refactor outline api to become a response modifier

cmd/main.go

@@ -183,12 +183,12 @@ func main() {
 		RulesetPath: *ruleset,
 	}
 
-	app.Get("api/outline/*", handlers.NewAPIOutlineHandler("api/outline/*", proxyOpts))
+	app.Get("api/content/*", handlers.NewAPIOutlineHandler("api/outline/*", proxyOpts))
-	app.Get("outline/*", handlers.Outline("outline/*", proxyOpts))
 
-	app.Get("/*", handlers.NewProxySiteHandler(proxyOpts))
+	app.Get("outline/*", handlers.NewOutlineHandler("outline/*", proxyOpts))
-	app.Post("/*", handlers.NewProxySiteHandler(proxyOpts))
 
-	fmt.Println(cli.StartupMessage("1.0.1", *port, *ruleset))
+	app.All("/*", handlers.NewProxySiteHandler(proxyOpts))
+
+	fmt.Println(cli.StartupMessage("1.0.0", *port, *ruleset))
 	log.Fatal(app.Listen(":" + *port))
 }

handlers/api_outline.go

@@ -34,7 +34,7 @@ func NewAPIOutlineHandler(path string, opts *ProxyOptions) fiber.Handler {
 			AddResponseModifications(
 				tx.DeleteIncomingCookies(),
 				tx.RewriteHTMLResourceURLs(),
-				tx.APIOutline(),
+				tx.APIContent(),
 			).
 			SetFiberCtx(c).
 			Execute()

handlers/outline.go

@@ -4,27 +4,14 @@ import (
 	"ladder/proxychain"
 	rx "ladder/proxychain/requestmodifers"
 	tx "ladder/proxychain/responsemodifers"
-	"log"
 
 	"github.com/gofiber/fiber/v2"
 )
 
-func Outline(path string, opts *ProxyOptions) fiber.Handler {
+func NewOutlineHandler(path string, opts *ProxyOptions) fiber.Handler {
-
-	// TODO: implement ruleset logic
-	/*
-		var rs ruleset.RuleSet
-		if opts.RulesetPath != "" {
-			r, err := ruleset.NewRuleset(opts.RulesetPath)
-			if err != nil {
-				panic(err)
-			}
-			rs = r
-		}
-	*/
-
 	return func(c *fiber.Ctx) error {
-		result, err := proxychain.
+
+		return proxychain.
 			NewProxyChain().
 			WithAPIPath(path).
 			SetDebugLogging(opts.Verbose).
@@ -36,20 +23,10 @@ func Outline(path string, opts *ProxyOptions) fiber.Handler {
 			AddResponseModifications(
 				tx.DeleteIncomingCookies(),
 				tx.RewriteHTMLResourceURLs(),
-				tx.APIOutline(),
+				tx.GenerateReadableOutline(), // <-- this response modification does the outline rendering
 			).
 			SetFiberCtx(c).
-			ExecuteForOutline()
+			Execute()
 
-		if err != nil {
-			log.Fatal(err)
-		}
-
-		return c.Render("outline", fiber.Map{
-			"Success": true,
-			"Params":  c.Params("*"),
-			"Title":   "Outline",
-			"Body":    result,
-		})
 	}
 }

proxychain/responsemodifers/api_content.go

@@ -1,21 +1,17 @@
 package responsemodifers
 
 import (
-	"io"
+	"bytes"
-	"strings"
+	"encoding/json"
-
-	//"github.com/go-shiori/dom"
-	"github.com/go-shiori/dom"
 	"github.com/markusmobius/go-trafilatura"
-
+	"io"
-	//"golang.org/x/net/html"
-
 	"ladder/proxychain"
 	"ladder/proxychain/responsemodifers/api"
 )
 
-// APIOutline creates an JSON representation of the article and returns it as an API response.
+// APIContent creates an JSON representation of the article and returns it as an API response.
-func APIOutline() proxychain.ResponseModification {
+func APIContent() proxychain.ResponseModification {
+
 	return func(chain *proxychain.ProxyChain) error {
 		// we set content-type twice here, in case another response modifier
 		// tries to forward over the original headers
@@ -38,9 +34,14 @@ func APIOutline() proxychain.ResponseModification {
 			return nil
 		}
 
-		doc := trafilatura.CreateReadableDocument(result)
+		res := api.ExtractResultToAPIResponse(result)
-		reader := io.NopCloser(strings.NewReader(dom.OuterHTML(doc)))
+		jsonData, err := json.MarshalIndent(res, "", "  ")
-		chain.Response.Body = reader
+		if err != nil {
+			return err
+		}
+
+		chain.Response.Body = io.NopCloser(bytes.NewReader(jsonData))
 		return nil
 	}
+
 }

proxychain/responsemodifers/generate_readable_outline.go

@@ -0,0 +1,84 @@
+package responsemodifers
+
+import (
+	"embed"
+	"fmt"
+	"html/template"
+	"io"
+	"ladder/proxychain"
+	"log"
+
+	"github.com/go-shiori/dom"
+	"github.com/markusmobius/go-trafilatura"
+)
+
+//go:embed generate_readable_outline.html
+var templateFS embed.FS
+
+// GenerateReadableOutline creates an reader-friendly distilled representation of the article.
+// This is a reliable way of bypassing soft-paywalled articles, where the content is hidden, but still present in the DOM.
+func GenerateReadableOutline() proxychain.ResponseModification {
+
+	// get template only once, and resuse for subsequent calls
+	f := "generate_readable_outline.html"
+	tmpl, err := template.ParseFS(templateFS, f)
+	if err != nil {
+		panic(fmt.Errorf("tx.GenerateReadableOutline Error: %s not found", f))
+	}
+
+	return func(chain *proxychain.ProxyChain) error {
+
+		// ===========================================================
+		// 1. extract dom contents using reading mode algo
+		// ===========================================================
+		opts := trafilatura.Options{
+			IncludeImages: true,
+			IncludeLinks:  true,
+			//FavorPrecision:     true,
+			FallbackCandidates: nil, // TODO: https://github.com/markusmobius/go-trafilatura/blob/main/examples/chained/main.go
+			// implement fallbacks from	"github.com/markusmobius/go-domdistiller" and 	"github.com/go-shiori/go-readability"
+			OriginalURL: chain.Request.URL,
+		}
+
+		result, err := trafilatura.Extract(chain.Response.Body, opts)
+		if err != nil {
+			return err
+		}
+
+		doc := trafilatura.CreateReadableDocument(result)
+		distilledHTML := dom.OuterHTML(doc)
+
+		// ============================================================================
+		// 2. render generate_readable_outline.html template using metadata from step 1
+		// ============================================================================
+		data := map[string]interface{}{
+			"Success": true,
+			"Params":  chain.Request.URL,
+			//"Title":   result.Metadata.Title, // todo: modify CreateReadableDocument so we don't have <h1> titles duplicated?
+			"Date":   result.Metadata.Date.String(),
+			"Author": result.Metadata.Author,
+			"Body":   distilledHTML,
+		}
+
+		// ============================================================================
+		// 3. queue sending the response back to the client by replacing the response body
+		// (the response body will be read as a stream in proxychain.Execute() later on.)
+		// ============================================================================
+		pr, pw := io.Pipe() // pipe io.writer contents into io.reader
+
+		// Use a goroutine for writing to the pipe so we don't deadlock the request
+		go func() {
+			defer pw.Close()
+
+			err := tmpl.Execute(pw, data) // <- render template
+
+			if err != nil {
+				log.Printf("WARN: GenerateReadableOutline template rendering error: %s\n", err)
+			}
+		}()
+
+		chain.Context.Set("content-type", "text/html")
+		chain.Response.Body = pr // <- replace reponse body reader with our new reader from pipe
+		return nil
+	}
+}

proxychain/responsemodifers/generate_readable_outline.html

@@ -356,7 +356,7 @@
 
         <div class="flex flex-col space-y-3">
         <div>{{ .Params }}</div>
-        <div>{{ unescape .Body }}</div>
+        <div>{{ .Body }}</div>
         {{ end }}
         <!-- Trick Tailwind into compiling these styles into styles.css -->
         <!-- <div class="hidden text-xs text-sm text-base text-xl text-2xl text-3xl text-4xl sm:text-3xl sm:text-4xl sm:text-5xl"></div> -->