12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262 |
- package cmd
- import (
- "archive/zip"
- "bytes"
- "context"
- "crypto/ed25519"
- "crypto/rand"
- "crypto/sha256"
- "encoding/pem"
- "errors"
- "fmt"
- "io"
- "log"
- "math"
- "net"
- "net/http"
- "os"
- "os/signal"
- "path/filepath"
- "regexp"
- "runtime"
- "slices"
- "strings"
- "syscall"
- "time"
- "github.com/containerd/console"
- "github.com/mattn/go-runewidth"
- "github.com/olekukonko/tablewriter"
- "github.com/spf13/cobra"
- "golang.org/x/crypto/ssh"
- "golang.org/x/term"
- "github.com/ollama/ollama/api"
- "github.com/ollama/ollama/auth"
- "github.com/ollama/ollama/envconfig"
- "github.com/ollama/ollama/format"
- "github.com/ollama/ollama/parser"
- "github.com/ollama/ollama/progress"
- "github.com/ollama/ollama/server"
- "github.com/ollama/ollama/types/errtypes"
- "github.com/ollama/ollama/types/model"
- "github.com/ollama/ollama/version"
- )
- func CreateHandler(cmd *cobra.Command, args []string) error {
- filename, _ := cmd.Flags().GetString("file")
- filename, err := filepath.Abs(filename)
- if err != nil {
- return err
- }
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- p := progress.NewProgress(os.Stderr)
- defer p.Stop()
- f, err := os.Open(filename)
- if err != nil {
- return err
- }
- defer f.Close()
- modelfile, err := parser.ParseFile(f)
- if err != nil {
- return err
- }
- home, err := os.UserHomeDir()
- if err != nil {
- return err
- }
- status := "transferring model data"
- spinner := progress.NewSpinner(status)
- p.Add(status, spinner)
- for i := range modelfile.Commands {
- switch modelfile.Commands[i].Name {
- case "model", "adapter":
- path := modelfile.Commands[i].Args
- if path == "~" {
- path = home
- } else if strings.HasPrefix(path, "~/") {
- path = filepath.Join(home, path[2:])
- }
- if !filepath.IsAbs(path) {
- path = filepath.Join(filepath.Dir(filename), path)
- }
- fi, err := os.Stat(path)
- if errors.Is(err, os.ErrNotExist) && modelfile.Commands[i].Name == "model" {
- continue
- } else if err != nil {
- return err
- }
- if fi.IsDir() {
- // this is likely a safetensors or pytorch directory
- // TODO make this work w/ adapters
- tempfile, err := tempZipFiles(path)
- if err != nil {
- return err
- }
- defer os.RemoveAll(tempfile)
- path = tempfile
- }
- digest, err := createBlob(cmd, client, path)
- if err != nil {
- return err
- }
- modelfile.Commands[i].Args = "@" + digest
- }
- }
- bars := make(map[string]*progress.Bar)
- fn := func(resp api.ProgressResponse) error {
- if resp.Digest != "" {
- spinner.Stop()
- bar, ok := bars[resp.Digest]
- if !ok {
- bar = progress.NewBar(fmt.Sprintf("pulling %s...", resp.Digest[7:19]), resp.Total, resp.Completed)
- bars[resp.Digest] = bar
- p.Add(resp.Digest, bar)
- }
- bar.Set(resp.Completed)
- } else if status != resp.Status {
- spinner.Stop()
- status = resp.Status
- spinner = progress.NewSpinner(status)
- p.Add(status, spinner)
- }
- return nil
- }
- quantize, _ := cmd.Flags().GetString("quantize")
- request := api.CreateRequest{Name: args[0], Modelfile: modelfile.String(), Quantize: quantize}
- if err := client.Create(cmd.Context(), &request, fn); err != nil {
- return err
- }
- return nil
- }
- func tempZipFiles(path string) (string, error) {
- tempfile, err := os.CreateTemp("", "ollama-tf")
- if err != nil {
- return "", err
- }
- defer tempfile.Close()
- zipfile := zip.NewWriter(tempfile)
- defer zipfile.Close()
- detectContentType := func(path string) (string, error) {
- f, err := os.Open(path)
- if err != nil {
- return "", err
- }
- defer f.Close()
- var b bytes.Buffer
- b.Grow(512)
- if _, err := io.CopyN(&b, f, 512); err != nil && !errors.Is(err, io.EOF) {
- return "", err
- }
- contentType, _, _ := strings.Cut(http.DetectContentType(b.Bytes()), ";")
- return contentType, nil
- }
- glob := func(pattern, contentType string) ([]string, error) {
- matches, err := filepath.Glob(pattern)
- if err != nil {
- return nil, err
- }
- for _, safetensor := range matches {
- if ct, err := detectContentType(safetensor); err != nil {
- return nil, err
- } else if ct != contentType {
- return nil, fmt.Errorf("invalid content type: expected %s for %s", ct, safetensor)
- }
- }
- return matches, nil
- }
- var files []string
- if st, _ := glob(filepath.Join(path, "model*.safetensors"), "application/octet-stream"); len(st) > 0 {
- // safetensors files might be unresolved git lfs references; skip if they are
- // covers model-x-of-y.safetensors, model.fp32-x-of-y.safetensors, model.safetensors
- files = append(files, st...)
- } else if pt, _ := glob(filepath.Join(path, "pytorch_model*.bin"), "application/zip"); len(pt) > 0 {
- // pytorch files might also be unresolved git lfs references; skip if they are
- // covers pytorch_model-x-of-y.bin, pytorch_model.fp32-x-of-y.bin, pytorch_model.bin
- files = append(files, pt...)
- } else if pt, _ := glob(filepath.Join(path, "consolidated*.pth"), "application/zip"); len(pt) > 0 {
- // pytorch files might also be unresolved git lfs references; skip if they are
- // covers consolidated.x.pth, consolidated.pth
- files = append(files, pt...)
- } else {
- return "", errors.New("no safetensors or torch files found")
- }
- // add configuration files, json files are detected as text/plain
- js, err := glob(filepath.Join(path, "*.json"), "text/plain")
- if err != nil {
- return "", err
- }
- files = append(files, js...)
- if tks, _ := glob(filepath.Join(path, "tokenizer.model"), "application/octet-stream"); len(tks) > 0 {
- // add tokenizer.model if it exists, tokenizer.json is automatically picked up by the previous glob
- // tokenizer.model might be a unresolved git lfs reference; error if it is
- files = append(files, tks...)
- } else if tks, _ := glob(filepath.Join(path, "**/tokenizer.model"), "text/plain"); len(tks) > 0 {
- // some times tokenizer.model is in a subdirectory (e.g. meta-llama/Meta-Llama-3-8B)
- files = append(files, tks...)
- }
- for _, file := range files {
- f, err := os.Open(file)
- if err != nil {
- return "", err
- }
- defer f.Close()
- fi, err := f.Stat()
- if err != nil {
- return "", err
- }
- zfi, err := zip.FileInfoHeader(fi)
- if err != nil {
- return "", err
- }
- zf, err := zipfile.CreateHeader(zfi)
- if err != nil {
- return "", err
- }
- if _, err := io.Copy(zf, f); err != nil {
- return "", err
- }
- }
- return tempfile.Name(), nil
- }
- func createBlob(cmd *cobra.Command, client *api.Client, path string) (string, error) {
- bin, err := os.Open(path)
- if err != nil {
- return "", err
- }
- defer bin.Close()
- hash := sha256.New()
- if _, err := io.Copy(hash, bin); err != nil {
- return "", err
- }
- if _, err := bin.Seek(0, io.SeekStart); err != nil {
- return "", err
- }
- digest := fmt.Sprintf("sha256:%x", hash.Sum(nil))
- if err = client.CreateBlob(cmd.Context(), digest, bin); err != nil {
- return "", err
- }
- return digest, nil
- }
- func RunHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- name := args[0]
- // check if the model exists on the server
- show, err := client.Show(cmd.Context(), &api.ShowRequest{Name: name})
- var statusError api.StatusError
- switch {
- case errors.As(err, &statusError) && statusError.StatusCode == http.StatusNotFound:
- if err := PullHandler(cmd, []string{name}); err != nil {
- return err
- }
- show, err = client.Show(cmd.Context(), &api.ShowRequest{Name: name})
- if err != nil {
- return err
- }
- case err != nil:
- return err
- }
- interactive := true
- opts := runOptions{
- Model: args[0],
- WordWrap: os.Getenv("TERM") == "xterm-256color",
- Options: map[string]interface{}{},
- MultiModal: slices.Contains(show.Details.Families, "clip"),
- ParentModel: show.Details.ParentModel,
- }
- format, err := cmd.Flags().GetString("format")
- if err != nil {
- return err
- }
- opts.Format = format
- keepAlive, err := cmd.Flags().GetString("keepalive")
- if err != nil {
- return err
- }
- if keepAlive != "" {
- d, err := time.ParseDuration(keepAlive)
- if err != nil {
- return err
- }
- opts.KeepAlive = &api.Duration{Duration: d}
- }
- prompts := args[1:]
- // prepend stdin to the prompt if provided
- if !term.IsTerminal(int(os.Stdin.Fd())) {
- in, err := io.ReadAll(os.Stdin)
- if err != nil {
- return err
- }
- prompts = append([]string{string(in)}, prompts...)
- opts.WordWrap = false
- interactive = false
- }
- opts.Prompt = strings.Join(prompts, " ")
- if len(prompts) > 0 {
- interactive = false
- }
- nowrap, err := cmd.Flags().GetBool("nowordwrap")
- if err != nil {
- return err
- }
- opts.WordWrap = !nowrap
- if !interactive {
- return generate(cmd, opts)
- }
- return generateInteractive(cmd, opts)
- }
- func errFromUnknownKey(unknownKeyErr error) error {
- // find SSH public key in the error message
- sshKeyPattern := `ssh-\w+ [^\s"]+`
- re := regexp.MustCompile(sshKeyPattern)
- matches := re.FindStringSubmatch(unknownKeyErr.Error())
- if len(matches) > 0 {
- serverPubKey := matches[0]
- localPubKey, err := auth.GetPublicKey()
- if err != nil {
- return unknownKeyErr
- }
- if runtime.GOOS == "linux" && serverPubKey != localPubKey {
- // try the ollama service public key
- svcPubKey, err := os.ReadFile("/usr/share/ollama/.ollama/id_ed25519.pub")
- if err != nil {
- return unknownKeyErr
- }
- localPubKey = strings.TrimSpace(string(svcPubKey))
- }
- // check if the returned public key matches the local public key, this prevents adding a remote key to the user's account
- if serverPubKey != localPubKey {
- return unknownKeyErr
- }
- var msg strings.Builder
- msg.WriteString(unknownKeyErr.Error())
- msg.WriteString("\n\nYour ollama key is:\n")
- msg.WriteString(localPubKey)
- msg.WriteString("\nAdd your key at:\n")
- msg.WriteString("https://ollama.com/settings/keys")
- return errors.New(msg.String())
- }
- return unknownKeyErr
- }
- func PushHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- insecure, err := cmd.Flags().GetBool("insecure")
- if err != nil {
- return err
- }
- p := progress.NewProgress(os.Stderr)
- defer p.Stop()
- bars := make(map[string]*progress.Bar)
- var status string
- var spinner *progress.Spinner
- fn := func(resp api.ProgressResponse) error {
- if resp.Digest != "" {
- if spinner != nil {
- spinner.Stop()
- }
- bar, ok := bars[resp.Digest]
- if !ok {
- bar = progress.NewBar(fmt.Sprintf("pushing %s...", resp.Digest[7:19]), resp.Total, resp.Completed)
- bars[resp.Digest] = bar
- p.Add(resp.Digest, bar)
- }
- bar.Set(resp.Completed)
- } else if status != resp.Status {
- if spinner != nil {
- spinner.Stop()
- }
- status = resp.Status
- spinner = progress.NewSpinner(status)
- p.Add(status, spinner)
- }
- return nil
- }
- request := api.PushRequest{Name: args[0], Insecure: insecure}
- if err := client.Push(cmd.Context(), &request, fn); err != nil {
- if spinner != nil {
- spinner.Stop()
- }
- if strings.Contains(err.Error(), "access denied") {
- return errors.New("you are not authorized to push to this namespace, create the model under a namespace you own")
- }
- host := model.ParseName(args[0]).Host
- isOllamaHost := strings.HasSuffix(host, ".ollama.ai") || strings.HasSuffix(host, ".ollama.com")
- if strings.Contains(err.Error(), errtypes.UnknownOllamaKeyErrMsg) && isOllamaHost {
- // the user has not added their ollama key to ollama.com
- // re-throw an error with a more user-friendly message
- return errFromUnknownKey(err)
- }
- return err
- }
- spinner.Stop()
- return nil
- }
- func ListHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- models, err := client.List(cmd.Context())
- if err != nil {
- return err
- }
- var data [][]string
- for _, m := range models.Models {
- if len(args) == 0 || strings.HasPrefix(m.Name, args[0]) {
- data = append(data, []string{m.Name, m.Digest[:12], format.HumanBytes(m.Size), format.HumanTime(m.ModifiedAt, "Never")})
- }
- }
- table := tablewriter.NewWriter(os.Stdout)
- table.SetHeader([]string{"NAME", "ID", "SIZE", "MODIFIED"})
- table.SetHeaderAlignment(tablewriter.ALIGN_LEFT)
- table.SetAlignment(tablewriter.ALIGN_LEFT)
- table.SetHeaderLine(false)
- table.SetBorder(false)
- table.SetNoWhiteSpace(true)
- table.SetTablePadding("\t")
- table.AppendBulk(data)
- table.Render()
- return nil
- }
- func ListRunningHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- models, err := client.ListRunning(cmd.Context())
- if err != nil {
- return err
- }
- var data [][]string
- for _, m := range models.Models {
- if len(args) == 0 || strings.HasPrefix(m.Name, args[0]) {
- var procStr string
- switch {
- case m.SizeVRAM == 0:
- procStr = "100% CPU"
- case m.SizeVRAM == m.Size:
- procStr = "100% GPU"
- case m.SizeVRAM > m.Size || m.Size == 0:
- procStr = "Unknown"
- default:
- sizeCPU := m.Size - m.SizeVRAM
- cpuPercent := math.Round(float64(sizeCPU) / float64(m.Size) * 100)
- procStr = fmt.Sprintf("%d%%/%d%% CPU/GPU", int(cpuPercent), int(100-cpuPercent))
- }
- data = append(data, []string{m.Name, m.Digest[:12], format.HumanBytes(m.Size), procStr, format.HumanTime(m.ExpiresAt, "Never")})
- }
- }
- table := tablewriter.NewWriter(os.Stdout)
- table.SetHeader([]string{"NAME", "ID", "SIZE", "PROCESSOR", "UNTIL"})
- table.SetHeaderAlignment(tablewriter.ALIGN_LEFT)
- table.SetAlignment(tablewriter.ALIGN_LEFT)
- table.SetHeaderLine(false)
- table.SetBorder(false)
- table.SetNoWhiteSpace(true)
- table.SetTablePadding("\t")
- table.AppendBulk(data)
- table.Render()
- return nil
- }
- func DeleteHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- for _, name := range args {
- req := api.DeleteRequest{Name: name}
- if err := client.Delete(cmd.Context(), &req); err != nil {
- return err
- }
- fmt.Printf("deleted '%s'\n", name)
- }
- return nil
- }
- func ShowHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- if len(args) != 1 {
- return errors.New("missing model name")
- }
- license, errLicense := cmd.Flags().GetBool("license")
- modelfile, errModelfile := cmd.Flags().GetBool("modelfile")
- parameters, errParams := cmd.Flags().GetBool("parameters")
- system, errSystem := cmd.Flags().GetBool("system")
- template, errTemplate := cmd.Flags().GetBool("template")
- for _, boolErr := range []error{errLicense, errModelfile, errParams, errSystem, errTemplate} {
- if boolErr != nil {
- return errors.New("error retrieving flags")
- }
- }
- flagsSet := 0
- showType := ""
- if license {
- flagsSet++
- showType = "license"
- }
- if modelfile {
- flagsSet++
- showType = "modelfile"
- }
- if parameters {
- flagsSet++
- showType = "parameters"
- }
- if system {
- flagsSet++
- showType = "system"
- }
- if template {
- flagsSet++
- showType = "template"
- }
- if flagsSet > 1 {
- return errors.New("only one of '--license', '--modelfile', '--parameters', '--system', or '--template' can be specified")
- } else if flagsSet == 0 {
- return errors.New("one of '--license', '--modelfile', '--parameters', '--system', or '--template' must be specified")
- }
- req := api.ShowRequest{Name: args[0]}
- resp, err := client.Show(cmd.Context(), &req)
- if err != nil {
- return err
- }
- switch showType {
- case "license":
- fmt.Println(resp.License)
- case "modelfile":
- fmt.Println(resp.Modelfile)
- case "parameters":
- fmt.Println(resp.Parameters)
- case "system":
- fmt.Println(resp.System)
- case "template":
- fmt.Println(resp.Template)
- }
- return nil
- }
- func CopyHandler(cmd *cobra.Command, args []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- req := api.CopyRequest{Source: args[0], Destination: args[1]}
- if err := client.Copy(cmd.Context(), &req); err != nil {
- return err
- }
- fmt.Printf("copied '%s' to '%s'\n", args[0], args[1])
- return nil
- }
- func PullHandler(cmd *cobra.Command, args []string) error {
- insecure, err := cmd.Flags().GetBool("insecure")
- if err != nil {
- return err
- }
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- p := progress.NewProgress(os.Stderr)
- defer p.Stop()
- bars := make(map[string]*progress.Bar)
- var status string
- var spinner *progress.Spinner
- fn := func(resp api.ProgressResponse) error {
- if resp.Digest != "" {
- if spinner != nil {
- spinner.Stop()
- }
- bar, ok := bars[resp.Digest]
- if !ok {
- bar = progress.NewBar(fmt.Sprintf("pulling %s...", resp.Digest[7:19]), resp.Total, resp.Completed)
- bars[resp.Digest] = bar
- p.Add(resp.Digest, bar)
- }
- bar.Set(resp.Completed)
- } else if status != resp.Status {
- if spinner != nil {
- spinner.Stop()
- }
- status = resp.Status
- spinner = progress.NewSpinner(status)
- p.Add(status, spinner)
- }
- return nil
- }
- request := api.PullRequest{Name: args[0], Insecure: insecure}
- if err := client.Pull(cmd.Context(), &request, fn); err != nil {
- return err
- }
- return nil
- }
- type generateContextKey string
- type runOptions struct {
- Model string
- ParentModel string
- Prompt string
- Messages []api.Message
- WordWrap bool
- Format string
- System string
- Template string
- Images []api.ImageData
- Options map[string]interface{}
- MultiModal bool
- KeepAlive *api.Duration
- }
- type displayResponseState struct {
- lineLength int
- wordBuffer string
- }
- func displayResponse(content string, wordWrap bool, state *displayResponseState) {
- termWidth, _, _ := term.GetSize(int(os.Stdout.Fd()))
- if wordWrap && termWidth >= 10 {
- for _, ch := range content {
- if state.lineLength+1 > termWidth-5 {
- if runewidth.StringWidth(state.wordBuffer) > termWidth-10 {
- fmt.Printf("%s%c", state.wordBuffer, ch)
- state.wordBuffer = ""
- state.lineLength = 0
- continue
- }
- // backtrack the length of the last word and clear to the end of the line
- a := runewidth.StringWidth(state.wordBuffer)
- if a > 0 {
- fmt.Printf("\x1b[%dD", a)
- }
- fmt.Printf("\x1b[K\n")
- fmt.Printf("%s%c", state.wordBuffer, ch)
- chWidth := runewidth.RuneWidth(ch)
- state.lineLength = runewidth.StringWidth(state.wordBuffer) + chWidth
- } else {
- fmt.Print(string(ch))
- state.lineLength += runewidth.RuneWidth(ch)
- if runewidth.RuneWidth(ch) >= 2 {
- state.wordBuffer = ""
- continue
- }
- switch ch {
- case ' ':
- state.wordBuffer = ""
- case '\n':
- state.lineLength = 0
- default:
- state.wordBuffer += string(ch)
- }
- }
- }
- } else {
- fmt.Printf("%s%s", state.wordBuffer, content)
- if len(state.wordBuffer) > 0 {
- state.wordBuffer = ""
- }
- }
- }
- func chat(cmd *cobra.Command, opts runOptions) (*api.Message, error) {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return nil, err
- }
- p := progress.NewProgress(os.Stderr)
- defer p.StopAndClear()
- spinner := progress.NewSpinner("")
- p.Add("", spinner)
- cancelCtx, cancel := context.WithCancel(cmd.Context())
- defer cancel()
- sigChan := make(chan os.Signal, 1)
- signal.Notify(sigChan, syscall.SIGINT)
- go func() {
- <-sigChan
- cancel()
- }()
- var state *displayResponseState = &displayResponseState{}
- var latest api.ChatResponse
- var fullResponse strings.Builder
- var role string
- fn := func(response api.ChatResponse) error {
- p.StopAndClear()
- latest = response
- role = response.Message.Role
- content := response.Message.Content
- fullResponse.WriteString(content)
- displayResponse(content, opts.WordWrap, state)
- return nil
- }
- req := &api.ChatRequest{
- Model: opts.Model,
- Messages: opts.Messages,
- Format: opts.Format,
- Options: opts.Options,
- }
- if opts.KeepAlive != nil {
- req.KeepAlive = opts.KeepAlive
- }
- if err := client.Chat(cancelCtx, req, fn); err != nil {
- if errors.Is(err, context.Canceled) {
- return nil, nil
- }
- return nil, err
- }
- if len(opts.Messages) > 0 {
- fmt.Println()
- fmt.Println()
- }
- verbose, err := cmd.Flags().GetBool("verbose")
- if err != nil {
- return nil, err
- }
- if verbose {
- latest.Summary()
- }
- return &api.Message{Role: role, Content: fullResponse.String()}, nil
- }
- func generate(cmd *cobra.Command, opts runOptions) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- p := progress.NewProgress(os.Stderr)
- defer p.StopAndClear()
- spinner := progress.NewSpinner("")
- p.Add("", spinner)
- var latest api.GenerateResponse
- generateContext, ok := cmd.Context().Value(generateContextKey("context")).([]int)
- if !ok {
- generateContext = []int{}
- }
- ctx, cancel := context.WithCancel(cmd.Context())
- defer cancel()
- sigChan := make(chan os.Signal, 1)
- signal.Notify(sigChan, syscall.SIGINT)
- go func() {
- <-sigChan
- cancel()
- }()
- var state *displayResponseState = &displayResponseState{}
- fn := func(response api.GenerateResponse) error {
- p.StopAndClear()
- latest = response
- content := response.Response
- displayResponse(content, opts.WordWrap, state)
- return nil
- }
- if opts.MultiModal {
- opts.Prompt, opts.Images, err = extractFileData(opts.Prompt)
- if err != nil {
- return err
- }
- }
- request := api.GenerateRequest{
- Model: opts.Model,
- Prompt: opts.Prompt,
- Context: generateContext,
- Images: opts.Images,
- Format: opts.Format,
- System: opts.System,
- Template: opts.Template,
- Options: opts.Options,
- KeepAlive: opts.KeepAlive,
- }
- if err := client.Generate(ctx, &request, fn); err != nil {
- if errors.Is(err, context.Canceled) {
- return nil
- }
- return err
- }
- if opts.Prompt != "" {
- fmt.Println()
- fmt.Println()
- }
- if !latest.Done {
- return nil
- }
- verbose, err := cmd.Flags().GetBool("verbose")
- if err != nil {
- return err
- }
- if verbose {
- latest.Summary()
- }
- ctx = context.WithValue(cmd.Context(), generateContextKey("context"), latest.Context)
- cmd.SetContext(ctx)
- return nil
- }
- func RunServer(cmd *cobra.Command, _ []string) error {
- // retrieve the OLLAMA_HOST environment variable
- ollamaHost, err := api.GetOllamaHost()
- if err != nil {
- return err
- }
- if err := initializeKeypair(); err != nil {
- return err
- }
- ln, err := net.Listen("tcp", net.JoinHostPort(ollamaHost.Host, ollamaHost.Port))
- if err != nil {
- return err
- }
- err = server.Serve(ln)
- if errors.Is(err, http.ErrServerClosed) {
- return nil
- }
- return err
- }
- func initializeKeypair() error {
- home, err := os.UserHomeDir()
- if err != nil {
- return err
- }
- privKeyPath := filepath.Join(home, ".ollama", "id_ed25519")
- pubKeyPath := filepath.Join(home, ".ollama", "id_ed25519.pub")
- _, err = os.Stat(privKeyPath)
- if os.IsNotExist(err) {
- fmt.Printf("Couldn't find '%s'. Generating new private key.\n", privKeyPath)
- cryptoPublicKey, cryptoPrivateKey, err := ed25519.GenerateKey(rand.Reader)
- if err != nil {
- return err
- }
- privateKeyBytes, err := ssh.MarshalPrivateKey(cryptoPrivateKey, "")
- if err != nil {
- return err
- }
- if err := os.MkdirAll(filepath.Dir(privKeyPath), 0o755); err != nil {
- return fmt.Errorf("could not create directory %w", err)
- }
- if err := os.WriteFile(privKeyPath, pem.EncodeToMemory(privateKeyBytes), 0o600); err != nil {
- return err
- }
- sshPublicKey, err := ssh.NewPublicKey(cryptoPublicKey)
- if err != nil {
- return err
- }
- publicKeyBytes := ssh.MarshalAuthorizedKey(sshPublicKey)
- if err := os.WriteFile(pubKeyPath, publicKeyBytes, 0o644); err != nil {
- return err
- }
- fmt.Printf("Your new public key is: \n\n%s\n", publicKeyBytes)
- }
- return nil
- }
- func checkServerHeartbeat(cmd *cobra.Command, _ []string) error {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return err
- }
- if err := client.Heartbeat(cmd.Context()); err != nil {
- if !strings.Contains(err.Error(), " refused") {
- return err
- }
- if err := startApp(cmd.Context(), client); err != nil {
- return fmt.Errorf("could not connect to ollama app, is it running?")
- }
- }
- return nil
- }
- func versionHandler(cmd *cobra.Command, _ []string) {
- client, err := api.ClientFromEnvironment()
- if err != nil {
- return
- }
- serverVersion, err := client.Version(cmd.Context())
- if err != nil {
- fmt.Println("Warning: could not connect to a running Ollama instance")
- }
- if serverVersion != "" {
- fmt.Printf("ollama version is %s\n", serverVersion)
- }
- if serverVersion != version.Version {
- fmt.Printf("Warning: client version is %s\n", version.Version)
- }
- }
- func appendEnvDocs(cmd *cobra.Command, envs []envconfig.EnvVar) {
- if len(envs) == 0 {
- return
- }
- envUsage := `
- Environment Variables:
- `
- for _, e := range envs {
- envUsage += fmt.Sprintf(" %-24s %s\n", e.Name, e.Description)
- }
- cmd.SetUsageTemplate(cmd.UsageTemplate() + envUsage)
- }
- func NewCLI() *cobra.Command {
- log.SetFlags(log.LstdFlags | log.Lshortfile)
- cobra.EnableCommandSorting = false
- if runtime.GOOS == "windows" {
- console.ConsoleFromFile(os.Stdin) //nolint:errcheck
- }
- rootCmd := &cobra.Command{
- Use: "ollama",
- Short: "Large language model runner",
- SilenceUsage: true,
- SilenceErrors: true,
- CompletionOptions: cobra.CompletionOptions{
- DisableDefaultCmd: true,
- },
- Run: func(cmd *cobra.Command, args []string) {
- if version, _ := cmd.Flags().GetBool("version"); version {
- versionHandler(cmd, args)
- return
- }
- cmd.Print(cmd.UsageString())
- },
- }
- rootCmd.Flags().BoolP("version", "v", false, "Show version information")
- createCmd := &cobra.Command{
- Use: "create MODEL",
- Short: "Create a model from a Modelfile",
- Args: cobra.ExactArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: CreateHandler,
- }
- createCmd.Flags().StringP("file", "f", "Modelfile", "Name of the Modelfile")
- createCmd.Flags().StringP("quantize", "q", "", "Quantize model to this level (e.g. q4_0)")
- showCmd := &cobra.Command{
- Use: "show MODEL",
- Short: "Show information for a model",
- Args: cobra.ExactArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: ShowHandler,
- }
- showCmd.Flags().Bool("license", false, "Show license of a model")
- showCmd.Flags().Bool("modelfile", false, "Show Modelfile of a model")
- showCmd.Flags().Bool("parameters", false, "Show parameters of a model")
- showCmd.Flags().Bool("template", false, "Show template of a model")
- showCmd.Flags().Bool("system", false, "Show system message of a model")
- runCmd := &cobra.Command{
- Use: "run MODEL [PROMPT]",
- Short: "Run a model",
- Args: cobra.MinimumNArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: RunHandler,
- }
- runCmd.Flags().String("keepalive", "", "Duration to keep a model loaded (e.g. 5m)")
- runCmd.Flags().Bool("verbose", false, "Show timings for response")
- runCmd.Flags().Bool("insecure", false, "Use an insecure registry")
- runCmd.Flags().Bool("nowordwrap", false, "Don't wrap words to the next line automatically")
- runCmd.Flags().String("format", "", "Response format (e.g. json)")
- serveCmd := &cobra.Command{
- Use: "serve",
- Aliases: []string{"start"},
- Short: "Start ollama",
- Args: cobra.ExactArgs(0),
- RunE: RunServer,
- }
- pullCmd := &cobra.Command{
- Use: "pull MODEL",
- Short: "Pull a model from a registry",
- Args: cobra.ExactArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: PullHandler,
- }
- pullCmd.Flags().Bool("insecure", false, "Use an insecure registry")
- pushCmd := &cobra.Command{
- Use: "push MODEL",
- Short: "Push a model to a registry",
- Args: cobra.ExactArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: PushHandler,
- }
- pushCmd.Flags().Bool("insecure", false, "Use an insecure registry")
- listCmd := &cobra.Command{
- Use: "list",
- Aliases: []string{"ls"},
- Short: "List models",
- PreRunE: checkServerHeartbeat,
- RunE: ListHandler,
- }
- psCmd := &cobra.Command{
- Use: "ps",
- Short: "List running models",
- PreRunE: checkServerHeartbeat,
- RunE: ListRunningHandler,
- }
- copyCmd := &cobra.Command{
- Use: "cp SOURCE DESTINATION",
- Short: "Copy a model",
- Args: cobra.ExactArgs(2),
- PreRunE: checkServerHeartbeat,
- RunE: CopyHandler,
- }
- deleteCmd := &cobra.Command{
- Use: "rm MODEL [MODEL...]",
- Short: "Remove a model",
- Args: cobra.MinimumNArgs(1),
- PreRunE: checkServerHeartbeat,
- RunE: DeleteHandler,
- }
- envVars := envconfig.AsMap()
- envs := []envconfig.EnvVar{envVars["OLLAMA_HOST"]}
- for _, cmd := range []*cobra.Command{
- createCmd,
- showCmd,
- runCmd,
- pullCmd,
- pushCmd,
- listCmd,
- psCmd,
- copyCmd,
- deleteCmd,
- serveCmd,
- } {
- switch cmd {
- case runCmd:
- appendEnvDocs(cmd, []envconfig.EnvVar{envVars["OLLAMA_HOST"], envVars["OLLAMA_NOHISTORY"]})
- case serveCmd:
- appendEnvDocs(cmd, []envconfig.EnvVar{
- envVars["OLLAMA_DEBUG"],
- envVars["OLLAMA_HOST"],
- envVars["OLLAMA_KEEP_ALIVE"],
- envVars["OLLAMA_MAX_LOADED_MODELS"],
- envVars["OLLAMA_MAX_QUEUE"],
- envVars["OLLAMA_MODELS"],
- envVars["OLLAMA_NUM_PARALLEL"],
- envVars["OLLAMA_NOPRUNE"],
- envVars["OLLAMA_ORIGINS"],
- envVars["OLLAMA_TMPDIR"],
- envVars["OLLAMA_FLASH_ATTENTION"],
- envVars["OLLAMA_LLM_LIBRARY"],
- envVars["OLLAMA_MAX_VRAM"],
- })
- default:
- appendEnvDocs(cmd, envs)
- }
- }
- rootCmd.AddCommand(
- serveCmd,
- createCmd,
- showCmd,
- runCmd,
- pullCmd,
- pushCmd,
- listCmd,
- psCmd,
- copyCmd,
- deleteCmd,
- )
- return rootCmd
- }
|