Files
llama.cpp/common/chat-parser.h
hksdpc255 1920345c3b common : Generalized XML-style tool-call parsing with streaming support (GLM 4.5/4.6 + MiniMax M2 + SeedOSS + Kimi-K2 + Qwen3-Coder + Apriel-1.5 + Xiaomi-MiMo) (#16932)
* Add files via upload

* fix unit test

* fix crashes for --reasoning-format=none

* Patch buggy official MiniMax-M2 chat template

* add upstream minja fix: https://github.com/ochafik/minja/pull/7

* Fix <think> token not generated

* add test copied from https://github.com/ggml-org/llama.cpp/pull/16946

* cleanup

* Hopes to fix the compilation error on CI

* Delete chat template patching since it’s fixed by upstream Minja

* Remove undeeded Minimax-M2 template patch

https://github.com/ochafik/minja/pull/7#issuecomment-3480356100

* Add proper handling of optional parameters with test
merged tests from: 23d4bb75c4

* Fix making all tool parameters optional

* Move xml tool parser to separate file

* cleanup & add tests for GLM4.5

* add streaming tests & enhancement & cleanups

Add streaming test for both GLM 4.5 and minimax-m2.
Cleanup for preserved_tokens.
Cleanup for grammar rule name.
Enhance the parser's stability.

* cleanup & add support for Kimi-K2 Qwen3-Coder Apriel-1.5 Xiaomi-MiMo

* apply suggestions from reviewers

* fix a misuse for data.grammar_lazy

* fix grammar when tool have no argument

* Fix `no triggers set for lazy grammar!` for GLM4.5/4.6. Insert additional stops for Kimi-K2

* update chat.cpp

* fix grammar for GLM 4.5/4.6

* Try fix Jinja template for GLM

* Try fix GLM-4.6.jinja

* Update common/chat-parser-xml-toolcall.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* Update tests/test-chat.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* improve chat template for GLM, rename Kimi-K2 template to Kimi-K2-Thinking

* Improve Kimi-K2 chat template

* Fix unit test

* Fix "Invalid tool call arguments passed" in a rare case.

In a rare case, the model may emit a raw string that begins with a valid JSON string. This commit adds unit tests to cover that scenario and fixes the regression introduced during the Kimi-K2 adaptation.

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
2025-11-18 18:54:15 +01:00

134 lines
4.8 KiB
C++

#pragma once
#include "chat.h"
#include "chat-parser-xml-toolcall.h"
#include "json-partial.h"
#include "regex-partial.h"
#include <nlohmann/json.hpp>
#include <optional>
#include <string>
#include <vector>
class common_chat_msg_partial_exception : public std::runtime_error {
public:
common_chat_msg_partial_exception(const std::string & message) : std::runtime_error(message) {}
};
class common_chat_msg_parser {
std::string input_;
bool is_partial_;
common_chat_syntax syntax_;
std::string healing_marker_;
size_t pos_ = 0;
common_chat_msg result_;
public:
common_chat_msg_parser(const std::string & input, bool is_partial, const common_chat_syntax & syntax);
const std::string & input() const { return input_; }
size_t pos() const { return pos_; }
const std::string & healing_marker() const { return healing_marker_; }
const bool & is_partial() const { return is_partial_; }
const common_chat_msg & result() const { return result_; }
const common_chat_syntax & syntax() const { return syntax_; }
void move_to(size_t pos) {
if (pos > input_.size()) {
throw std::runtime_error("Invalid position!");
}
pos_ = pos;
}
void move_back(size_t n) {
if (pos_ < n) {
throw std::runtime_error("Can't move back that far!");
}
pos_ -= n;
}
// Get the substring of the input at the given range
std::string str(const common_string_range & rng) const;
// Appends to the result.content field
void add_content(const std::string & content);
// Appends to the result.reasoning_content field
void add_reasoning_content(const std::string & reasoning_content);
// Adds a tool call to the result. If the tool call is too incomplete (e.g. name empty), it won't add anything.
bool add_tool_call(const std::string & name, const std::string & id, const std::string & arguments);
// Adds a tool call using the "name", "id" and "arguments" fields of the json object
bool add_tool_call(const nlohmann::ordered_json & tool_call);
// Adds an array of tool calls using their "name", "id" and "arguments" fields.
bool add_tool_calls(const nlohmann::ordered_json & arr);
// Adds a tool call using the short form: { "tool_name": { "arg1": val, "arg2": val } }
bool add_tool_call_short_form(const nlohmann::ordered_json & tool_call);
void finish();
bool consume_spaces();
void consume_literal(const std::string & literal);
bool try_parse_reasoning(const std::string & start_think, const std::string & end_think);
std::string consume_rest();
struct find_regex_result {
std::string prelude;
std::vector<common_string_range> groups;
};
std::optional<find_regex_result> try_find_regex(const common_regex & regex, size_t from = std::string::npos, bool add_prelude_to_content = true);
bool try_consume_literal(const std::string & literal);
std::optional<find_regex_result> try_find_literal(const std::string & literal);
find_regex_result consume_regex(const common_regex & regex);
std::optional<find_regex_result> try_consume_regex(const common_regex & regex);
std::optional<common_json> try_consume_json();
common_json consume_json();
struct consume_json_result {
nlohmann::ordered_json value;
bool is_partial;
};
/*
Consume (possibly partial) json and converts specific subtrees to (possibly truncated) JSON strings.
By default, object keys can't be truncated, nor can string values (their corresponding key is removed,
e.g. `{"foo": "bar", "baz": "b` -> `{"foo": "bar"}`
But one can allow subpaths to be kept truncated, and possibly json-dumped to truncated json strings
- with `content_paths={{"foo"}}` -> `{"foo": "b` -> {"foo": "b"}`
- with `args_paths={{"foo"}}` -> `{"foo": {"b` -> `{"foo": "{b"}`
*/
consume_json_result consume_json_with_dumped_args(
const std::vector<std::vector<std::string>> & args_paths = {},
const std::vector<std::vector<std::string>> & content_paths = {}
);
std::optional<consume_json_result> try_consume_json_with_dumped_args(
const std::vector<std::vector<std::string>> & args_paths = {},
const std::vector<std::vector<std::string>> & content_paths = {}
);
/**
* Parse XML-Style tool call for given xml_tool_call_format. Return false for invalid syntax and get the position untouched.
* form.scope_start, form.tool_sep and form.scope_end can be empty.
*/
bool try_consume_xml_tool_calls(const struct xml_tool_call_format & form);
// Parse content uses reasoning and XML-Style tool call
void consume_reasoning_with_xml_tool_calls(const struct xml_tool_call_format & form, const std::string & start_think = "<think>", const std::string & end_think = "</think>");
void clear_tools();
};