fix: fixed project loading from filesystem and added project metadata to metadata diff

lib/bds/maintenance.ex

@@ -4,6 +4,7 @@ defmodule BDS.Maintenance do
   import Ecto.Query
 
   alias BDS.Frontmatter
+  alias BDS.Metadata
   alias BDS.Media.Media
   alias BDS.Media.Translation, as: MediaTranslation
   alias BDS.Embeddings
@@ -30,7 +31,8 @@ defmodule BDS.Maintenance do
     project = Projects.get_project!(project_id)
 
     diff_reports =
-      post_diff_reports(project_id, project) ++
+      project_metadata_diff_reports(project_id) ++
+        post_diff_reports(project_id, project) ++
         post_translation_diff_reports(project_id, project) ++
         media_diff_reports(project_id, project) ++
         media_translation_diff_reports(project_id, project) ++
@@ -43,6 +45,71 @@ defmodule BDS.Maintenance do
     {:ok, %{diff_reports: diff_reports, orphan_reports: orphan_reports}}
   end
 
+  defp project_metadata_diff_reports(project_id) do
+    {:ok, db_state} = Metadata.get_project_metadata(project_id)
+    {:ok, filesystem_state} = Metadata.read_project_metadata_from_filesystem(project_id)
+
+    [
+      build_diff_report("project", project_id, [
+        diff_field("name", db_state.name, filesystem_state.name),
+        diff_field("description", db_state.description, filesystem_state.description),
+        diff_field("public_url", db_state.public_url, filesystem_state.public_url),
+        diff_field("main_language", db_state.main_language, filesystem_state.main_language),
+        diff_field("default_author", db_state.default_author, filesystem_state.default_author),
+        diff_field(
+          "max_posts_per_page",
+          db_state.max_posts_per_page,
+          filesystem_state.max_posts_per_page
+        ),
+        diff_field(
+          "blogmark_category",
+          db_state.blogmark_category,
+          filesystem_state.blogmark_category
+        ),
+        diff_field("pico_theme", db_state.pico_theme, filesystem_state.pico_theme),
+        diff_field(
+          "semantic_similarity_enabled",
+          db_state.semantic_similarity_enabled,
+          filesystem_state.semantic_similarity_enabled
+        ),
+        diff_field("blog_languages", db_state.blog_languages, filesystem_state.blog_languages)
+      ]),
+      build_diff_report("categories", project_id, [
+        diff_field("categories", db_state.categories, filesystem_state.categories)
+      ]),
+      build_diff_report("category_meta", project_id, [
+        diff_field(
+          "category_settings",
+          db_state.category_settings,
+          filesystem_state.category_settings
+        )
+      ]),
+      build_diff_report("publishing", project_id, [
+        diff_field(
+          "ssh_host",
+          Map.get(db_state.publishing_preferences, "ssh_host"),
+          Map.get(filesystem_state.publishing_preferences, "ssh_host")
+        ),
+        diff_field(
+          "ssh_user",
+          Map.get(db_state.publishing_preferences, "ssh_user"),
+          Map.get(filesystem_state.publishing_preferences, "ssh_user")
+        ),
+        diff_field(
+          "ssh_remote_path",
+          Map.get(db_state.publishing_preferences, "ssh_remote_path"),
+          Map.get(filesystem_state.publishing_preferences, "ssh_remote_path")
+        ),
+        diff_field(
+          "ssh_mode",
+          Map.get(db_state.publishing_preferences, "ssh_mode"),
+          Map.get(filesystem_state.publishing_preferences, "ssh_mode")
+        )
+      ])
+    ]
+    |> Enum.reject(&is_nil/1)
+  end
+
   defp normalize_entity_type(:post), do: :post
   defp normalize_entity_type(:media), do: :media
   defp normalize_entity_type(:script), do: :script
@@ -366,6 +433,16 @@ defmodule BDS.Maintenance do
     |> Enum.map(&%{file_path: &1})
   end
 
+  defp build_diff_report(entity_type, entity_id, differences) do
+    normalized = Enum.reject(differences, &is_nil/1)
+
+    if normalized == [] do
+      nil
+    else
+      %{entity_type: entity_type, entity_id: entity_id, differences: normalized}
+    end
+  end
+
   defp diff_field(name, db_value, file_value) do
     if equal_diff_values?(db_value, file_value) do
       nil
@@ -378,6 +455,10 @@ defmodule BDS.Maintenance do
     normalize_list_diff_values(left) == normalize_list_diff_values(right)
   end
 
+  defp equal_diff_values?(left, right) when is_map(left) and is_map(right) do
+    normalize_map_diff_values(left) == normalize_map_diff_values(right)
+  end
+
   defp equal_diff_values?(left, right), do: stringify_value(left) == stringify_value(right)
 
   defp normalize_list_diff_values(values) do
@@ -392,11 +473,26 @@ defmodule BDS.Maintenance do
   defp stringify_value(value) when is_integer(value), do: Integer.to_string(value)
   defp stringify_value(value) when is_binary(value), do: value
 
+  defp stringify_value(value) when is_map(value),
+    do: value |> normalize_map_diff_values() |> Jason.encode!()
+
   defp stringify_value(value) when is_list(value),
     do: Enum.map_join(value, ",", &stringify_value/1)
 
   defp stringify_value(value), do: to_string(value)
 
+  defp normalize_map_diff_values(values) when is_map(values) do
+    values
+    |> Enum.map(fn {key, value} -> {to_string(key), normalize_nested_diff_value(value)} end)
+    |> Enum.sort_by(&elem(&1, 0))
+    |> Map.new()
+  end
+
+  defp normalize_nested_diff_value(value) when is_map(value), do: normalize_map_diff_values(value)
+  defp normalize_nested_diff_value(value) when is_list(value), do: Enum.map(value, &normalize_nested_diff_value/1)
+  defp normalize_nested_diff_value(value) when is_atom(value), do: Atom.to_string(value)
+  defp normalize_nested_diff_value(value), do: value
+
   defp read_frontmatter_document(project, relative_path) do
     full_path = Path.join(Projects.project_data_dir(project), relative_path)
 

lib/bds/metadata.ex

@@ -41,6 +41,11 @@ defmodule BDS.Metadata do
     {:ok, load_state(project)}
   end
 
+  def read_project_metadata_from_filesystem(project_id) do
+    project = Projects.get_project!(project_id)
+    {:ok, load_state_from_filesystem(project)}
+  end
+
   def update_project_metadata(project_id, attrs) do
     project = Projects.get_project!(project_id)
     state = load_state(project)
@@ -131,35 +136,32 @@ defmodule BDS.Metadata do
   def sync_project_metadata_from_filesystem(project_id) do
     project = Projects.get_project!(project_id)
     now = Persistence.now_ms()
-
+    filesystem_state = load_state_from_filesystem(project)
-    project_metadata_from_files =
-      read_json(project, "project.json") ||
-        stringify_project_metadata(default_project_metadata(project))
-
-    categories_from_files =
-      read_json(project, "categories.json") || %{"categories" => @default_categories}
-
-    category_meta_from_files = read_json(project, "category-meta.json") || %{"categories" => %{}}
-    publishing_from_files = read_json(project, "publishing.json") || %{"ssh_mode" => "scp"}
 
     Repo.transaction(fn ->
       updated_project =
         project
         |> Project.changeset(%{
-          name: Map.get(project_metadata_from_files, "name", project.name),
+          name: filesystem_state.name,
-          description: Map.get(project_metadata_from_files, "description"),
+          description: filesystem_state.description,
           updated_at: now
         })
         |> Repo.update!()
 
-      persist_setting(project_id, "project", project_metadata_from_files, now)
+      persist_setting(project_id, "project", stringify_project_metadata(filesystem_state), now)
-      persist_setting(project_id, "categories", categories_from_files, now)
+      persist_setting(project_id, "categories", %{"categories" => filesystem_state.categories}, now)
-      persist_setting(project_id, "category_meta", category_meta_from_files, now)
+      persist_setting(
-      persist_setting(project_id, "publishing", publishing_from_files, now)
+        project_id,
-      write_project_json(updated_project, project_metadata_from_files)
+        "category_meta",
-      write_categories_json(updated_project, normalized_categories(categories_from_files))
+        %{"categories" => filesystem_state.category_settings},
-      write_category_meta_json(updated_project, normalized_category_settings(category_meta_from_files))
+        now
-      write_publishing_json(updated_project, publishing_from_files)
+      )
+
+      persist_setting(project_id, "publishing", filesystem_state.publishing_preferences, now)
+      write_project_json(updated_project, stringify_project_metadata(filesystem_state))
+      write_categories_json(updated_project, filesystem_state.categories)
+      write_category_meta_json(updated_project, filesystem_state.category_settings)
+      write_publishing_json(updated_project, filesystem_state.publishing_preferences)
       load_state(updated_project)
     end)
     |> unwrap_transaction()
@@ -210,6 +212,37 @@ defmodule BDS.Metadata do
     }
   end
 
+  defp load_state_from_filesystem(project) do
+    project_metadata =
+      read_json(project, "project.json") ||
+        stringify_project_metadata(default_project_metadata(project))
+
+    categories = normalized_categories(read_json(project, "categories.json") || %{"categories" => @default_categories})
+
+    category_settings =
+      normalized_category_settings(read_json(project, "category-meta.json") || %{"categories" => %{}})
+
+    publishing_preferences = read_json(project, "publishing.json") || %{"ssh_mode" => "scp"}
+
+    %{
+      name: Map.get(project_metadata, "name", project.name),
+      description: Map.get(project_metadata, "description"),
+      public_url: Map.get(project_metadata, "public_url"),
+      main_language: Map.get(project_metadata, "main_language"),
+      default_author: Map.get(project_metadata, "default_author"),
+      max_posts_per_page:
+        Map.get(project_metadata, "max_posts_per_page", @default_max_posts_per_page),
+      blogmark_category: Map.get(project_metadata, "blogmark_category"),
+      pico_theme: Map.get(project_metadata, "pico_theme"),
+      semantic_similarity_enabled:
+        Map.get(project_metadata, "semantic_similarity_enabled", false),
+      blog_languages: Map.get(project_metadata, "blog_languages", []),
+      categories: categories,
+      category_settings: category_settings,
+      publishing_preferences: publishing_preferences
+    }
+  end
+
   defp default_project_metadata(project) do
     %{
       name: project.name,

lib/bds/projects.ex

@@ -3,6 +3,7 @@ defmodule BDS.Projects do
 
   import Ecto.Query
 
+  alias BDS.Metadata
   alias BDS.Persistence
   alias BDS.Projects.Project
   alias BDS.Repo
@@ -96,7 +97,7 @@ defmodule BDS.Projects do
       project
     end)
     |> case do
-      {:ok, project} -> {:ok, project}
+      {:ok, project} -> sync_filesystem_metadata(project)
       {:error, reason} -> {:error, reason}
     end
   end
@@ -168,6 +169,15 @@ defmodule BDS.Projects do
     }
   end
 
+  defp sync_filesystem_metadata(%Project{data_path: nil} = project), do: {:ok, project}
+
+  defp sync_filesystem_metadata(%Project{} = project) do
+    case Metadata.sync_project_metadata_from_filesystem(project.id) do
+      {:ok, _metadata} -> {:ok, get_project!(project.id)}
+      {:error, reason} -> {:error, reason}
+    end
+  end
+
   defp unique_slug(base_slug) do
     normalized = if base_slug in [nil, ""], do: "project", else: base_slug
 

test/bds/maintenance_test.exs

@@ -749,6 +749,70 @@ defmodule BDS.MaintenanceTest do
            end)
   end
 
+  test "metadata_diff includes project-level metadata drift", %{project: project, temp_dir: temp_dir} do
+    assert {:ok, _metadata} =
+             BDS.Metadata.update_project_metadata(project.id, %{
+               name: "Database Blog",
+               description: "Database description",
+               public_url: "https://database.example",
+               main_language: "en",
+               default_author: "Database Author",
+               max_posts_per_page: 25,
+               blogmark_category: "article",
+               pico_theme: "blue",
+               semantic_similarity_enabled: false,
+               blog_languages: ["de"]
+             })
+
+    assert {:ok, _metadata} =
+             BDS.Metadata.set_publishing_preferences(project.id, %{
+               ssh_host: "db.example",
+               ssh_user: "db-user",
+               ssh_remote_path: "/srv/db",
+               ssh_mode: "scp"
+             })
+
+    meta_dir = Path.join(temp_dir, "meta")
+
+    File.write!(
+      Path.join(meta_dir, "project.json"),
+      Jason.encode!(%{
+        "name" => "Filesystem Blog",
+        "description" => "Filesystem description",
+        "publicUrl" => "https://filesystem.example",
+        "mainLanguage" => "fr",
+        "defaultAuthor" => "Filesystem Author",
+        "maxPostsPerPage" => 12,
+        "blogmarkCategory" => "notes",
+        "picoTheme" => "slate",
+        "semanticSimilarityEnabled" => true,
+        "blogLanguages" => ["it"]
+      })
+    )
+
+    File.write!(
+      Path.join(meta_dir, "publishing.json"),
+      Jason.encode!(%{
+        "sshHost" => "files.example",
+        "sshUser" => "files-user",
+        "sshRemotePath" => "/srv/files",
+        "sshMode" => "rsync"
+      })
+    )
+
+    assert {:ok, diff} = BDS.Maintenance.metadata_diff(project.id)
+
+    assert Enum.any?(diff.diff_reports, fn report ->
+             report.entity_type == "project" and
+               Enum.any?(report.differences, &(&1.name == "main_language" and &1.db_value == "en" and &1.file_value == "fr"))
+           end)
+
+    assert Enum.any?(diff.diff_reports, fn report ->
+             report.entity_type == "publishing" and
+               Enum.any?(report.differences, &(&1.name == "ssh_mode" and &1.db_value == "scp" and &1.file_value == "rsync"))
+           end)
+  end
+
   defp collect_progress_events(acc \\ []) do
     receive do
       {:rebuild_progress, value, message} -> collect_progress_events([{value, message} | acc])

test/bds/metadata_test.exs

@@ -221,6 +221,9 @@ defmodule BDS.MetadataTest do
   test "sync_project_metadata_from_filesystem materializes the canonical metadata files when missing",
        %{project: project, temp_dir: temp_dir} do
     meta_dir = Path.join(temp_dir, "meta")
+
+    File.rm_rf!(meta_dir)
+
     refute File.exists?(Path.join(meta_dir, "project.json"))
     refute File.exists?(Path.join(meta_dir, "categories.json"))
     refute File.exists?(Path.join(meta_dir, "category-meta.json"))

test/bds/projects_test.exs

@@ -3,6 +3,7 @@ defmodule BDS.ProjectsTest do
 
   import Ecto.Query
 
+  alias BDS.Metadata
   alias BDS.Projects.Project
   alias BDS.Repo
   alias BDS.Templates.Template
@@ -157,4 +158,60 @@ defmodule BDS.ProjectsTest do
     assert BDS.Projects.get_project(external_project.id) == nil
     assert File.read!(marker_path) == "preserve me"
   end
+
+  test "create_project loads project metadata from an existing filesystem-backed blog", %{temp_root: temp_root} do
+    external_dir = Path.join(temp_root, "imported-blog")
+    meta_dir = Path.join(external_dir, "meta")
+    File.mkdir_p!(meta_dir)
+
+    File.write!(
+      Path.join(meta_dir, "project.json"),
+      Jason.encode!(%{
+        "name" => "Imported Blog",
+        "description" => "Filesystem metadata",
+        "publicUrl" => "https://imported.example",
+        "mainLanguage" => "de",
+        "defaultAuthor" => "Importer",
+        "maxPostsPerPage" => 17,
+        "blogmarkCategory" => "notes",
+        "picoTheme" => "slate",
+        "semanticSimilarityEnabled" => true,
+        "blogLanguages" => ["en", "fr"]
+      })
+    )
+
+    File.write!(
+      Path.join(meta_dir, "publishing.json"),
+      Jason.encode!(%{
+        "sshHost" => "upload.example",
+        "sshUser" => "deploy",
+        "sshRemotePath" => "/srv/imported",
+        "sshMode" => "rsync"
+      })
+    )
+
+    assert {:ok, project} =
+             BDS.Projects.create_project(%{name: "Placeholder", data_path: external_dir})
+
+    assert BDS.Projects.get_project!(project.id).name == "Imported Blog"
+
+    assert {:ok, metadata} = Metadata.get_project_metadata(project.id)
+    assert metadata.name == "Imported Blog"
+    assert metadata.description == "Filesystem metadata"
+    assert metadata.public_url == "https://imported.example"
+    assert metadata.main_language == "de"
+    assert metadata.default_author == "Importer"
+    assert metadata.max_posts_per_page == 17
+    assert metadata.blogmark_category == "notes"
+    assert metadata.pico_theme == "slate"
+    assert metadata.semantic_similarity_enabled == true
+    assert metadata.blog_languages == ["en", "fr"]
+
+    assert metadata.publishing_preferences == %{
+             "ssh_host" => "upload.example",
+             "ssh_user" => "deploy",
+             "ssh_remote_path" => "/srv/imported",
+             "ssh_mode" => "rsync"
+           }
+  end
 end