docs: Add workflow for extrinsic metadata + mention storage on the path from loader to journal

docs/images/tasks-extrinsic-metadata-indexers.uml

+@startuml
+  participant LOADERS as "Metadata Loaders"
+  participant STORAGE as "Graph Storage"
+  participant JOURNAL as "Journal"
+  participant IDX_REM_META as "REM Indexer"
+  participant IDX_STORAGE as "Indexer Storage"
+
+  activate IDX_STORAGE
+  activate STORAGE
+  activate JOURNAL
+  activate LOADERS
+
+  LOADERS->>STORAGE: new REM (Raw Extrinsic Metadata) object\n for Origin http://example.org/repo.git\nor object swh:1:dir:...
+  STORAGE->>JOURNAL: new REM object
+  deactivate LOADERS
+
+  JOURNAL->>IDX_REM_META: run indexers on REM object
+  activate IDX_REM_META
+
+  IDX_REM_META->>IDX_REM_META: recognize REM object (gitea/github/deposit/...)
+
+  IDX_REM_META->>IDX_REM_META: parse REM object
+
+  alt If the REM object describe an origin
+    IDX_REM_META->>IDX_STORAGE: origin_extrinsic_metadata_add(id="http://example.org/repo.git", {author: "Jane Doe", ...})
+    IDX_STORAGE->>IDX_REM_META: ok
+  end
+
+  alt If the REM object describe a directory
+    IDX_REM_META->>IDX_STORAGE: directory_extrinsic_metadata_add(id="swh:1:dir:...", {author: "Jane Doe", ...})
+    IDX_STORAGE->>IDX_REM_META: ok
+  end
+
+  deactivate IDX_REM_META
+
+
+@enduml

docs/images/tasks-metadata-indexers.uml → docs/images/tasks-intrinsic-metadata-indexers.uml

 @startuml
   participant LOADERS as "Loaders"
+  participant STORAGE as "Graph Storage"
   participant JOURNAL as "Journal"
   participant IDX_ORIG_META as "Origin Metadata Indexer"
   participant IDX_ORIG_HEAD as "Origin-Head Indexer"
   participant IDX_DIR_META as "Directory Metadata Indexer"
   participant IDX_CONT_META as "Content Metadata Indexer"
   participant IDX_STORAGE as "Indexer Storage"
-  participant STORAGE as "Graph Storage"
   participant OBJ_STORAGE as "Object Storage"
 
   activate OBJ_STORAGE
@@ -17,7 +17,9 @@
 
   activate LOADERS
 
-  LOADERS->>JOURNAL: Origin http://example.org/repo.git\nwas added/revisited
+  LOADERS->>STORAGE: Repository content
+  LOADERS->>STORAGE: Origin http://example.org/repo.git\nwas added/revisited
+  STORAGE->>JOURNAL: Origin http://example.org/repo.git\nwas added/revisited
   deactivate LOADERS
 
   JOURNAL->>IDX_ORIG_META: run indexers on origin\nhttp://example.org/repo.git

docs/metadata-workflow.rst

@@ -14,7 +14,7 @@ at each step in the indexer storage.
 Indexer architecture
 ^^^^^^^^^^^^^^^^^^^^
 
-.. thumbnail:: images/tasks-metadata-indexers.svg
+.. thumbnail:: images/tasks-intrinsic-metadata-indexers.svg
 
 
 Origin-Head Indexer
@@ -96,6 +96,11 @@ This normalization makes up for most of the code of
 Extrinsic metadata
 ------------------
 
+Indexer architecture
+^^^^^^^^^^^^^^^^^^^^
+
+.. thumbnail:: images/tasks-extrinsic-metadata-indexers.svg
+
 The :term:`extrinsic metadata` indexer works very differently from
 the :term:`intrinsic metadata` indexers we saw above.
 While the latter extract metadata from software artefacts (files and directories)