mirror of
https://github.com/ollama/ollama.git
synced 2025-12-28 10:10:00 -05:00
Compare commits
5 Commits
main
...
grace/deep
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
e1878e6e33 | ||
|
|
f0733c13b5 | ||
|
|
07162c509f | ||
|
|
5be8277683 | ||
|
|
ec65cc3690 |
292
model/parsers/deepseek.go
Normal file
292
model/parsers/deepseek.go
Normal file
@@ -0,0 +1,292 @@
|
||||
package parsers
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"errors"
|
||||
"log/slog"
|
||||
"strings"
|
||||
"unicode"
|
||||
|
||||
"github.com/ollama/ollama/api"
|
||||
)
|
||||
|
||||
type DeepSeekParserState int
|
||||
|
||||
const (
|
||||
DeepSeekCollectingThinking DeepSeekParserState = iota
|
||||
DeepSeekCollectingContent
|
||||
DeepSeekCollectingToolCalls
|
||||
DeepSeekCollectingToolOutput
|
||||
)
|
||||
|
||||
const (
|
||||
deepseekThinkingCloseTag = "</think>"
|
||||
deepseekToolCallsBeginTag = "<|tool▁calls▁begin|>"
|
||||
deepseekToolCallsEndTag = "<|tool▁calls▁end|>"
|
||||
deepseekToolCallBeginTag = "<|tool▁call▁begin|>"
|
||||
deepseekToolCallEndTag = "<|tool▁call▁end|>"
|
||||
deepseekToolSepTag = "<|tool▁sep|>"
|
||||
deepseekToolOutputBeginTag = "<|tool▁output▁begin|>"
|
||||
deepseekToolOutputEndTag = "<|tool▁output▁end|>"
|
||||
)
|
||||
|
||||
type DeepSeekParser struct {
|
||||
state DeepSeekParserState
|
||||
buffer strings.Builder
|
||||
hasThinkingSupport bool
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) HasToolSupport() bool {
|
||||
return true
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) HasThinkingSupport() bool {
|
||||
return p.hasThinkingSupport
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) setInitialState(lastMessage *api.Message, tools []api.Tool, thinkValue *api.ThinkValue) {
|
||||
prefill := lastMessage != nil && lastMessage.Role == "assistant"
|
||||
|
||||
// Check both model capability AND request preference
|
||||
thinkingEnabled := p.HasThinkingSupport() && (thinkValue == nil || thinkValue.Bool())
|
||||
|
||||
if !thinkingEnabled {
|
||||
p.state = DeepSeekCollectingContent
|
||||
return
|
||||
}
|
||||
|
||||
if prefill && lastMessage.Content != "" {
|
||||
p.state = DeepSeekCollectingContent
|
||||
return
|
||||
}
|
||||
|
||||
p.state = DeepSeekCollectingThinking
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) Init(tools []api.Tool, lastMessage *api.Message, thinkValue *api.ThinkValue) []api.Tool {
|
||||
p.setInitialState(lastMessage, tools, thinkValue)
|
||||
return tools
|
||||
}
|
||||
|
||||
type deepseekEvent interface {
|
||||
isDeepSeekEvent()
|
||||
}
|
||||
|
||||
type deepseekEventThinkingContent struct {
|
||||
content string
|
||||
}
|
||||
|
||||
type deepseekEventContent struct {
|
||||
content string
|
||||
}
|
||||
|
||||
type deepseekEventToolCall struct {
|
||||
toolCall api.ToolCall
|
||||
}
|
||||
|
||||
func (deepseekEventThinkingContent) isDeepSeekEvent() {}
|
||||
func (deepseekEventContent) isDeepSeekEvent() {}
|
||||
func (deepseekEventToolCall) isDeepSeekEvent() {}
|
||||
|
||||
func (p *DeepSeekParser) Add(s string, done bool) (content string, thinking string, calls []api.ToolCall, err error) {
|
||||
p.buffer.WriteString(s)
|
||||
events := p.parseEvents()
|
||||
|
||||
var toolCalls []api.ToolCall
|
||||
var contentSb strings.Builder
|
||||
var thinkingSb strings.Builder
|
||||
for _, event := range events {
|
||||
switch event := event.(type) {
|
||||
case deepseekEventToolCall:
|
||||
toolCalls = append(toolCalls, event.toolCall)
|
||||
case deepseekEventThinkingContent:
|
||||
thinkingSb.WriteString(event.content)
|
||||
case deepseekEventContent:
|
||||
contentSb.WriteString(event.content)
|
||||
}
|
||||
}
|
||||
|
||||
return contentSb.String(), thinkingSb.String(), toolCalls, nil
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) parseEvents() []deepseekEvent {
|
||||
var all []deepseekEvent
|
||||
|
||||
keepLooping := true
|
||||
for keepLooping {
|
||||
var events []deepseekEvent
|
||||
events, keepLooping = p.eat()
|
||||
if len(events) > 0 {
|
||||
all = append(all, events...)
|
||||
}
|
||||
}
|
||||
|
||||
return all
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) eat() ([]deepseekEvent, bool) {
|
||||
var events []deepseekEvent
|
||||
bufStr := p.buffer.String()
|
||||
if bufStr == "" {
|
||||
return events, false
|
||||
}
|
||||
|
||||
switch p.state {
|
||||
case DeepSeekCollectingThinking:
|
||||
if strings.Contains(bufStr, deepseekThinkingCloseTag) { // thinking[</think>] -> content
|
||||
split := strings.SplitN(bufStr, deepseekThinkingCloseTag, 2)
|
||||
thinking := split[0]
|
||||
thinking = strings.TrimRightFunc(thinking, unicode.IsSpace)
|
||||
|
||||
remaining := split[1]
|
||||
remaining = strings.TrimLeftFunc(remaining, unicode.IsSpace)
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
p.state = DeepSeekCollectingContent
|
||||
|
||||
if len(thinking) > 0 {
|
||||
events = append(events, deepseekEventThinkingContent{content: thinking})
|
||||
}
|
||||
return events, true
|
||||
} else if overlapLen := overlap(bufStr, deepseekThinkingCloseTag); overlapLen > 0 { // partial </think>
|
||||
beforePartialTag := bufStr[:len(bufStr)-overlapLen]
|
||||
trailingLen := trailingWhitespaceLen(beforePartialTag)
|
||||
ambiguousStart := len(beforePartialTag) - trailingLen
|
||||
|
||||
unambiguous := bufStr[:ambiguousStart]
|
||||
ambiguous := bufStr[ambiguousStart:]
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(ambiguous)
|
||||
if len(unambiguous) > 0 {
|
||||
events = append(events, deepseekEventThinkingContent{content: unambiguous})
|
||||
}
|
||||
return events, false
|
||||
} else { // otherwise its thinking content
|
||||
whitespaceLen := trailingWhitespaceLen(bufStr)
|
||||
ambiguousStart := len(bufStr) - whitespaceLen
|
||||
|
||||
unambiguous := bufStr[:ambiguousStart]
|
||||
ambiguous := bufStr[ambiguousStart:]
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(ambiguous)
|
||||
if len(unambiguous) > 0 {
|
||||
events = append(events, deepseekEventThinkingContent{content: unambiguous})
|
||||
}
|
||||
return events, false
|
||||
}
|
||||
|
||||
case DeepSeekCollectingContent:
|
||||
switch {
|
||||
case strings.Contains(bufStr, deepseekToolCallsBeginTag): // content[<|tool▁calls▁begin|>] -> tool calls
|
||||
split := strings.SplitN(bufStr, deepseekToolCallsBeginTag, 2)
|
||||
contentBefore := strings.TrimRightFunc(split[0], unicode.IsSpace)
|
||||
remaining := split[1]
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
p.state = DeepSeekCollectingToolCalls
|
||||
|
||||
if len(contentBefore) > 0 {
|
||||
events = append(events, deepseekEventContent{content: contentBefore})
|
||||
}
|
||||
return events, true
|
||||
case strings.Contains(bufStr, deepseekToolOutputBeginTag): // content[<|tool▁output▁begin|>] -> tool output
|
||||
split := strings.SplitN(bufStr, deepseekToolOutputBeginTag, 2)
|
||||
contentBefore := split[0] // Don't trim whitespace - preserve spaces
|
||||
remaining := split[1]
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
p.state = DeepSeekCollectingToolOutput
|
||||
|
||||
if len(contentBefore) > 0 {
|
||||
events = append(events, deepseekEventContent{content: contentBefore})
|
||||
}
|
||||
return events, true
|
||||
default: // otherwise its content
|
||||
p.buffer.Reset()
|
||||
if len(bufStr) > 0 {
|
||||
events = append(events, deepseekEventContent{content: bufStr})
|
||||
}
|
||||
return events, false
|
||||
}
|
||||
|
||||
case DeepSeekCollectingToolCalls:
|
||||
if idx := strings.Index(bufStr, deepseekToolCallBeginTag); idx != -1 {
|
||||
startIdx := idx + len(deepseekToolCallBeginTag)
|
||||
if endIdx := strings.Index(bufStr[startIdx:], deepseekToolCallEndTag); endIdx != -1 {
|
||||
toolCallContent := bufStr[startIdx : startIdx+endIdx]
|
||||
|
||||
if toolCall, err := p.parseToolCallContent(toolCallContent); err == nil {
|
||||
remaining := bufStr[startIdx+endIdx+len(deepseekToolCallEndTag):]
|
||||
remaining = strings.TrimLeftFunc(remaining, unicode.IsSpace)
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
|
||||
events = append(events, deepseekEventToolCall{toolCall: toolCall})
|
||||
return events, true
|
||||
} else {
|
||||
slog.Warn("deepseek tool call parsing failed", "error", err)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if idx := strings.Index(bufStr, deepseekToolCallsEndTag); idx != -1 {
|
||||
remaining := bufStr[idx+len(deepseekToolCallsEndTag):]
|
||||
remaining = strings.TrimLeftFunc(remaining, unicode.IsSpace)
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
p.state = DeepSeekCollectingContent
|
||||
|
||||
return events, true
|
||||
}
|
||||
|
||||
return events, false
|
||||
|
||||
case DeepSeekCollectingToolOutput:
|
||||
if idx := strings.Index(bufStr, deepseekToolOutputEndTag); idx != -1 {
|
||||
toolOutputContent := bufStr[:idx]
|
||||
remaining := bufStr[idx+len(deepseekToolOutputEndTag):]
|
||||
// Don't trim whitespace - preserve spaces after tool output tags
|
||||
|
||||
p.buffer.Reset()
|
||||
p.buffer.WriteString(remaining)
|
||||
p.state = DeepSeekCollectingContent
|
||||
|
||||
if len(toolOutputContent) > 0 {
|
||||
events = append(events, deepseekEventContent{content: toolOutputContent})
|
||||
}
|
||||
return events, true
|
||||
}
|
||||
|
||||
return events, false
|
||||
}
|
||||
|
||||
return events, false
|
||||
}
|
||||
|
||||
func (p *DeepSeekParser) parseToolCallContent(content string) (api.ToolCall, error) {
|
||||
// Expected format: tool_name<|tool▁sep|>{args}
|
||||
parts := strings.SplitN(content, deepseekToolSepTag, 2)
|
||||
if len(parts) < 2 {
|
||||
return api.ToolCall{}, errors.New("invalid format")
|
||||
}
|
||||
|
||||
toolName := strings.TrimSpace(parts[0])
|
||||
argsJSON := strings.TrimSpace(parts[1])
|
||||
|
||||
var args api.ToolCallFunctionArguments
|
||||
if err := json.Unmarshal([]byte(argsJSON), &args); err != nil {
|
||||
return api.ToolCall{}, err
|
||||
}
|
||||
|
||||
return api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: toolName,
|
||||
Arguments: args,
|
||||
},
|
||||
}, nil
|
||||
}
|
||||
721
model/parsers/deepseek_test.go
Normal file
721
model/parsers/deepseek_test.go
Normal file
@@ -0,0 +1,721 @@
|
||||
package parsers
|
||||
|
||||
import (
|
||||
"testing"
|
||||
|
||||
"github.com/google/go-cmp/cmp"
|
||||
|
||||
"github.com/ollama/ollama/api"
|
||||
)
|
||||
|
||||
func TestDeepSeekParser(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
input string
|
||||
expectedContent string
|
||||
expectedThinking string
|
||||
expectedCalls []api.ToolCall
|
||||
hasThinking bool
|
||||
}{
|
||||
{
|
||||
name: "simple_content",
|
||||
input: "Hello, how are you?",
|
||||
expectedContent: "Hello, how are you?",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "thinking_content",
|
||||
input: "I need to think about this...</think>The answer is 42.",
|
||||
expectedThinking: "I need to think about this...",
|
||||
expectedContent: "The answer is 42.",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "no_thinking_simple",
|
||||
input: "Just a regular response.",
|
||||
expectedContent: "Just a regular response.",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "thinking_with_newlines",
|
||||
input: "Let me think:\n- Point 1\n- Point 2</think>\n\nHere's my answer.",
|
||||
expectedThinking: "Let me think:\n- Point 1\n- Point 2",
|
||||
expectedContent: "Here's my answer.",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "tool_call_simple",
|
||||
input: "I'll check the weather.<|tool▁calls▁begin|><|tool▁call▁begin|>get_weather<|tool▁sep|>{\"location\":\"Paris\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "I'll check the weather.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Paris",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "multiple_tool_calls",
|
||||
input: "Getting weather for both cities.<|tool▁calls▁begin|><|tool▁call▁begin|>get_weather<|tool▁sep|>{\"location\":\"Paris\"}<|tool▁call▁end|><|tool▁call▁begin|>get_weather<|tool▁sep|>{\"location\":\"London\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Getting weather for both cities.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Paris",
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "London",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "tool_output",
|
||||
input: "Here's the weather: <|tool▁output▁begin|>Temperature: 22°C, Sunny<|tool▁output▁end|> Hope that helps!",
|
||||
expectedContent: "Here's the weather: Temperature: 22°C, Sunny Hope that helps!",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "complex_tool_arguments",
|
||||
input: "Processing data.<|tool▁calls▁begin|><|tool▁call▁begin|>process_data<|tool▁sep|>{\"items\":[\"item1\",\"item2\"],\"config\":{\"enabled\":true,\"threshold\":0.95}}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Processing data.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "process_data",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"items": []interface{}{"item1", "item2"},
|
||||
"config": map[string]interface{}{"enabled": true, "threshold": 0.95},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "thinking_with_tool_call", // technically this can't happen, but the parser can handle it
|
||||
input: "Let me check the weather...</think>I'll get that for you.<|tool▁calls▁begin|><|tool▁call▁begin|>get_weather<|tool▁sep|>{\"location\":\"Paris\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedThinking: "Let me check the weather...",
|
||||
expectedContent: "I'll get that for you.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Paris",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "empty_content",
|
||||
input: "",
|
||||
expectedContent: "",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "only_thinking",
|
||||
input: "Just thinking content</think>",
|
||||
expectedThinking: "Just thinking content",
|
||||
expectedContent: "",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "multiple_tool_outputs",
|
||||
input: "Results: <|tool▁output▁begin|>Paris: 22°C<|tool▁output▁end|> and <|tool▁output▁begin|>London: 18°C<|tool▁output▁end|>",
|
||||
expectedContent: "Results: Paris: 22°C and London: 18°C",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "unicode_content",
|
||||
input: "مرحبا بالعالم! 你好世界! 🌍",
|
||||
expectedContent: "مرحبا بالعالم! 你好世界! 🌍",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "emoji_passthrough",
|
||||
input: "Task completed ✅ 🎉",
|
||||
expectedContent: "Task completed ✅ 🎉",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "emoji_after_tool_call",
|
||||
input: "I'll help you.<|tool▁calls▁begin|><|tool▁call▁begin|>get_weather<|tool▁sep|>{\"location\":\"Tokyo\"}<|tool▁call▁end|><|tool▁calls▁end|>完成 ✅",
|
||||
expectedContent: "I'll help you.完成 ✅",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Tokyo",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "newlines_and_whitespace",
|
||||
input: "Line 1\n\nLine 3\t\tTabbed content",
|
||||
expectedContent: "Line 1\n\nLine 3\t\tTabbed content",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "thinking_with_unicode",
|
||||
input: "我在思考这个问题...</think>答案是42。",
|
||||
expectedThinking: "我在思考这个问题...",
|
||||
expectedContent: "答案是42。",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "tool_call_with_unicode_args",
|
||||
input: "Searching for information.<|tool▁calls▁begin|><|tool▁call▁begin|>search<|tool▁sep|>{\"query\":\"北京天气\",\"language\":\"中文\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Searching for information.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "search",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"query": "北京天气",
|
||||
"language": "中文",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "tool_output_with_unicode",
|
||||
input: "天气信息: <|tool▁output▁begin|>北京: 25°C, 晴天<|tool▁output▁end|> 希望对您有帮助!",
|
||||
expectedContent: "天气信息: 北京: 25°C, 晴天 希望对您有帮助!",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "mixed_content_with_special_chars",
|
||||
input: "Price: $100 & tax @ 10% = $110 <|tool▁output▁begin|>Total: $110<|tool▁output▁end|> (final)",
|
||||
expectedContent: "Price: $100 & tax @ 10% = $110 Total: $110 (final)",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "tool_call_with_special_chars",
|
||||
input: "Processing data.<|tool▁calls▁begin|><|tool▁call▁begin|>execute_command<|tool▁sep|>{\"command\":\"ls && echo \\\"done\\\"\",\"path\":\"/home/user\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Processing data.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "execute_command",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"command": "ls && echo \"done\"",
|
||||
"path": "/home/user",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "thinking_with_special_chars",
|
||||
input: "Let me calculate: 2+2=4 & 3*3=9...</think>The results are correct!",
|
||||
expectedThinking: "Let me calculate: 2+2=4 & 3*3=9...",
|
||||
expectedContent: "The results are correct!",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "empty_tool_call_args",
|
||||
input: "Pinging server.<|tool▁calls▁begin|><|tool▁call▁begin|>ping<|tool▁sep|>{}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Pinging server.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "ping",
|
||||
Arguments: api.ToolCallFunctionArguments{},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "empty_tool_output",
|
||||
input: "Checking status: <|tool▁output▁begin|><|tool▁output▁end|> No output received.",
|
||||
expectedContent: "Checking status: No output received.",
|
||||
hasThinking: false,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
parser := &DeepSeekParser{hasThinkingSupport: tt.hasThinking}
|
||||
parser.Init([]api.Tool{}, nil, &api.ThinkValue{Value: tt.hasThinking})
|
||||
|
||||
content, thinking, calls, err := parser.Add(tt.input, true)
|
||||
if err != nil {
|
||||
t.Fatalf("Add() error = %v", err)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedContent, content); diff != "" {
|
||||
t.Errorf("Content mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedThinking, thinking); diff != "" {
|
||||
t.Errorf("Thinking mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedCalls, calls); diff != "" {
|
||||
t.Errorf("Tool calls mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_Streaming(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
chunks []string
|
||||
expectedContent string
|
||||
expectedThinking string
|
||||
expectedCalls []api.ToolCall
|
||||
hasThinking bool
|
||||
}{
|
||||
{
|
||||
name: "streaming_simple_content",
|
||||
chunks: []string{"Hello, ", "how are ", "you?"},
|
||||
expectedContent: "Hello, how are you?",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_thinking",
|
||||
chunks: []string{"I need to ", "think about this", "...</think>", "The answer is 42."},
|
||||
expectedThinking: "I need to think about this...",
|
||||
expectedContent: "The answer is 42.",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "streaming_tool_call",
|
||||
chunks: []string{"I'll check weather.", "<|tool▁calls▁begin|>", "<|tool▁call▁begin|>get_weather", "<|tool▁sep|>{\"location\":\"Paris\"}", "<|tool▁call▁end|><|tool▁calls▁end|>"},
|
||||
expectedContent: "I'll check weather.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Paris",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_thinking_with_partial_tag",
|
||||
chunks: []string{"Thinking about this", "...</", "think>", "Done thinking."},
|
||||
expectedThinking: "Thinking about this...",
|
||||
expectedContent: "Done thinking.",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "streaming_tool_output",
|
||||
chunks: []string{"Weather info: ", "<|tool▁output▁begin|>", "25°C, Sunny", "<|tool▁output▁end|>", " Enjoy!"},
|
||||
expectedContent: "Weather info: 25°C, Sunny Enjoy!",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_with_split_tags",
|
||||
chunks: []string{"Content before ", "<|tool▁calls▁begin|><|tool▁call▁begin|>test", "<|tool▁sep|>{}", "<|tool▁call▁end|><|tool▁calls▁end|>", " after"},
|
||||
expectedContent: "Content before after",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "test",
|
||||
Arguments: api.ToolCallFunctionArguments{},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_thinking_with_split_end_tag",
|
||||
chunks: []string{"Thinking content", "</th", "ink>", "Regular content"},
|
||||
expectedThinking: "Thinking content",
|
||||
expectedContent: "Regular content",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "streaming_unicode_content",
|
||||
chunks: []string{"مرحبا ", "بالعالم! ", "你好", "世界!"},
|
||||
expectedContent: "مرحبا بالعالم! 你好世界!",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_multiple_tool_outputs",
|
||||
chunks: []string{"Results: ", "<|tool▁output▁begin|>", "Paris: 22°C", "<|tool▁output▁end|>", " and ", "<|tool▁output▁begin|>", "London: 18°C", "<|tool▁output▁end|>"},
|
||||
expectedContent: "Results: Paris: 22°C and London: 18°C",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "streaming_tool_call_with_split_json",
|
||||
chunks: []string{"Processing.", "<|tool▁calls▁begin|><|tool▁call▁begin|>calc<|tool▁sep|>{\"x\":", "42,\"y\":", "24}<|tool▁call▁end|><|tool▁calls▁end|>"},
|
||||
expectedContent: "Processing.",
|
||||
expectedCalls: []api.ToolCall{
|
||||
{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "calc",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"x": float64(42),
|
||||
"y": float64(24),
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
hasThinking: false,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
parser := &DeepSeekParser{hasThinkingSupport: tt.hasThinking}
|
||||
parser.Init([]api.Tool{}, nil, &api.ThinkValue{Value: tt.hasThinking})
|
||||
|
||||
var allContent, allThinking string
|
||||
var allCalls []api.ToolCall
|
||||
|
||||
for i, chunk := range tt.chunks {
|
||||
done := i == len(tt.chunks)-1
|
||||
content, thinking, calls, err := parser.Add(chunk, done)
|
||||
if err != nil {
|
||||
t.Fatalf("Add() error = %v", err)
|
||||
}
|
||||
|
||||
allContent += content
|
||||
allThinking += thinking
|
||||
allCalls = append(allCalls, calls...)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedContent, allContent); diff != "" {
|
||||
t.Errorf("Content mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedThinking, allThinking); diff != "" {
|
||||
t.Errorf("Thinking mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedCalls, allCalls); diff != "" {
|
||||
t.Errorf("Tool calls mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_HasThinkingSupport(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
hasThinking bool
|
||||
expectedSupport bool
|
||||
}{
|
||||
{
|
||||
name: "thinking_enabled",
|
||||
hasThinking: true,
|
||||
expectedSupport: true,
|
||||
},
|
||||
{
|
||||
name: "thinking_disabled",
|
||||
hasThinking: false,
|
||||
expectedSupport: false,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
parser := &DeepSeekParser{hasThinkingSupport: tt.hasThinking}
|
||||
if got := parser.HasThinkingSupport(); got != tt.expectedSupport {
|
||||
t.Errorf("HasThinkingSupport() = %v, want %v", got, tt.expectedSupport)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_HasToolSupport(t *testing.T) {
|
||||
parser := &DeepSeekParser{}
|
||||
if !parser.HasToolSupport() {
|
||||
t.Error("HasToolSupport() should return true")
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_Init(t *testing.T) {
|
||||
parser := &DeepSeekParser{hasThinkingSupport: true}
|
||||
tools := []api.Tool{
|
||||
{
|
||||
Type: "function",
|
||||
Function: api.ToolFunction{
|
||||
Name: "test_tool",
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
returnedTools := parser.Init(tools, nil, &api.ThinkValue{Value: true})
|
||||
|
||||
if diff := cmp.Diff(tools, returnedTools); diff != "" {
|
||||
t.Errorf("Init() returned tools mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
// Test initial state is set to thinking when enabled
|
||||
if parser.state != DeepSeekCollectingThinking {
|
||||
t.Errorf("Expected initial state to be DeepSeekCollectingThinking, got %v", parser.state)
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_parseToolCallContent(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
content string
|
||||
expected api.ToolCall
|
||||
expectError bool
|
||||
}{
|
||||
{
|
||||
name: "valid_tool_call",
|
||||
content: "get_weather<|tool▁sep|>{\"location\":\"Paris\"}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "get_weather",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"location": "Paris",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "complex_arguments",
|
||||
content: "process_data<|tool▁sep|>{\"items\":[\"a\",\"b\"],\"config\":{\"enabled\":true}}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "process_data",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"items": []interface{}{"a", "b"},
|
||||
"config": map[string]interface{}{"enabled": true},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "empty_arguments",
|
||||
content: "ping<|tool▁sep|>{}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "ping",
|
||||
Arguments: api.ToolCallFunctionArguments{},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "unicode_in_tool_name",
|
||||
content: "获取天气<|tool▁sep|>{\"城市\":\"北京\"}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "获取天气",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"城市": "北京",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "special_chars_in_arguments",
|
||||
content: "execute<|tool▁sep|>{\"command\":\"ls && echo \\\"done\\\"\",\"path\":\"/home/user\"}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "execute",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"command": "ls && echo \"done\"",
|
||||
"path": "/home/user",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "numeric_arguments",
|
||||
content: "calculate<|tool▁sep|>{\"x\":3.14,\"y\":42,\"enabled\":true}",
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "calculate",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"x": 3.14,
|
||||
"y": float64(42),
|
||||
"enabled": true,
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "invalid_format_no_separator",
|
||||
content: "get_weather{\"location\":\"Paris\"}",
|
||||
expectError: true,
|
||||
},
|
||||
{
|
||||
name: "invalid_json",
|
||||
content: "get_weather<|tool▁sep|>{invalid json}",
|
||||
expectError: true,
|
||||
},
|
||||
{
|
||||
name: "empty_tool_name",
|
||||
content: "<|tool▁sep|>{\"arg\":\"value\"}",
|
||||
expectError: false, // This should work, just empty name
|
||||
expected: api.ToolCall{
|
||||
Function: api.ToolCallFunction{
|
||||
Name: "",
|
||||
Arguments: api.ToolCallFunctionArguments{
|
||||
"arg": "value",
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "missing_json_part",
|
||||
content: "tool_name<|tool▁sep|>",
|
||||
expectError: true,
|
||||
},
|
||||
}
|
||||
|
||||
parser := &DeepSeekParser{}
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
result, err := parser.parseToolCallContent(tt.content)
|
||||
|
||||
if tt.expectError {
|
||||
if err == nil {
|
||||
t.Error("Expected error but got none")
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
if err != nil {
|
||||
t.Fatalf("Unexpected error: %v", err)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expected, result); diff != "" {
|
||||
t.Errorf("parseToolCallContent() mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestDeepSeekParser_EdgeCases(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
input string
|
||||
expectedContent string
|
||||
expectedThinking string
|
||||
hasThinking bool
|
||||
}{
|
||||
{
|
||||
name: "nested_think_tags_in_thinking",
|
||||
input: "Outer thinking <think>inner</think> content</think>Final content",
|
||||
expectedThinking: "Outer thinking <think>inner",
|
||||
expectedContent: "content</think>Final content",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "multiple_think_close_tags",
|
||||
input: "First thought</think>Second thought</think>Final content",
|
||||
expectedThinking: "First thought",
|
||||
expectedContent: "Second thought</think>Final content",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "empty_thinking_content",
|
||||
input: "</think>Just content",
|
||||
expectedThinking: "",
|
||||
expectedContent: "Just content",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "thinking_disabled_with_think_tags",
|
||||
input: "Some content</think>More content",
|
||||
expectedContent: "Some content</think>More content",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "malformed_tool_call_missing_sep",
|
||||
input: "Testing.<|tool▁calls▁begin|><|tool▁call▁begin|>bad_tool{\"arg\":\"value\"}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Testing.",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "malformed_tool_call_invalid_json",
|
||||
input: "Testing.<|tool▁calls▁begin|><|tool▁call▁begin|>bad_tool<|tool▁sep|>{invalid json}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "Testing.",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "partial_tool_tag_at_end",
|
||||
input: "Content with partial <|tool▁calls▁",
|
||||
expectedContent: "Content with partial <|tool▁calls▁",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "partial_think_tag_at_end",
|
||||
input: "Thinking content</th",
|
||||
expectedContent: "Thinking content</th",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "partial_think_tag_at_end_with_thinking",
|
||||
input: "Thinking content</th",
|
||||
expectedThinking: "Thinking content",
|
||||
expectedContent: "",
|
||||
hasThinking: true,
|
||||
},
|
||||
{
|
||||
name: "whitespace_only_content",
|
||||
input: " \n\t ",
|
||||
expectedContent: " \n\t ",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "tool_output_with_newlines",
|
||||
input: "Output:\n<|tool▁output▁begin|>Line 1\nLine 2\nLine 3<|tool▁output▁end|>\nDone.",
|
||||
expectedContent: "Output:\nLine 1\nLine 2\nLine 3\nDone.",
|
||||
hasThinking: false,
|
||||
},
|
||||
{
|
||||
name: "consecutive_tool_calls",
|
||||
input: "First.<|tool▁calls▁begin|><|tool▁call▁begin|>tool1<|tool▁sep|>{}<|tool▁call▁end|><|tool▁calls▁end|>Second.<|tool▁calls▁begin|><|tool▁call▁begin|>tool2<|tool▁sep|>{}<|tool▁call▁end|><|tool▁calls▁end|>",
|
||||
expectedContent: "First.",
|
||||
hasThinking: false,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
parser := &DeepSeekParser{hasThinkingSupport: tt.hasThinking}
|
||||
parser.Init([]api.Tool{}, nil, &api.ThinkValue{Value: tt.hasThinking})
|
||||
|
||||
content, thinking, _, err := parser.Add(tt.input, true)
|
||||
if err != nil {
|
||||
t.Fatalf("Add() error = %v", err)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedContent, content); diff != "" {
|
||||
t.Errorf("Content mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
|
||||
if diff := cmp.Diff(tt.expectedThinking, thinking); diff != "" {
|
||||
t.Errorf("Thinking mismatch (-want +got):\n%s", diff)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
@@ -58,6 +58,8 @@ func ParserForName(name string) Parser {
|
||||
return harmony.NewHarmonyMessageHandler()
|
||||
case "cogito":
|
||||
return &CogitoParser{}
|
||||
case "deepseek":
|
||||
return &DeepSeekParser{hasThinkingSupport: true}
|
||||
case "olmo3":
|
||||
return &Olmo3Parser{}
|
||||
case "olmo3-think":
|
||||
|
||||
Reference in New Issue
Block a user