dearwolf/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2025-05-20 10:35:01 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 21

refactor: backend/service split, channel-based llm flow (#1963)

Browse source 
Refactor: channel based llm flow and services split

---------

Signed-off-by: Dave Lee <dave@gray101.com>
This commit is contained in:
Dave 2024-04-13 03:45:34 -04:00 • committed by GitHub
parent 1981154f49
commit eed5706994
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
52 changed files with 3064 additions and 2279 deletions
Download patch file Download diff file Expand all files Collapse all files

78
core/http/endpoints/openai/edit.go
View file

@ -3,92 +3,36 @@ package openai
import (
"encoding/json"
"fmt"
"time"
"github.com/go-skynet/LocalAI/core/backend"
"github.com/go-skynet/LocalAI/core/config"
fiberContext "github.com/go-skynet/LocalAI/core/http/ctx"
"github.com/go-skynet/LocalAI/core/services"
"github.com/go-skynet/LocalAI/core/schema"
model "github.com/go-skynet/LocalAI/pkg/model"
"github.com/gofiber/fiber/v2"
"github.com/google/uuid"
"github.com/rs/zerolog/log"
)
func EditEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *config.ApplicationConfig) func(c *fiber.Ctx) error {
func EditEndpoint(fce *fiberContext.FiberContextExtractor, oais *services.OpenAIService) func(c *fiber.Ctx) error {
return func(c *fiber.Ctx) error {
modelFile, input, err := readRequest(c, ml, appConfig, true)
_, request, err := fce.OpenAIRequestFromContext(c, false)
if err != nil {
return fmt.Errorf("failed reading parameters from request:%w", err)
}
config, input, err := mergeRequestWithConfig(modelFile, input, cl, ml, appConfig.Debug, appConfig.Threads, appConfig.ContextSize, appConfig.F16)
_, finalResultChannel, _, _, _, err := oais.Edit(request, false, request.Stream)
if err != nil {
return fmt.Errorf("failed reading parameters from request:%w", err)
return err
}
log.Debug().Msgf("Parameter Config: %+v", config)
templateFile := ""
// A model can have a "file.bin.tmpl" file associated with a prompt template prefix
if ml.ExistsInModelPath(fmt.Sprintf("%s.tmpl", config.Model)) {
templateFile = config.Model
rawResponse := <-finalResultChannel
if rawResponse.Error != nil {
return rawResponse.Error
}
if config.TemplateConfig.Edit != "" {
templateFile = config.TemplateConfig.Edit
}
var result []schema.Choice
totalTokenUsage := backend.TokenUsage{}
for _, i := range config.InputStrings {
if templateFile != "" {
templatedInput, err := ml.EvaluateTemplateForPrompt(model.EditPromptTemplate, templateFile, model.PromptTemplateData{
Input: i,
Instruction: input.Instruction,
SystemPrompt: config.SystemPrompt,
})
if err == nil {
i = templatedInput
log.Debug().Msgf("Template found, input modified to: %s", i)
}
}
r, tokenUsage, err := ComputeChoices(input, i, config, appConfig, ml, func(s string, c *[]schema.Choice) {
*c = append(*c, schema.Choice{Text: s})
}, nil)
if err != nil {
return err
}
totalTokenUsage.Prompt += tokenUsage.Prompt
totalTokenUsage.Completion += tokenUsage.Completion
result = append(result, r...)
}
id := uuid.New().String()
created := int(time.Now().Unix())
resp := &schema.OpenAIResponse{
ID: id,
Created: created,
Model: input.Model, // we have to return what the user sent here, due to OpenAI spec.
Choices: result,
Object: "edit",
Usage: schema.OpenAIUsage{
PromptTokens: totalTokenUsage.Prompt,
CompletionTokens: totalTokenUsage.Completion,
TotalTokens: totalTokenUsage.Prompt + totalTokenUsage.Completion,
},
}
jsonResult, _ := json.Marshal(resp)
jsonResult, _ := json.Marshal(rawResponse.Value)
log.Debug().Msgf("Response: %s", jsonResult)
// Return the prediction in the response body
return c.JSON(resp)
return c.JSON(rawResponse.Value)
}
}