[tracker/wip/carlosg/hotdoc: 4/40] docs: Move overview to markdown

[Carlos Garnacho <carlosg src gnome org>]

commit 7a2cb88e8b0ff1a0e9ad9adfe3b4fb2a4e4589cd
Author: Carlos Garnacho <carlosg gnome org>
Date:   Thu May 27 18:22:00 2021 +0200

    docs: Move overview to markdown

 docs/reference/libtracker-sparql/meson.build  |  1 +
 docs/reference/libtracker-sparql/overview.md  | 59 ++++++++++++++++++
 docs/reference/libtracker-sparql/overview.xml | 89 ---------------------------
 3 files changed, 60 insertions(+), 89 deletions(-)
---
diff --git a/docs/reference/libtracker-sparql/meson.build b/docs/reference/libtracker-sparql/meson.build
index e30f0fad1..ba22fb2e5 100644
--- a/docs/reference/libtracker-sparql/meson.build
+++ b/docs/reference/libtracker-sparql/meson.build
@@ -1,4 +1,5 @@
 content = [
+  'overview.md',
   'tutorial.md',
 ]
 
diff --git a/docs/reference/libtracker-sparql/overview.md b/docs/reference/libtracker-sparql/overview.md
new file mode 100644
index 000000000..949d81c3f
--- /dev/null
+++ b/docs/reference/libtracker-sparql/overview.md
@@ -0,0 +1,59 @@
+Title: Overview
+
+Tracker SPARQL allows creating and connecting to one or more
+triplestore databases. It is used by the
+[Tracker Miners filesystem indexer](https://gitlab.gnome.org/GNOME/tracker-miners/),
+and can also store and publish any kind of app data.
+
+Querying data is done using the SPARQL graph query language. See the
+[examples](examples.html) to find out how this works.
+Storing data can also be done using SPARQL, or using the [class@Tracker.Resource]
+API.
+
+You can share a database over D-Bus using the [class@Tracker.Endpoint] API,
+allowing other libtracker-sparql users to query from it, either
+by referencing it in a `SELECT { SERVICE ... }` query, or by connecting
+directly with [ctor@Tracker.SparqlConnection.bus_new].
+
+Tracker SPARQL partitions the database into multiple graphs.
+You can implementing access control restrictions based on
+graphs, and we provide a Flatpak portal that does so.
+The number of graphs is [limited](limits.html).
+
+## Connection methods
+
+You can create and access a private store using
+[ctor Tracker SparqlConnection new]. This is useful to store
+app-specific data.
+
+To connect to another database on the same local machine, such as the
+one exposed by Tracker Miner FS, use [ctor@Tracker.SparqlConnection.bus_new].
+
+To connect to another a database on a remote machine, use
+[ctor@Tracker.SparqlConnection.remote_new]. This can be used to query online
+databases that provide a SPARQL endpoint, such as [DBpedia](https://wiki.dbpedia.org/about).
+        .
+## Connecting from Flatpak
+
+Tracker SPARQL provides a portal for the [Flatpak](https://flatpak.org/)
+application sandboxing system. This lets you control which parts of a
+database each app can query.
+
+The app's Flatpak manifest needs to specify which graph(s) the app will
+access. See the [example 
app](https://gitlab.gnome.org/GNOME/tracker/-/blob/master/examples/flatpak/org.example.TrackerSandbox.json)
+and the [portal 
documentation](https://gnome.pages.gitlab.gnome.org/tracker/docs/commandline/#tracker-xdg-portal-3) to see 
how.
+
+No code changes are needed in the app, as [ctor@Tracker.SparqlConnection.bus_new]
+will automatically try connect via the portal if it can't talk to the
+given D-Bus name directly.
+
+Tracker SPARQL is included in the
+[GNOME Flatpak SDK and runtime](https://docs.flatpak.org/en/latest/available-runtimes.html#gnome).
+If the app uses a different runtime, you may need to build Tracker
+SPARQL in the app manifest.
+
+The app might use the Tracker Miner FS search index. Ideally this is done via the portal,
+using the search index maintained on the host. Since not all OSes ship tracker-miner-fs-3,
+we recommend that apps bundle it inside the Flatpak and connect to the bundled version as
+a fallback. See [GNOME Music](https://gitlab.gnome.org/GNOME/gnome-music/)
+for an example of how to do this.