huge refactoring

2026-08-17 20:52:14 +01:00 · 2018-03-29 20:13:19 +02:00
parent 2f97a44bed
commit e847c86baa
24 changed files with 1039 additions and 783 deletions
--- a/app/handlers/context/context.go
+++ b/app/handlers/context/context.go
@@ -0,0 +1,51 @@
+package context
+
+import (
+	"context"
+	"net/http"
+
+	"github.com/gulien/gotenberg/app/handlers/converter"
+
+	ghttp "github.com/gulien/gotenberg/app/handlers/http"
+)
+
+type key uint32
+
+const (
+	contentTypeKey key = iota
+	converterKey
+)
+
+func WithContentType(r *http.Request, contentType ghttp.ContentType) *http.Request {
+	ctx := r.Context()
+	ctx = context.WithValue(ctx, contentTypeKey, contentType)
+	r = r.WithContext(ctx)
+
+	return r
+}
+
+func GetContentType(r *http.Request) (ghttp.ContentType, error) {
+	ct, ok := r.Context().Value(contentTypeKey).(ghttp.ContentType)
+	if !ok {
+		return "", &contentTypeNotFoundError{}
+	}
+
+	return ct, nil
+}
+
+func WithConverter(r *http.Request, converter *converter.Converter) *http.Request {
+	ctx := r.Context()
+	ctx = context.WithValue(ctx, converterKey, converter)
+	r = r.WithContext(ctx)
+
+	return r
+}
+
+func GetConverter(r *http.Request) (*converter.Converter, error) {
+	c, ok := r.Context().Value(converterKey).(*converter.Converter)
+	if !ok {
+		return nil, &converterNotFoundError{}
+	}
+
+	return c, nil
+}
--- a/app/handlers/context/errors.go
+++ b/app/handlers/context/errors.go
@@ -0,0 +1,13 @@
+package context
+
+type contentTypeNotFoundError struct{}
+
+func (e *contentTypeNotFoundError) Error() string {
+	return "The 'Content-Type' was not found in request context"
+}
+
+type converterNotFoundError struct{}
+
+func (e *converterNotFoundError) Error() string {
+	return "The converter was not found in request context"
+}
--- a/app/handlers/converter/converter.go
+++ b/app/handlers/converter/converter.go
@@ -0,0 +1,118 @@
+package converter
+
+import (
+	"net/http"
+	"os"
+
+	gfile "github.com/gulien/gotenberg/app/handlers/converter/file"
+	"github.com/gulien/gotenberg/app/handlers/converter/process"
+	ghttp "github.com/gulien/gotenberg/app/handlers/http"
+)
+
+type Converter struct {
+	files            []*gfile.File
+	resultFilesPaths []string
+	FinalFilePath    string
+}
+
+func NewConverter(r *http.Request, contentType ghttp.ContentType) (*Converter, error) {
+	c := &Converter{}
+
+	switch contentType {
+	case ghttp.MultipartFormDataContentType:
+		err := r.ParseMultipartForm(32 << 20)
+		if err != nil {
+			return nil, err
+		}
+
+		formData := r.MultipartForm
+		files, ok := formData.File["files"]
+		if !ok {
+			// TODO
+			return nil, nil
+		}
+
+		for i := range files {
+			file, err := files[i].Open()
+			if err != nil {
+				return nil, err
+			}
+
+			defer file.Close()
+
+			f, err := gfile.NewFile(file)
+			if err != nil {
+				// todo err
+				return nil, err
+			}
+
+			c.files = append(c.files, f)
+		}
+		break
+	default:
+		f, err := gfile.NewFile(r.Body)
+		if err != nil {
+			// todo err
+			return nil, err
+		}
+
+		c.files = append(c.files, f)
+	}
+
+	return c, nil
+}
+
+func (c *Converter) Convert() error {
+	if len(c.files) == 0 {
+		return &noFileToConvertError{}
+	}
+
+	var filesPaths []string
+	for _, f := range c.files {
+		if f.Type != gfile.PDFType {
+			path, err := process.ExecConversion(f)
+			if err != nil {
+				return err
+			}
+
+			filesPaths = append(filesPaths, path)
+		} else {
+			filesPaths = append(filesPaths, f.Path)
+		}
+	}
+
+	path, err := process.ExecMerge(filesPaths)
+	if err != nil {
+		return err
+	}
+
+	c.resultFilesPaths = filesPaths
+	c.FinalFilePath = path
+
+	return nil
+}
+
+func (c *Converter) Clear() error {
+	for _, f := range c.files {
+		err := os.Remove(f.Path)
+		if err != nil {
+			return err
+		}
+	}
+
+	for _, path := range c.resultFilesPaths {
+		err := os.Remove(path)
+		if err != nil {
+			return err
+		}
+	}
+
+	if c.FinalFilePath != "" {
+		err := os.Remove(c.FinalFilePath)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
--- a/app/handlers/converter/errors.go
+++ b/app/handlers/converter/errors.go
@@ -0,0 +1,7 @@
+package converter
+
+type noFileToConvertError struct{}
+
+func (e *noFileToConvertError) Error() string {
+	return "There is no file to convert"
+}
--- a/app/handlers/converter/file/file.go
+++ b/app/handlers/converter/file/file.go
@@ -0,0 +1,120 @@
+package file
+
+import (
+	"fmt"
+	"io"
+	"os"
+
+	ghttp "github.com/gulien/gotenberg/app/handlers/http"
+
+	"github.com/satori/go.uuid"
+)
+
+type (
+	File struct {
+		Type FileType
+		Path string
+	}
+
+	FileType string
+
+	FileExt string
+)
+
+const (
+	PDFType    FileType = "PDF"
+	HTMLType   FileType = "HTML"
+	OfficeType FileType = "Office"
+
+	PDFExt    FileExt = ".pdf"
+	HTMLExt   FileExt = ".html"
+	OfficeExt FileExt = ""
+)
+
+func NewFile(r io.Reader) (*File, error) {
+	f := &File{
+		Path: MakeFilePath(),
+	}
+
+	file, err := os.Create(f.Path)
+	if err != nil {
+		return nil, err
+	}
+
+	defer file.Close()
+
+	_, err = io.Copy(file, r)
+	if err != nil {
+		return nil, err
+	}
+
+	// resets the read pointer.
+	file.Seek(0, 0)
+
+	t, err := findFileType(file)
+	if err != nil {
+		return nil, err
+	}
+
+	f.Type = t
+
+	f, err = reworkFilePath(f)
+	if err != nil {
+		return nil, err
+	}
+
+	return f, nil
+}
+
+func MakeFilePath() string {
+	return fmt.Sprintf("./%s", uuid.NewV4().String())
+}
+
+var filesTypes = map[ghttp.ContentType]FileType{
+	ghttp.PDFContentType:         PDFType,
+	ghttp.HTMLContentType:        HTMLType,
+	ghttp.OctetStreamContentType: OfficeType,
+	ghttp.ZipContentType:         OfficeType,
+}
+
+func findFileType(f *os.File) (FileType, error) {
+	ct, err := ghttp.SniffContentType(f)
+	if err != nil {
+		return "", err
+	}
+
+	t, ok := filesTypes[ct]
+	if !ok {
+		// TODO error
+		return "", nil
+	}
+
+	return t, nil
+}
+
+var filesExtensions = map[FileType]FileExt{
+	PDFType:    PDFExt,
+	HTMLType:   HTMLExt,
+	OfficeType: OfficeExt,
+}
+
+func reworkFilePath(f *File) (*File, error) {
+	ext, ok := filesExtensions[f.Type]
+	if !ok {
+		// TODO error
+		return nil, nil
+	}
+
+	if ext != OfficeExt {
+		newPath := fmt.Sprintf("./%s%s", MakeFilePath(), ext)
+
+		err := os.Rename(f.Path, newPath)
+		if err != nil {
+			return nil, err
+		}
+
+		f.Path = newPath
+	}
+
+	return f, nil
+}
--- a/app/handlers/converter/process/errors.go
+++ b/app/handlers/converter/process/errors.go
@@ -0,0 +1,13 @@
+package process
+
+type impossibleConversionError struct{}
+
+func (e *impossibleConversionError) Error() string {
+	return "Impossible conversion"
+}
+
+type commandTimeoutError struct{}
+
+func (e *commandTimeoutError) Error() string {
+	return "The command has reached timeout"
+}
--- a/app/handlers/converter/process/process.go
+++ b/app/handlers/converter/process/process.go
@@ -0,0 +1,119 @@
+package process
+
+import (
+	"bytes"
+	"fmt"
+	"os/exec"
+	"text/template"
+	"time"
+
+	"github.com/gulien/gotenberg/app/config"
+	gfile "github.com/gulien/gotenberg/app/handlers/converter/file"
+)
+
+var commandsConfig *config.CommandsConfig
+
+func Load(config *config.CommandsConfig) {
+	commandsConfig = config
+}
+
+func Reset() {
+	commandsConfig = nil
+}
+
+type conversionData struct {
+	FilePath       string
+	ResultFilePath string
+}
+
+func ExecConversion(file *gfile.File) (string, error) {
+	cmdData := &conversionData{
+		FilePath:       file.Path,
+		ResultFilePath: fmt.Sprintf("%s%s", gfile.MakeFilePath(), gfile.PDFExt),
+	}
+
+	var (
+		cmdTemplate *template.Template
+		cmdTimeout  int
+	)
+
+	switch file.Type {
+	case gfile.HTMLType:
+		cmdTemplate = commandsConfig.HTML.Template
+		cmdTimeout = commandsConfig.HTML.Timeout
+		break
+	case gfile.OfficeType:
+		cmdTemplate = commandsConfig.Office.Template
+		cmdTimeout = commandsConfig.Office.Timeout
+		break
+	default:
+		return "", &impossibleConversionError{}
+	}
+
+	var data bytes.Buffer
+	if err := cmdTemplate.Execute(&data, cmdData); err != nil {
+		return "", err
+	}
+
+	err := execCommand(data.String(), cmdTimeout)
+	if err != nil {
+		return "", err
+	}
+
+	return cmdData.ResultFilePath, nil
+}
+
+type mergeData struct {
+	FilesPaths     []string
+	ResultFilePath string
+}
+
+func ExecMerge(filesPaths []string) (string, error) {
+	cmdData := &mergeData{
+		FilesPaths:     filesPaths,
+		ResultFilePath: fmt.Sprintf("%s%s", gfile.MakeFilePath(), gfile.PDFExt),
+	}
+
+	cmdTemplate := commandsConfig.Merge.Template
+	cmdTimeout := commandsConfig.Merge.Timeout
+
+	var data bytes.Buffer
+	if err := cmdTemplate.Execute(&data, cmdData); err != nil {
+		return "", err
+	}
+
+	err := execCommand(data.String(), cmdTimeout)
+	if err != nil {
+		return "", err
+	}
+
+	return cmdData.ResultFilePath, nil
+}
+
+func execCommand(command string, timeout int) error {
+	// Wait for the process to finish or kill it after a timeout.
+	cmd := exec.Command("/bin/sh", "-c", command)
+	if err := cmd.Start(); err != nil {
+		return err
+	}
+
+	done := make(chan error, 1)
+	go func() {
+		done <- cmd.Wait()
+	}()
+
+	select {
+	case <-time.After(time.Duration(timeout) * time.Second):
+		if err := cmd.Process.Kill(); err != nil {
+			return err
+		}
+
+		return &commandTimeoutError{}
+	case err := <-done:
+		if err != nil {
+			return err
+		}
+
+		return nil
+	}
+}
--- a/app/handlers/errors.go
+++ b/app/handlers/errors.go
@@ -0,0 +1,7 @@
+package handlers
+
+type requestHasNoContentError struct{}
+
+func (e *requestHasNoContentError) Error() string {
+	return "Request has not content"
+}
--- a/app/handlers/handlers.go
+++ b/app/handlers/handlers.go
@@ -0,0 +1,126 @@
+package handlers
+
+import (
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+
+	"github.com/gulien/gotenberg/app/handlers/context"
+	"github.com/gulien/gotenberg/app/handlers/converter"
+	ghttp "github.com/gulien/gotenberg/app/handlers/http"
+	"github.com/gulien/gotenberg/app/logger"
+
+	"github.com/justinas/alice"
+)
+
+func GetHandlersChain() http.Handler {
+	return alice.New(enforceContentLengthHandler, enforceContentTypeHandler, convertHandler, serveHandler).ThenFunc(clearHandler)
+}
+
+// enforeContentLengthHandler checks if the request has content.
+func enforceContentLengthHandler(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if r.ContentLength == 0 {
+			e := &requestHasNoContentError{}
+			http.Error(w, e.Error(), http.StatusBadRequest)
+			logger.Error(e)
+			return
+		}
+
+		next.ServeHTTP(w, r)
+	})
+}
+
+// enforceContentTypeHandler checks if the "Content-Type" entry
+// from the request's header matches one of the allowed content types.
+func enforceContentTypeHandler(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		ct, err := ghttp.FindAuthorizedContentType(r.Header)
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusUnsupportedMediaType)
+			logger.Error(err)
+			return
+		}
+
+		r = context.WithContentType(r, ct)
+
+		next.ServeHTTP(w, r)
+	})
+}
+
+// convertHandler is in charge of converting the file(s) from the request to PDF.
+func convertHandler(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		ct, err := context.GetContentType(r)
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+			return
+		}
+
+		c, err := converter.NewConverter(r, ct)
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+			return
+		}
+
+		err = c.Convert()
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+			return
+		}
+
+		r = context.WithConverter(r, c)
+
+		next.ServeHTTP(w, r)
+	})
+}
+
+// serveHandler simply serves the created PDF.
+func serveHandler(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		c, err := context.GetConverter(r)
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+		}
+
+		reader, err := os.Open(c.FinalFilePath)
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+			return
+		}
+
+		defer reader.Close()
+
+		resultFileInfo, err := reader.Stat()
+		if err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+			logger.Error(err)
+			return
+		}
+
+		w.Header().Set("Content-Disposition", fmt.Sprintf("attachment; filename=\"%s\"", resultFileInfo.Name()))
+		w.Header().Set("Content-Type", "application/pdf")
+		w.Header().Set("Content-Length", fmt.Sprintf("%d", resultFileInfo.Size()))
+		io.Copy(w, reader)
+
+		next.ServeHTTP(w, r)
+	})
+}
+
+// clearHandler removes all files created during the conversion.
+func clearHandler(w http.ResponseWriter, r *http.Request) {
+	c, err := context.GetConverter(r)
+	if err != nil {
+		logger.Warn(err.Error())
+	}
+
+	if err := c.Clear(); err != nil {
+		logger.Warn(err.Error())
+	}
+}
--- a/app/handlers/http/errors.go
+++ b/app/handlers/http/errors.go
@@ -0,0 +1,15 @@
+package http
+
+import "fmt"
+
+type notAuthorizedContentTypeError struct{}
+
+func (e *notAuthorizedContentTypeError) Error() string {
+	return fmt.Sprintf("Accepted values for 'Content-Type': %s, %s, %s, %s", PDFContentType, HTMLContentType, OctetStreamContentType, MultipartFormDataContentType)
+}
+
+type notAuthorizedFileContentTypeError struct{}
+
+func (e *notAuthorizedFileContentTypeError) Error() string {
+	return fmt.Sprintf("Unable to detect a file 'Content-Type'")
+}
--- a/app/handlers/http/http.go
+++ b/app/handlers/http/http.go
@@ -0,0 +1,60 @@
+package http
+
+import (
+	"net/http"
+	"os"
+	"strings"
+)
+
+type ContentType string
+
+const (
+	PDFContentType               ContentType = "application/pdf"
+	HTMLContentType              ContentType = "text/html"
+	OctetStreamContentType       ContentType = "application/octet-stream"
+	ZipContentType               ContentType = "application/zip"
+	MultipartFormDataContentType ContentType = "multipart/form-data"
+)
+
+func FindAuthorizedContentType(h http.Header) (ContentType, error) {
+	ct := findContentType(h.Get("Content-Type"), HTMLContentType, OctetStreamContentType, MultipartFormDataContentType)
+	if ct == "" {
+		return "", &notAuthorizedContentTypeError{}
+	}
+
+	return ct, nil
+}
+
+func SniffContentType(f *os.File) (ContentType, error) {
+	// only the first 512 bytes are used to sniff the content type.
+	buffer := make([]byte, 512)
+	n, err := f.Read(buffer)
+	if err != nil {
+		return "", err
+	}
+
+	// resets the read pointer.
+	f.Seek(0, 0)
+
+	// using n if size of buffer < 512 bytes.
+	ct := findContentType(http.DetectContentType(buffer[:n]), PDFContentType, HTMLContentType, OctetStreamContentType, ZipContentType)
+	if ct == "" {
+		return "", &notAuthorizedFileContentTypeError{}
+	}
+
+	return ct, nil
+}
+
+func findContentType(contentType string, contentTypes ...ContentType) ContentType {
+	for _, ct := range contentTypes {
+		if i := strings.IndexRune(contentType, ';'); i != -1 {
+			contentType = contentType[0:i]
+		}
+
+		if contentType == string(ct) {
+			return ct
+		}
+	}
+
+	return ""
+}