docd: return JSON errors when things fail (#127)

docd/convert.go

@@ -0,0 +1,122 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"log"
+	"net/http"
+	"os"
+
+	"cloud.google.com/go/errorreporting"
+
+	"code.sajari.com/docconv"
+	"code.sajari.com/docconv/docd/internal"
+)
+
+type convertServer struct {
+	er internal.ErrorReporter
+}
+
+func (s *convertServer) convert(w http.ResponseWriter, r *http.Request) {
+	ctx := r.Context()
+
+	// Readability flag. Currently only used for HTML
+	var readability bool
+	if r.FormValue("readability") == "1" {
+		readability = true
+		if *logLevel >= 2 {
+			log.Println("Readability is on")
+		}
+	}
+
+	path := r.FormValue("path")
+	if path != "" {
+		mimeType := docconv.MimeTypeByExtension(path)
+
+		f, err := os.Open(path)
+		if err != nil {
+			s.serverError(ctx, w, r, fmt.Errorf("could not open file: %w", err))
+			return
+		}
+		defer f.Close()
+
+		data, err := docconv.Convert(f, mimeType, readability)
+		if err != nil {
+			s.serverError(ctx, w, r, fmt.Errorf("could not convert file from path %v: %w", path, err))
+			return
+		}
+
+		s.respond(ctx, w, r, http.StatusOK, data)
+		return
+	}
+
+	// Get uploaded file
+	file, info, err := r.FormFile("input")
+	if err != nil {
+		s.serverError(ctx, w, r, fmt.Errorf("could not get input file: %w", err))
+		return
+	}
+	defer file.Close()
+
+	// Abort if file doesn't have a mime type
+	if len(info.Header["Content-Type"]) == 0 {
+		s.clientError(ctx, w, r, http.StatusUnprocessableEntity, "input file %v does not have a Content-Type header", info.Filename)
+		return
+	}
+
+	// If a generic mime type was provided then use file extension to determine mimetype
+	mimeType := info.Header["Content-Type"][0]
+	if mimeType == "application/octet-stream" {
+		mimeType = docconv.MimeTypeByExtension(info.Filename)
+	}
+
+	if *logLevel >= 1 {
+		log.Printf("Received file: %v (%v)", info.Filename, mimeType)
+	}
+
+	data, err := docconv.Convert(file, mimeType, readability)
+	if err != nil {
+		s.serverError(ctx, w, r, fmt.Errorf("could not convert file: %w", err))
+		return
+	}
+
+	s.respond(ctx, w, r, http.StatusOK, data)
+}
+
+func (s *convertServer) clientError(ctx context.Context, w http.ResponseWriter, r *http.Request, code int, pattern string, args ...interface{}) {
+	s.respond(ctx, w, r, code, &docconv.Response{
+		Error: fmt.Sprintf(pattern, args...),
+	})
+
+	log.Printf(pattern, args...)
+}
+
+func (s *convertServer) serverError(ctx context.Context, w http.ResponseWriter, r *http.Request, err error) {
+	w.WriteHeader(http.StatusInternalServerError)
+	w.Write([]byte(`{"error":"internal server error"}`))
+
+	e := errorreporting.Entry{
+		Error: err,
+		Req:   r,
+	}
+	s.er.Report(e)
+
+	log.Printf("%v", err)
+}
+
+func (s *convertServer) respond(ctx context.Context, w http.ResponseWriter, r *http.Request, code int, resp interface{}) {
+	buf := &bytes.Buffer{}
+	err := json.NewEncoder(buf).Encode(resp)
+	if err != nil {
+		s.serverError(ctx, w, r, fmt.Errorf("could not marshal JSON response: %w", err))
+		return
+	}
+	w.WriteHeader(code)
+	n, err := io.Copy(w, buf)
+	if err != nil {
+		panic(fmt.Errorf("could not write to response (failed after %d bytes): %w", n, err))
+	}
+}

docd/main.go

@@ -2,7 +2,6 @@ package main
 
 import (
 	"context"
-	"encoding/json"
 	"flag"
 	"fmt"
 	"log"
@@ -53,6 +52,10 @@ func main() {
 		}
 	}
 
+	cs := &convertServer{
+		er: er,
+	}
+
 	// TODO: Improve this (remove the need for it!)
 	docconv.HTMLReadabilityOptionsValues = docconv.HTMLReadabilityOptions{
 		LengthLow:             *readabilityLengthLow,
@@ -73,73 +76,13 @@ func main() {
 		return
 	}
 
-	serve(er)
-}
-
-func convert(w http.ResponseWriter, r *http.Request) {
-	// Readability flag. Currently only used for HTML
-	var readability bool
-	if r.FormValue("readability") == "1" {
-		readability = true
-		if *logLevel >= 2 {
-			log.Println("Readability is on")
-		}
-	}
-
-	path := r.FormValue("path")
-	if path != "" {
-		b, err := docconv.ConvertPathReadability(path, readability)
-		if err != nil {
-			// TODO: return a sensible status code for errors like this.
-			log.Printf("error converting path '%v': %v", path, err)
-			return
-		}
-		w.Write(b)
-		return
-	}
-
-	// Get uploaded file
-	file, info, err := r.FormFile("input")
-	if err != nil {
-		log.Println("File upload", err)
-		return
-	}
-	defer file.Close()
-
-	// Abort if file doesn't have a mime type
-	if len(info.Header["Content-Type"]) == 0 {
-		log.Println("No content type", info.Filename)
-		return
-	}
-
-	// If a generic mime type was provided then use file extension to determine mimetype
-	mimeType := info.Header["Content-Type"][0]
-	if mimeType == "application/octet-stream" {
-		mimeType = docconv.MimeTypeByExtension(info.Filename)
-	}
-
-	if *logLevel >= 1 {
-		log.Println("Received file: " + info.Filename + " (" + mimeType + ")")
-	}
-
-	data, err := docconv.Convert(file, mimeType, readability)
-	if err != nil {
-		log.Printf("error converting data: %v", err)
-		data = &docconv.Response{
-			Error: err.Error(),
-		}
-	}
-
-	if err := json.NewEncoder(w).Encode(data); err != nil {
-		log.Printf("error marshaling JSON data: %v", err)
-		return
-	}
+	serve(er, cs)
 }
 
 // Start the conversion web service
-func serve(er internal.ErrorReporter) {
+func serve(er internal.ErrorReporter, cs *convertServer) {
 	r := mux.NewRouter()
-	r.HandleFunc("/convert", convert)
+	r.HandleFunc("/convert", cs.convert)
 
 	// Start webserver
 	log.Println("Setting log level to", *logLevel)