From 8551b7e98ecaa4ea543755ef6d585f9e8f79dde6 Mon Sep 17 00:00:00 2001
From: Matthieu Gautier <mgautier@kymeria.fr>
Date: Wed, 8 Dec 2021 16:15:39 +0100
Subject: [PATCH] Add `raw` endpoint.

As the name suggests it, this endpoint is not smart :
It returns the content as it is and only if it is present
(no compatibility or whatever).

The only "smart" thing is to return a redirect if the entry is a redirect.
---
 src/server/internalServer.cpp | 57 +++++++++++++++++++++++++++++++++++
 src/server/internalServer.h   |  1 +
 test/server.cpp               | 29 ++++++++++++++++++
 3 files changed, 87 insertions(+)

diff --git a/src/server/internalServer.cpp b/src/server/internalServer.cpp
index 6ff949380..06cfa3c58 100644
--- a/src/server/internalServer.cpp
+++ b/src/server/internalServer.cpp
@@ -283,6 +283,9 @@ std::unique_ptr<Response> InternalServer::handle_request(const RequestContext& r
     if (startsWith(request.get_url(), "/catalog/"))
       return handle_catalog(request);
 
+    if (startsWith(request.get_url(), "/raw/"))
+      return handle_raw(request);
+
     if (request.get_url() == "/meta")
       return handle_meta(request);
 
@@ -891,4 +894,58 @@ std::unique_ptr<Response> InternalServer::handle_content(const RequestContext& r
   }
 }
 
+
+std::unique_ptr<Response> InternalServer::handle_raw(const RequestContext& request)
+{
+  if (m_verbose.load()) {
+    printf("** running handle_raw\n");
+  }
+
+  std::string bookName;
+  std::string kind;
+  try {
+     bookName = request.get_url_part(1);
+     kind = request.get_url_part(2);
+  } catch (const std::out_of_range& e) {
+     return Response::build_404(*this, request, bookName, "", "");
+  }
+
+  if (kind != "meta" && kind!= "content") {
+    return Response::build_404(*this, request, bookName, "", "");
+  }
+
+  std::shared_ptr<zim::Archive> archive;
+  try {
+    const std::string bookId = mp_nameMapper->getIdForName(bookName);
+    archive = mp_library->getArchiveById(bookId);
+  } catch (const std::out_of_range& e) {}
+
+  if (archive == nullptr) {
+    return Response::build_404(*this, request, bookName, "", "");
+  }
+
+  auto itemPath = request.get_url().substr(bookName.size()+kind.size()+6);
+  if (itemPath[0] == '/') {
+    itemPath = itemPath.substr(1);
+  }
+
+  try {
+    if (kind == "meta") {
+      auto item = archive->getMetadataItem(itemPath);
+      return ItemResponse::build(*this, request, item, true);
+    } else {
+      auto entry = archive->getEntryByPath(itemPath);
+      if (entry.isRedirect()) {
+        return build_redirect(bookName, entry.getItem(true));
+      }
+      return ItemResponse::build(*this, request, entry.getItem(), true);
+    }
+  } catch (zim::EntryNotFound& e ) {
+    if (m_verbose.load()) {
+      printf("Failed to find %s\n", itemPath.c_str());
+    }
+    return Response::build_404(*this, request, bookName, getArchiveTitle(*archive), "");
+  }
+}
+
 }
diff --git a/src/server/internalServer.h b/src/server/internalServer.h
index 7b935a126..658fe5eae 100644
--- a/src/server/internalServer.h
+++ b/src/server/internalServer.h
@@ -86,6 +86,7 @@ class InternalServer {
     std::unique_ptr<Response> handle_random(const RequestContext& request);
     std::unique_ptr<Response> handle_captured_external(const RequestContext& request);
     std::unique_ptr<Response> handle_content(const RequestContext& request);
+    std::unique_ptr<Response> handle_raw(const RequestContext& request);
 
     std::vector<std::string> search_catalog(const RequestContext& request,
                                             kiwix::OPDSDumper& opdsDumper);
diff --git a/test/server.cpp b/test/server.cpp
index afffee224..d71f1cb99 100644
--- a/test/server.cpp
+++ b/test/server.cpp
@@ -185,6 +185,9 @@ const ResourceCollection resources200Compressible{
 
   { WITH_ETAG, "/zimfile/A/index" },
   { WITH_ETAG, "/zimfile/A/Ray_Charles" },
+
+  { WITH_ETAG, "/raw/zimfile/content/A/index" },
+  { WITH_ETAG, "/raw/zimfile/content/A/Ray_Charles" },
 };
 
 const ResourceCollection resources200Uncompressible{
@@ -207,6 +210,10 @@ const ResourceCollection resources200Uncompressible{
   { WITH_ETAG, "/corner_cases/A/empty.html" },
   { WITH_ETAG, "/corner_cases/-/empty.css" },
   { WITH_ETAG, "/corner_cases/-/empty.js" },
+
+  // The title and creator are too small to be compressed
+  { WITH_ETAG, "/raw/zimfile/meta/Creator" },
+  { WITH_ETAG, "/raw/zimfile/meta/Title" },
 };
 
 ResourceCollection all200Resources()
@@ -310,6 +317,28 @@ TEST_F(ServerTest, BookMainPageIsRedirectedToArticleIndex)
   ASSERT_EQ("/zimfile/A/index", g->get_header_value("Location"));
 }
 
+
+TEST_F(ServerTest, RawEntry)
+{
+  auto p = zfs1_->GET("/raw/zimfile/meta/Title");
+  EXPECT_EQ(200, p->status);
+  EXPECT_EQ(p->body, std::string("Ray Charles"));
+
+  p = zfs1_->GET("/raw/zimfile/meta/Creator");
+  EXPECT_EQ(200, p->status);
+  EXPECT_EQ(p->body, std::string("Wikipedia"));
+
+  // The raw content of Ray_Charles doesn't contain a taskbar...
+  p = zfs1_->GET("/raw/zimfile/content/A/Ray_Charles");
+  EXPECT_EQ(200, p->status);
+  EXPECT_TRUE(p->body.find("taskbar") == std::string::npos);
+
+  // ... but the "normal" content does.
+  p = zfs1_->GET("/zimfile/A/Ray_Charles");
+  EXPECT_EQ(200, p->status);
+  EXPECT_TRUE(p->body.find("taskbar") != std::string::npos);
+}
+
 TEST_F(ServerTest, HeadMethodIsSupported)
 {
   for ( const Resource& res : all200Resources() )