94 lines
2.8 KiB
Go
94 lines
2.8 KiB
Go
package model
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
|
|
"git.yeij.top/AskaEth/Cyrene_v2/internal/config"
|
|
)
|
|
|
|
// Step binds a model name to an Adapter for one entry in a fallback chain. An
|
|
// empty Model keeps the incoming Request.Model.
|
|
type Step struct {
|
|
Model string
|
|
Adapter Adapter
|
|
}
|
|
|
|
// Chain is an Adapter that tries its steps in order and returns the first
|
|
// successful response. It implements Adapter, so it can be used anywhere a
|
|
// single provider is expected.
|
|
type Chain struct {
|
|
steps []Step
|
|
}
|
|
|
|
// NewChain builds a fallback chain from the given steps.
|
|
func NewChain(steps ...Step) *Chain {
|
|
return &Chain{steps: steps}
|
|
}
|
|
|
|
// NewOpenAIChatChain builds a fallback chain from a single supplier config: the
|
|
// primary model first, then cfg.Fallback (when set and distinct), both served
|
|
// by an OpenAIChatAdapter against the same OpenAI-compatible endpoint. It
|
|
// delegates to the provider-aware builder so the OpenAI path and the factory stay
|
|
// in sync. It returns ErrNoModel when no base URL or primary model is configured.
|
|
func NewOpenAIChatChain(cfg config.LLMConfig, opts ...OpenAIOption) (*Chain, error) {
|
|
return newOpenAIChain(cfg, opts...)
|
|
}
|
|
|
|
// Name returns the chain identifier.
|
|
func (c *Chain) Name() string { return "fallback" }
|
|
|
|
// Chat tries each step in order and returns the first success. If every step
|
|
// fails, the last error is returned wrapped as ErrUpstream.
|
|
func (c *Chain) Chat(ctx context.Context, req Request, opts Options) (Response, error) {
|
|
if len(c.steps) == 0 {
|
|
return Response{}, ErrNoSteps
|
|
}
|
|
var lastErr error
|
|
for _, s := range c.steps {
|
|
stepReq := req
|
|
if s.Model != "" {
|
|
stepReq.Model = s.Model
|
|
}
|
|
resp, err := s.Adapter.Chat(ctx, stepReq, opts)
|
|
if err == nil {
|
|
return resp, nil
|
|
}
|
|
lastErr = err
|
|
}
|
|
return Response{}, fmt.Errorf("model: %w: %v", ErrUpstream, lastErr)
|
|
}
|
|
|
|
// ChatStream tries each step in order. If a step fails before delivering any
|
|
// delta, the chain moves to the next step; once a step has delivered output it
|
|
// is committed, so a mid-stream failure surfaces as an error rather than mixing
|
|
// content across providers.
|
|
func (c *Chain) ChatStream(ctx context.Context, req Request, onChunk func(delta string)) (Response, error) {
|
|
if len(c.steps) == 0 {
|
|
return Response{}, ErrNoSteps
|
|
}
|
|
var lastErr error
|
|
for _, s := range c.steps {
|
|
stepReq := req
|
|
if s.Model != "" {
|
|
stepReq.Model = s.Model
|
|
}
|
|
delivered := false
|
|
wrapped := func(delta string) {
|
|
delivered = true
|
|
if onChunk != nil {
|
|
onChunk(delta)
|
|
}
|
|
}
|
|
resp, err := s.Adapter.ChatStream(ctx, stepReq, wrapped)
|
|
if err == nil {
|
|
return resp, nil
|
|
}
|
|
if delivered {
|
|
return Response{}, fmt.Errorf("model: %w: stream step %q failed after partial output: %v", ErrUpstream, s.Model, err)
|
|
}
|
|
lastErr = err
|
|
}
|
|
return Response{}, fmt.Errorf("model: %w: %v", ErrUpstream, lastErr)
|
|
}
|