is defined -%}
-        {%- set formatted_id = tool_call['id'] -%}
-      {%- else -%}
-        {%- set formatted_id = 'functions.' + tool_call['function']['name'] + ':' + (tool_call_counter.value | string) -%}
-        {%- set tool_call_counter.value = tool_call_counter.value + 1 -%}
-      {%- endif -%}
+      {%- set formatted_id = 'functions.' + tool_call['function']['name'] + ':' + (tool_call_counter.value | string) -%}
+      {%- set tool_call_counter.value = tool_call_counter.value + 1 -%}
       {%- set _ = tool_response_queue.ids.append(formatted_id) -%}
       <|tool_call_begin|>{{ formatted_id }}<|tool_call_argument_begin|>{% if tool_call['function']['arguments'] is string %}{{ tool_call['function']['arguments'] }}{% else %}{{ tool_call['function']['arguments'] | tojson }}{% endif %}<|tool_call_end|>
     {%- endfor -%}

models/templates/Kimi-K2-Thinking.jinja

@@ -25,17 +25,13 @@
 {%- endmacro -%}
 
 {%- set tool_response_queue = namespace(ids=[]) -%}
-{%- set tool_call_counter = namespace(value=1) -%}
+{%- set tool_call_counter = namespace(value=0) -%}
 
 {%- macro render_toolcalls(message) -%}
   <|tool_calls_section_begin|>
   {%- for tool_call in message['tool_calls'] -%}
-    {%- if tool_call['id'] is defined -%}
-      {%- set formatted_id = tool_call['id'] -%}
-    {%- else -%}
-      {%- set formatted_id = 'functions.' + tool_call['function']['name'] + ':' + (tool_call_counter.value | string) -%}
-      {%- set tool_call_counter.value = tool_call_counter.value + 1 -%}
-    {%- endif -%}
+    {%- set formatted_id = 'functions.' + tool_call['function']['name'] + ':' + (tool_call_counter.value | string) -%}
+    {%- set tool_call_counter.value = tool_call_counter.value + 1 -%}
     {%- set _ = tool_response_queue.ids.append(formatted_id) -%}
     <|tool_call_begin|>{{ formatted_id }}<|tool_call_argument_begin|>{% if tool_call['function']['arguments'] is string %}{{ tool_call['function']['arguments'] }}{% else %}{{ tool_call['function']['arguments'] | tojson }}{% endif %}<|tool_call_end|>
   {%- endfor -%}