fyp/logic/models/train/train.go

482 lines
13 KiB
Go
Raw Normal View History

package models_train
import (
"database/sql"
"errors"
"fmt"
2023-09-27 21:20:39 +01:00
"io"
"net/http"
2023-09-27 21:20:39 +01:00
"os"
"os/exec"
"path"
"strconv"
"text/template"
model_classes "git.andr3h3nriqu3s.com/andr3/fyp/logic/models/classes"
. "git.andr3h3nriqu3s.com/andr3/fyp/logic/models/utils"
. "git.andr3h3nriqu3s.com/andr3/fyp/logic/utils"
)
func MakeDefenition(db *sql.DB, model_id string, target_accuracy int) (id string, err error) {
2023-09-27 21:20:39 +01:00
id = ""
2023-10-10 12:28:49 +01:00
rows, err := db.Query("insert into model_definition (model_id, target_accuracy) values ($1, $2) returning id;", model_id, target_accuracy)
if err != nil {
return
}
defer rows.Close()
if !rows.Next() {
return id, errors.New("Something wrong!")
}
2023-09-27 21:20:39 +01:00
err = rows.Scan(&id)
2023-10-10 12:28:49 +01:00
return
}
type ModelDefinitionStatus int
const (
2023-09-27 21:20:39 +01:00
MODEL_DEFINITION_STATUS_FAILED_TRAINING = -3
MODEL_DEFINITION_STATUS_PRE_INIT ModelDefinitionStatus = 1
MODEL_DEFINITION_STATUS_INIT = 2
MODEL_DEFINITION_STATUS_TRAINING = 3
MODEL_DEFINITION_STATUS_TRANIED = 4
MODEL_DEFINITION_STATUS_READY = 5
)
type LayerType int
const (
2023-10-10 12:28:49 +01:00
LAYER_INPUT LayerType = 1
LAYER_DENSE = 2
LAYER_FLATTEN = 3
)
func ModelDefinitionUpdateStatus(c *Context, id string, status ModelDefinitionStatus) (err error) {
_, err = c.Db.Exec("update model_definition set status = $1 where id = $2", status, id)
2023-09-27 21:20:39 +01:00
return
}
2023-09-27 21:20:39 +01:00
func MakeLayer(db *sql.DB, def_id string, layer_order int, layer_type LayerType, shape string) (err error) {
_, err = db.Exec("insert into model_definition_layer (def_id, layer_order, layer_type, shape) values ($1, $2, $3, $4)", def_id, layer_order, layer_type, shape)
return
}
2023-10-10 12:28:49 +01:00
func generateCvs(c *Context, run_path string, model_id string) (count int, err error) {
2023-10-02 21:15:31 +01:00
2023-10-10 12:28:49 +01:00
classes, err := c.Db.Query("select count(*) from model_classes where model_id=$1;", model_id)
if err != nil {
return
}
defer classes.Close()
if !classes.Next() {
return
}
if err = classes.Scan(&count); err != nil {
return
}
data, err := c.Db.Query("select mdp.id, mc.class_order, mdp.file_path from model_data_point as mdp inner join model_classes as mc on mc.id = mdp.class_id where mc.model_id = $1;", model_id)
if err != nil {
return
}
defer data.Close()
f, err := os.Create(path.Join(run_path, "train.csv"))
if err != nil {
return
}
defer f.Close()
f.Write([]byte("Id,Index\n"))
for data.Next() {
var id string
var class_order int
var file_path string
if err = data.Scan(&id, &class_order, &file_path); err != nil {
return
}
if file_path == "id://" {
f.Write([]byte(id + "," + strconv.Itoa(class_order) + "\n"))
} else {
return count, errors.New("TODO generateCvs to file_path " + file_path)
}
}
return
2023-10-02 21:15:31 +01:00
}
func trainDefinition(c *Context, model *BaseModel, definition_id string) (accuracy float64, err error) {
2023-09-27 21:20:39 +01:00
accuracy = 0
layers, err := c.Db.Query("select layer_type, shape from model_definition_layer where def_id=$1 order by layer_order asc;", definition_id)
2023-09-27 21:20:39 +01:00
if err != nil {
return
}
defer layers.Close()
type layerrow struct {
LayerType int
Shape string
}
got := []layerrow{}
for layers.Next() {
var row = layerrow{}
if err = layers.Scan(&row.LayerType, &row.Shape); err != nil {
return
}
row.Shape = shapeToSize(row.Shape)
got = append(got, row)
}
// Generate run folder
2023-10-03 19:02:02 +01:00
run_path := path.Join("/tmp", model.Id, "defs", definition_id)
2023-09-27 21:20:39 +01:00
err = os.MkdirAll(run_path, os.ModePerm)
if err != nil {
return
}
2023-10-10 12:28:49 +01:00
_, err = generateCvs(c, run_path, model.Id)
if err != nil {
return
}
2023-10-02 21:15:31 +01:00
2023-09-27 21:20:39 +01:00
// Create python script
f, err := os.Create(path.Join(run_path, "run.py"))
if err != nil {
return
}
defer f.Close()
tmpl, err := template.New("python_model_template.py").ParseFiles("views/py/python_model_template.py")
if err != nil {
return
}
if err = tmpl.Execute(f, AnyMap{
2023-10-10 12:28:49 +01:00
"Layers": got,
"Size": got[0].Shape,
"DataDir": path.Join(getDir(), "savedData", model.Id, "data"),
"RunPath": run_path,
"ColorMode": model.ImageMode,
2023-09-27 21:20:39 +01:00
}); err != nil {
return
}
// Run the command
2023-10-10 12:28:49 +01:00
out, err := exec.Command("bash", "-c", fmt.Sprintf("cd %s && python run.py", run_path)).Output()
if err != nil {
fmt.Println(string(out))
2023-09-27 21:20:39 +01:00
return
}
// Copy result around
2023-10-03 19:02:02 +01:00
result_path := path.Join("savedData", model.Id, "defs", definition_id)
2023-09-27 21:20:39 +01:00
if err = os.MkdirAll(result_path, os.ModePerm); err != nil {
return
}
if err = exec.Command("cp", "-r", path.Join(run_path, "model"), path.Join(result_path, "model")).Run(); err != nil {
return
}
accuracy_file, err := os.Open(path.Join(run_path, "accuracy.val"))
if err != nil {
return
}
defer accuracy_file.Close()
accuracy_file_bytes, err := io.ReadAll(accuracy_file)
if err != nil {
return
}
2023-10-10 12:28:49 +01:00
fmt.Println(string(accuracy_file_bytes))
2023-09-27 21:20:39 +01:00
accuracy, err = strconv.ParseFloat(string(accuracy_file_bytes), 64)
if err != nil {
return
}
2023-10-10 12:28:49 +01:00
os.RemoveAll(run_path)
2023-09-27 21:20:39 +01:00
return
}
func trainModel(c *Context, model *BaseModel) {
definitionsRows, err := c.Db.Query("select id, target_accuracy from model_definition where status=$1 and model_id=$2", MODEL_DEFINITION_STATUS_INIT, model.Id)
2023-09-27 21:20:39 +01:00
if err != nil {
c.Logger.Error("Failed to trainModel!Err:")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
}
defer definitionsRows.Close()
type row struct {
id string
target_accuracy int
}
definitions := []row{}
for definitionsRows.Next() {
var rowv row
if err = definitionsRows.Scan(&rowv.id, &rowv.target_accuracy); err != nil {
2023-10-10 12:28:49 +01:00
c.Logger.Error("Failed to train Model Could not read definition from db!Err:")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
}
definitions = append(definitions, rowv)
}
if len(definitions) == 0 {
2023-10-10 12:28:49 +01:00
c.Logger.Error("No Definitions defined!")
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
}
for _, def := range definitions {
accuracy, err := trainDefinition(c, model, def.id)
2023-09-27 21:20:39 +01:00
if err != nil {
2023-10-10 12:28:49 +01:00
c.Logger.Error("Failed to train definition!Err:")
c.Logger.Error(err)
ModelDefinitionUpdateStatus(c, def.id, MODEL_DEFINITION_STATUS_FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
continue
}
int_accuracy := int(accuracy * 100)
if int_accuracy < def.target_accuracy {
fmt.Printf("Failed to train definition! Accuracy less %d < %d\n", int_accuracy, def.target_accuracy)
ModelDefinitionUpdateStatus(c, def.id, MODEL_DEFINITION_STATUS_FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
continue
}
_, err = c.Db.Exec("update model_definition set accuracy=$1, status=$2 where id=$3", int_accuracy, MODEL_DEFINITION_STATUS_TRANIED, def.id)
2023-09-27 21:20:39 +01:00
if err != nil {
fmt.Printf("Failed to train definition!Err:\n")
fmt.Println(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
}
}
rows, err := c.Db.Query("select id from model_definition where model_id=$1 and status=$2 order by accuracy desc limit 1;", model.Id, MODEL_DEFINITION_STATUS_TRANIED)
2023-09-27 21:20:39 +01:00
if err != nil {
2023-10-10 12:28:49 +01:00
c.Logger.Error("DB: failed to read definition")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
}
2023-10-10 12:28:49 +01:00
defer rows.Close()
2023-09-27 21:20:39 +01:00
2023-10-10 12:28:49 +01:00
if !rows.Next() {
// TODO Make the Model status have a message
c.Logger.Error("All definitions failed to train!")
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
2023-10-10 12:28:49 +01:00
}
2023-09-27 21:20:39 +01:00
2023-10-10 12:28:49 +01:00
var id string
if err = rows.Scan(&id); err != nil {
c.Logger.Error("Failed to read id:")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
2023-10-10 12:28:49 +01:00
}
2023-10-10 12:28:49 +01:00
if _, err = c.Db.Exec("update model_definition set status=$1 where id=$2;", MODEL_DEFINITION_STATUS_READY, id); err != nil {
c.Logger.Error("Failed to update model definition")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
2023-10-10 12:28:49 +01:00
}
2023-09-27 21:20:39 +01:00
2023-10-10 12:28:49 +01:00
to_delete, err := c.Db.Query("select id from model_definition where status != $1 and model_id=$2", MODEL_DEFINITION_STATUS_READY, model.Id)
if err != nil {
c.Logger.Error("Failed to select model_definition to delete")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
2023-10-10 12:28:49 +01:00
}
defer to_delete.Close()
for to_delete.Next() {
var id string
if to_delete.Scan(&id); err != nil {
c.Logger.Error("Failed to scan the id of a model_definition to delete")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
return
}
os.RemoveAll(path.Join("savedData", model.Id, "defs", id))
}
// TODO Check if returning also works here
if _, err = c.Db.Exec("delete from model_definition where status!=$1 and model_id=$2;", MODEL_DEFINITION_STATUS_READY, model.Id); err != nil {
c.Logger.Error("Failed to delete model_definition")
c.Logger.Error(err)
ModelUpdateStatus(c, model.Id, FAILED_TRAINING)
2023-09-27 21:20:39 +01:00
return
2023-10-10 12:28:49 +01:00
}
ModelUpdateStatus(c, model.Id, READY)
}
2023-10-10 12:28:49 +01:00
func removeFailedDataPoints(db *sql.DB, model *BaseModel) (err error) {
rows, err := db.Query("select id from model_data_point as mdp join model_classes as mc on mc.id=mpd.class_id where mc.model_id=$1 and mdp.status=-1;", model.Id)
if err != nil {
return
}
defer rows.Close()
base_path := path.Join("savedData", model.Id, "data")
for rows.Next() {
var dataPointId string
err = rows.Scan(&dataPointId)
if err != nil {
return
}
err = os.RemoveAll(path.Join(base_path, dataPointId + model.Format))
if err != nil {
return
}
}
return
}
func handleTrain(handle *Handle) {
handle.Post("/models/train", func(w http.ResponseWriter, r *http.Request, c *Context) *Error {
if !CheckAuthLevel(1, w, r, c) {
return nil
}
if c.Mode == JSON {
panic("TODO /models/train JSON")
}
r.ParseForm()
f := r.Form
number_of_models := 0
accuracy := 0
if !CheckId(f, "id") || CheckEmpty(f, "model_type") || !CheckNumber(f, "number_of_models", &number_of_models) || !CheckNumber(f, "accuracy", &accuracy) {
fmt.Println(
!CheckId(f, "id"), CheckEmpty(f, "model_type"), !CheckNumber(f, "number_of_models", &number_of_models), !CheckNumber(f, "accuracy", &accuracy),
)
// TODO improve this response
return ErrorCode(nil, 400, c.AddMap(nil))
}
id := f.Get("id")
model_type := f.Get("model_type")
// Its not used rn
_ = model_type
2023-09-27 21:20:39 +01:00
// TODO check if the model has data
/*rows, err := handle.Db.Query("select mc.name, mdp.file_path from model_classes as mc join model_data_point as mdp on mdp.class_id = mc.id where mdp.model_mode = 1 and mc.model_id = $1 limit 1;", id)
if err != nil {
return Error500(err)
}
defer rows.Close()
if !rows.Next() {
return Error500(err)
}
var name string
var file_path string
err = rows.Scan(&name, &file_path)
if err != nil {
return Error500(err)
}*/
model, err := GetBaseModel(handle.Db, id)
if err == ModelNotFoundError {
return ErrorCode(nil, http.StatusNotFound, c.AddMap(AnyMap{
"NotFoundMessage": "Model not found",
"GoBackLink": "/models",
}))
} else if err != nil {
// TODO improve this response
return Error500(err)
}
if model.Status != CONFIRM_PRE_TRAINING {
// TODO improve this response
return ErrorCode(nil, 400, c.AddMap(nil))
}
2023-09-27 21:20:39 +01:00
cls, err := model_classes.ListClasses(handle.Db, model.Id)
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
2023-09-27 21:20:39 +01:00
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
2023-10-10 12:28:49 +01:00
err = removeFailedDataPoints(c.Db, model.Id)
if err != nil {
return c.Error500(err)
}
2023-09-27 21:20:39 +01:00
var fid string
for i := 0; i < number_of_models; i++ {
2023-09-27 21:20:39 +01:00
def_id, err := MakeDefenition(handle.Db, model.Id, accuracy)
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
// TODO improve this response
return c.Error500(err)
}
2023-09-27 21:20:39 +01:00
if fid == "" {
fid = def_id
}
// TODO change shape of it depends on the type of the image
err = MakeLayer(handle.Db, def_id, 1, LAYER_INPUT, fmt.Sprintf("%d,%d,1", model.Width, model.Height))
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
2023-09-27 21:20:39 +01:00
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
err = MakeLayer(handle.Db, def_id, 4, LAYER_FLATTEN, fmt.Sprintf("%d,1", len(cls)))
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
2023-10-10 12:28:49 +01:00
err = MakeLayer(handle.Db, def_id, 5, LAYER_DENSE, fmt.Sprintf("%d,1", len(cls)*3))
2023-09-27 21:20:39 +01:00
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
2023-10-10 12:28:49 +01:00
// Using sparce
2023-10-03 11:55:22 +01:00
err = MakeLayer(handle.Db, def_id, 5, LAYER_DENSE, fmt.Sprintf("%d, 1", len(cls)))
2023-09-27 21:20:39 +01:00
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
err = ModelDefinitionUpdateStatus(c, def_id, MODEL_DEFINITION_STATUS_INIT)
2023-09-27 21:20:39 +01:00
if err != nil {
ModelUpdateStatus(c, model.Id, FAILED_PREPARING_TRAINING)
// TODO improve this response
return c.Error500(err)
2023-09-27 21:20:39 +01:00
}
}
2023-09-27 21:20:39 +01:00
// TODO start training with id fid
go trainModel(c, model)
ModelUpdateStatus(c, model.Id, TRAINING)
2023-09-27 21:20:39 +01:00
Redirect("/models/edit?id="+model.Id, c.Mode, w, r)
return nil
})
}