From 2388d0371340700125b094c187ca7f1b838c6bd0 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 12 May 2026 14:22:45 +0300
Subject: [PATCH 01/20] Replaces erlang's record tuple w/ json as search vector
 source

---
 apps/dmt/src/dmt_repository.erl | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/apps/dmt/src/dmt_repository.erl b/apps/dmt/src/dmt_repository.erl
index 4fa58ce..c35ad4d 100644
--- a/apps/dmt/src/dmt_repository.erl
+++ b/apps/dmt/src/dmt_repository.erl
@@ -618,7 +618,16 @@ get_new_version(Worker, AuthorID) ->
 insert_object(Worker, Type, ID0, Version, Data0) ->
     ID1 = dmt_mapper:ref_to_string(ID0),
     Data1 = dmt_mapper:object_to_string(Data0),
-    SearchVector = dmt_mapper:extract_searchable_text_from_term(Data0),
+    %% NOTE Turns string containing nested json into space-separated string of
+    %% words/lexems.
+    %% As example, this turns
+    %%   {
+    %%     "hello": "world",
+    %%     "test": 42
+    %%   }
+    %% into
+    %%   hello world test 42
+    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data1)),
 
     case dmt_database:insert_object(Worker, ID1, Type, Version, Data1, SearchVector) of
         ok ->

From 64533124dbcd801ba69098fe3c483b6577429b4b Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 30 Jun 2026 17:14:23 +0300
Subject: [PATCH 02/20] XYZ-250: Adds JSONish search vector and storage
 migrations

---
 apps/dmt/src/dmt_repository.erl               | 32 +++++++------
 .../dmt_repository_client_tests_SUITE.erl     |  2 +-
 .../test/dmt_search_objects_tests_SUITE.erl   |  2 +-
 elvis.config                                  |  2 +-
 ...2722010-create-search-vector-tmp-table.sql |  6 +++
 ...2011-fill-search-vector-fill-tmp-table.erl | 46 +++++++++++++++++++
 ...-entities-with-search-vector-tmp-table.sql |  4 ++
 ...782722013-drop-search-vector-tmp-table.sql |  1 +
 rebar.config                                  |  1 +
 9 files changed, 78 insertions(+), 18 deletions(-)
 create mode 100644 migrations/1782722010-create-search-vector-tmp-table.sql
 create mode 100644 migrations/1782722011-fill-search-vector-fill-tmp-table.erl
 create mode 100644 migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
 create mode 100644 migrations/1782722013-drop-search-vector-tmp-table.sql

diff --git a/apps/dmt/src/dmt_repository.erl b/apps/dmt/src/dmt_repository.erl
index 4578305..3de4d62 100644
--- a/apps/dmt/src/dmt_repository.erl
+++ b/apps/dmt/src/dmt_repository.erl
@@ -753,19 +753,8 @@ get_new_version(Worker, AuthorID) ->
 ) ->
     object_ref() | no_return().
 insert_object(Worker, Type, ID0, Version, Data0) ->
+    {ok, Data1, SearchVector} = object_to_serialized(Data0),
     ID1 = dmt_mapper:ref_to_string(ID0),
-    Data1 = dmt_mapper:object_to_string(Data0),
-    %% NOTE Turns string containing nested json into space-separated string of
-    %% words/lexems.
-    %% As example, this turns
-    %%   {
-    %%     "hello": "world",
-    %%     "test": 42
-    %%   }
-    %% into
-    %%   hello world test 42
-    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data1)),
-
     case dmt_database:insert_object(Worker, ID1, Type, Version, Data1, SearchVector) of
         ok ->
             ID0;
@@ -807,15 +796,28 @@ get_object_field({_, _, _, ref, _}, _Data, {_Type, Ref}) ->
 get_object_field({_, _, _, data, _}, Data, _Ref) ->
     Data.
 
+-spec object_to_serialized(domain_object()) -> {ok, JsonBinary :: binary(), SearchString :: string()}.
+object_to_serialized(Data0) ->
+    Data1 = dmt_mapper:object_to_string(Data0),
+    %% NOTE Turns string containing nested json into space-separated string of
+    %% words/lexems.
+    %% As example, this turns
+    %%   {
+    %%     "hello": "world",
+    %%     "test": 42
+    %%   }
+    %% into
+    %%   hello world test 42
+    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data1)),
+    {ok, Data1, SearchVector}.
+
 -spec update_object(
     dmt_database:worker(), dmt_database:entity_type(), object_ref(), boolean(), domain_object(), dmt_object:version()
 ) ->
     ok | no_return().
 update_object(Worker, Type, ID0, IsActive, Data0, Version) ->
-    Data1 = dmt_mapper:object_to_string(Data0),
+    {ok, Data1, SearchVector} = object_to_serialized(Data0),
     ID1 = dmt_mapper:ref_to_string(ID0),
-    SearchVector = dmt_mapper:extract_searchable_text_from_term(Data0),
-
     case
         dmt_database:update_object(
             Worker,
diff --git a/apps/dmt/test/dmt_repository_client_tests_SUITE.erl b/apps/dmt/test/dmt_repository_client_tests_SUITE.erl
index 53c66ac..73d0b28 100644
--- a/apps/dmt/test/dmt_repository_client_tests_SUITE.erl
+++ b/apps/dmt/test/dmt_repository_client_tests_SUITE.erl
@@ -36,7 +36,7 @@
 
 %% Initialize per suite
 init_per_suite(Config) ->
-    {Apps, _Ret} = dmt_ct_helper:start_apps([woody, scoper, epg_connector, dmt]),
+    {Apps, _Ret} = dmt_ct_helper:start_apps([woody, scoper, epg_connector, brod, dmt]),
     ApiClient = dmt_ct_helper:create_client(),
     [{client, ApiClient}, {apps, Apps} | Config].
 
diff --git a/apps/dmt/test/dmt_search_objects_tests_SUITE.erl b/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
index 8e6ce2b..d0a50a4 100644
--- a/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
+++ b/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
@@ -42,7 +42,7 @@
 
 %% Initialize per suite
 init_per_suite(Config) ->
-    {Apps, _Ret} = dmt_ct_helper:start_apps([woody, scoper, epg_connector, dmt]),
+    {Apps, _Ret} = dmt_ct_helper:start_apps([woody, scoper, epg_connector, brod, dmt]),
     ApiClient = dmt_ct_helper:create_client(),
     [{client, ApiClient}, {apps, Apps} | Config].
 
diff --git a/elvis.config b/elvis.config
index 2815340..e61cc07 100644
--- a/elvis.config
+++ b/elvis.config
@@ -2,7 +2,7 @@
     {elvis, [
         {config, [
             #{
-                dirs => ["apps/**/src", "apps/**/include"],
+                dirs => ["migrations/*.erl", "apps/**/src", "apps/**/include"],
                 filter => "*.erl",
                 ruleset => erl_files,
                 rules => [
diff --git a/migrations/1782722010-create-search-vector-tmp-table.sql b/migrations/1782722010-create-search-vector-tmp-table.sql
new file mode 100644
index 0000000..d2ea64e
--- /dev/null
+++ b/migrations/1782722010-create-search-vector-tmp-table.sql
@@ -0,0 +1,6 @@
+CREATE TABLE tmp_entity_search_vector (
+    id TEXT NOT NULL,
+    version BIGINT NOT NULL REFERENCES version(version),
+    search_vector tsvector,
+    PRIMARY KEY (id, version)
+);
diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
new file mode 100644
index 0000000..53e21af
--- /dev/null
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -0,0 +1,46 @@
+-module('1782722011-fill-search-vector-fill-tmp-table').
+
+-export([perform/2]).
+
+-define(BATCH_SIZE, 100).
+
+perform(Conn, _MigrationOpts) ->
+    ObjectsCount = count_objects(Conn),
+    BatchesCount = round(math:ceil(ObjectsCount / ?BATCH_SIZE)),
+    ok = lists:foreach(
+        fun(N) ->
+            Objects0 = collect_objects(Conn, Offset, ?BATCH_SIZE),
+            Objects1 = lists:map(fun form_search_vector/1, Objects0),
+            insert_objects_into_buffer(Conn, Objects1)
+        end,
+        lists:seq(0, BatchesCount - 1)
+    ).
+
+%%
+
+count_objects(Conn) ->
+    {ok, _Cols, Count} = epgsql:squery(Conn, "SELECT COUNT(*) FROM entity"),
+    binary_to_integer(Count).
+
+collect_objects(Conn, Offset, Limit) ->
+    Query = """
+    SELECT id, version, entity_type, data
+    FROM entity
+    ORDER BY id, version ASC
+    LIMIT $1 OFFSET $2
+    """,
+    {ok, _Cols, Objects} = epgsql:equery(Conn, Query, [Limit, Offset]),
+    Objects.
+
+form_search_vector({ID, Version, _Type, Data}) ->
+    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data)),
+    [ID, Version, SearchVector].
+
+insert_objects_into_buffer(Conn, Objects) ->
+    epgsql:execute_batch(
+        """
+        INSERT INTO tmp_entity_search_vector (id, version, search_vector)
+        VALUES ($1, $2, to_tsvector('multilingual', $3))
+        """,
+        Objects
+    ).
diff --git a/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql b/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
new file mode 100644
index 0000000..fbf9375
--- /dev/null
+++ b/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
@@ -0,0 +1,4 @@
+UPDATE entity AS entity
+SET entity.search_vector = buffer.search_vector
+FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS buffer
+WHERE entity.id = buffer.id AND entity.version = buffer.version;
diff --git a/migrations/1782722013-drop-search-vector-tmp-table.sql b/migrations/1782722013-drop-search-vector-tmp-table.sql
new file mode 100644
index 0000000..2290f96
--- /dev/null
+++ b/migrations/1782722013-drop-search-vector-tmp-table.sql
@@ -0,0 +1 @@
+DROP TABLE tmp_entity_search_vector;
diff --git a/rebar.config b/rebar.config
index 36d20c7..1f52e8b 100644
--- a/rebar.config
+++ b/rebar.config
@@ -148,6 +148,7 @@
     {print_width, 120},
     {files, [
         "apps/dmt*/{src,include,test}/*.{hrl,erl,app.src}",
+        "migrations/*.erl",
         "rebar.config",
         "elvis.config",
         "config/sys.config",

From 96fabae9a5b8e303c7345bbddde51925f9532b5d Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 30 Jun 2026 17:48:31 +0300
Subject: [PATCH 03/20] Fixes migration's `badarg` when counting objects

---
 ...2011-fill-search-vector-fill-tmp-table.erl | 35 ++++++++++++-------
 ...-entities-with-search-vector-tmp-table.sql |  8 ++---
 2 files changed, 26 insertions(+), 17 deletions(-)

diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index 53e21af..63f5329 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -7,9 +7,9 @@
 perform(Conn, _MigrationOpts) ->
     ObjectsCount = count_objects(Conn),
     BatchesCount = round(math:ceil(ObjectsCount / ?BATCH_SIZE)),
-    ok = lists:foreach(
+    lists:foreach(
         fun(N) ->
-            Objects0 = collect_objects(Conn, Offset, ?BATCH_SIZE),
+            Objects0 = collect_objects(Conn, N * ?BATCH_SIZE, ?BATCH_SIZE),
             Objects1 = lists:map(fun form_search_vector/1, Objects0),
             insert_objects_into_buffer(Conn, Objects1)
         end,
@@ -19,8 +19,10 @@ perform(Conn, _MigrationOpts) ->
 %%
 
 count_objects(Conn) ->
-    {ok, _Cols, Count} = epgsql:squery(Conn, "SELECT COUNT(*) FROM entity"),
-    binary_to_integer(Count).
+    case epgsql:squery(Conn, "SELECT COUNT(*) FROM entity") of
+        {ok, _Cols, [{Count}]} -> binary_to_integer(Count);
+        {error, Error} -> erlang:throw(Error)
+    end.
 
 collect_objects(Conn, Offset, Limit) ->
     Query = """
@@ -29,18 +31,25 @@ collect_objects(Conn, Offset, Limit) ->
     ORDER BY id, version ASC
     LIMIT $1 OFFSET $2
     """,
-    {ok, _Cols, Objects} = epgsql:equery(Conn, Query, [Limit, Offset]),
-    Objects.
+    case epgsql:equery(Conn, Query, [Limit, Offset]) of
+        {ok, _Cols, Objects} -> Objects;
+        {error, Error} -> erlang:throw(Error)
+    end.
 
 form_search_vector({ID, Version, _Type, Data}) ->
     SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data)),
     [ID, Version, SearchVector].
 
 insert_objects_into_buffer(Conn, Objects) ->
-    epgsql:execute_batch(
-        """
-        INSERT INTO tmp_entity_search_vector (id, version, search_vector)
-        VALUES ($1, $2, to_tsvector('multilingual', $3))
-        """,
-        Objects
-    ).
+    case
+        epgsql:execute_batch(
+            """
+            INSERT INTO tmp_entity_search_vector (id, version, search_vector)
+            VALUES ($1, $2, to_tsvector('multilingual', $3))
+            """,
+            Objects
+        )
+    of
+        {error, Error} -> erlang:throw(Error);
+        _ -> ok
+    end.
diff --git a/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql b/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
index fbf9375..1205760 100644
--- a/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
+++ b/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
@@ -1,4 +1,4 @@
-UPDATE entity AS entity
-SET entity.search_vector = buffer.search_vector
-FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS buffer
-WHERE entity.id = buffer.id AND entity.version = buffer.version;
+UPDATE entity
+SET search_vector = tmp.search_vector
+FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS tmp
+WHERE entity.id = tmp.id AND entity.version = tmp.version;

From 9212c4f0f84c31fcf33323080be302eefd5b715e Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Wed, 1 Jul 2026 15:40:08 +0300
Subject: [PATCH 04/20] Refactors filling migration into bulk insert w/
 prepared statement

---
 ...2011-fill-search-vector-fill-tmp-table.erl | 32 +++++++++++++------
 1 file changed, 23 insertions(+), 9 deletions(-)

diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index 63f5329..c7cc80b 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -2,7 +2,7 @@
 
 -export([perform/2]).
 
--define(BATCH_SIZE, 100).
+-define(BATCH_SIZE, 500).
 
 perform(Conn, _MigrationOpts) ->
     ObjectsCount = count_objects(Conn),
@@ -40,16 +40,30 @@ form_search_vector({ID, Version, _Type, Data}) ->
     SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data)),
     [ID, Version, SearchVector].
 
+insert_objects_into_buffer(_Conn, []) ->
+    ok;
 insert_objects_into_buffer(Conn, Objects) ->
-    case
-        epgsql:execute_batch(
-            """
-            INSERT INTO tmp_entity_search_vector (id, version, search_vector)
-            VALUES ($1, $2, to_tsvector('multilingual', $3))
-            """,
-            Objects
+    QueryHead = """
+    INSERT INTO tmp_entity_search_vector (id, version, search_vector)
+    VALUES
+
+    """,
+    Values = lists:join(
+        $,,
+        lists:map(
+            fun({I, [ID, Version, SearchVector]}) ->
+                PH = [ph(I, 1), ph(I, 2), ["to_tsvector('multilingual',", ph(I, 3), ")"]],
+                [$(, lists:join($,, PH), $)]
+            end,
+            lists:zip(lists:seq(1, length(Objects)), Objects)
         )
-    of
+    ),
+    Params = lists:append(Objects),
+    case epgsql:equery(Conn, [QueryHead | Values], Params) of
         {error, Error} -> erlang:throw(Error);
         _ -> ok
     end.
+
+%% Placeholder helper
+ph(I, N) ->
+    [$$, integer_to_binary((I - 1) * 3 + N)].

From 3fb109d269ac4bf599107d464cdb0a1fc82a58af Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Wed, 1 Jul 2026 15:49:15 +0300
Subject: [PATCH 05/20] Consolidates migration commands into single module

---
 apps/dmt/src/dmt_sup.erl                      |  6 +--
 ...2722010-create-search-vector-tmp-table.sql |  6 ---
 ...2011-fill-search-vector-fill-tmp-table.erl | 42 ++++++++++++++++++-
 ...-entities-with-search-vector-tmp-table.sql |  4 --
 ...782722013-drop-search-vector-tmp-table.sql |  1 -
 5 files changed, 43 insertions(+), 16 deletions(-)
 delete mode 100644 migrations/1782722010-create-search-vector-tmp-table.sql
 delete mode 100644 migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
 delete mode 100644 migrations/1782722013-drop-search-vector-tmp-table.sql

diff --git a/apps/dmt/src/dmt_sup.erl b/apps/dmt/src/dmt_sup.erl
index 9f02c07..b34b880 100644
--- a/apps/dmt/src/dmt_sup.erl
+++ b/apps/dmt/src/dmt_sup.erl
@@ -67,10 +67,10 @@ dbinit() ->
     MigrationsPath = WorkDir ++ "/migrations",
     case dmt_db_migration:run(MigrationsPath) of
         ok ->
-            _ = logger:warning("entity_type: ~p", [
-                epg_pool:query(default_pool, "SELECT * FROM entity_type;")
-            ]),
             ok;
+        %% TODO Pass through stacktrace to make migration debug easier.
+        %% Also consider reraise option in migrator transaction and use separate
+        %% transactions per migration.
         {error, Reason} ->
             throw({migrations_error, Reason})
     end.
diff --git a/migrations/1782722010-create-search-vector-tmp-table.sql b/migrations/1782722010-create-search-vector-tmp-table.sql
deleted file mode 100644
index d2ea64e..0000000
--- a/migrations/1782722010-create-search-vector-tmp-table.sql
+++ /dev/null
@@ -1,6 +0,0 @@
-CREATE TABLE tmp_entity_search_vector (
-    id TEXT NOT NULL,
-    version BIGINT NOT NULL REFERENCES version(version),
-    search_vector tsvector,
-    PRIMARY KEY (id, version)
-);
diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index c7cc80b..9e93177 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -5,19 +5,57 @@
 -define(BATCH_SIZE, 500).
 
 perform(Conn, _MigrationOpts) ->
+    ok = create_tmp_table(Conn),
     ObjectsCount = count_objects(Conn),
     BatchesCount = round(math:ceil(ObjectsCount / ?BATCH_SIZE)),
-    lists:foreach(
+    ok = lists:foreach(
         fun(N) ->
             Objects0 = collect_objects(Conn, N * ?BATCH_SIZE, ?BATCH_SIZE),
             Objects1 = lists:map(fun form_search_vector/1, Objects0),
             insert_objects_into_buffer(Conn, Objects1)
         end,
         lists:seq(0, BatchesCount - 1)
-    ).
+    ),
+    ok = update_entities(Conn),
+    ok = drop_tmp_table(Conn).
 
 %%
 
+create_tmp_table(Conn) ->
+    Query = """
+    CREATE TABLE tmp_entity_search_vector (
+        id TEXT NOT NULL,
+        version BIGINT NOT NULL REFERENCES version(version),
+        search_vector tsvector,
+        PRIMARY KEY (id, version)
+    )
+    """,
+    case epgsql:squery(Conn, Query) of
+        {error, Error} -> erlang:throw(Error);
+        _ -> ok
+    end.
+
+update_entities(Conn) ->
+    Query = """
+    UPDATE entity
+    SET search_vector = tmp.search_vector
+    FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS tmp
+    WHERE entity.id = tmp.id AND entity.version = tmp.version
+    """,
+    case epgsql:squery(Conn, Query) of
+        {error, Error} -> erlang:throw(Error);
+        _ -> ok
+    end.
+
+drop_tmp_table(Conn) ->
+    Query = """
+    DROP TABLE tmp_entity_search_vector
+    """,
+    case epgsql:squery(Conn, Query) of
+        {error, Error} -> erlang:throw(Error);
+        _ -> ok
+    end.
+
 count_objects(Conn) ->
     case epgsql:squery(Conn, "SELECT COUNT(*) FROM entity") of
         {ok, _Cols, [{Count}]} -> binary_to_integer(Count);
diff --git a/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql b/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
deleted file mode 100644
index 1205760..0000000
--- a/migrations/1782722012-update-entities-with-search-vector-tmp-table.sql
+++ /dev/null
@@ -1,4 +0,0 @@
-UPDATE entity
-SET search_vector = tmp.search_vector
-FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS tmp
-WHERE entity.id = tmp.id AND entity.version = tmp.version;
diff --git a/migrations/1782722013-drop-search-vector-tmp-table.sql b/migrations/1782722013-drop-search-vector-tmp-table.sql
deleted file mode 100644
index 2290f96..0000000
--- a/migrations/1782722013-drop-search-vector-tmp-table.sql
+++ /dev/null
@@ -1 +0,0 @@
-DROP TABLE tmp_entity_search_vector;

From 591a2d979bc3bdd07d6a7a0dddea46ea5bab1076 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Mon, 6 Jul 2026 17:12:26 +0300
Subject: [PATCH 06/20] Refactors string to search vector formation

---
 apps/dmt/src/dmt_mapper.erl                   | 120 +++++++++++++++++-
 apps/dmt/src/dmt_repository.erl               |  13 +-
 ...2011-fill-search-vector-fill-tmp-table.erl |  49 +++----
 3 files changed, 146 insertions(+), 36 deletions(-)

diff --git a/apps/dmt/src/dmt_mapper.erl b/apps/dmt/src/dmt_mapper.erl
index be8087e..c9ec8e7 100644
--- a/apps/dmt/src/dmt_mapper.erl
+++ b/apps/dmt/src/dmt_mapper.erl
@@ -10,9 +10,9 @@
 -export([string_to_ref/1]).
 -export([object_to_string/1]).
 -export([string_to_object/1]).
+-export([to_search_vector/2]).
 -export([to_string/1]).
 -export([from_string/1]).
--export([extract_searchable_text_from_term/1]).
 
 -type row() :: tuple().
 -type transform_fun() :: fun((map()) -> term()).
@@ -111,6 +111,32 @@ object_to_string({_Type, _} = Data) ->
 string_to_object(Str) ->
     string_to_thrift_term_(Str, ?OBJECT_TYPE).
 
+-spec to_search_vector(binary(), binary()) -> string().
+to_search_vector(RedundantRootKey, Str) ->
+    %% NOTE Turns string containing nested json into space-separated string of
+    %% words/lexems.
+    %% As example, this turns
+    %%   {
+    %%     "my-object": {
+    %%       "ref": {
+    %%         "id": "my-id"
+    %%       },
+    %%       "data": {
+    %%         "hello": "world",
+    %%         "test": 42
+    %%       }
+    %%     }
+    %%   }
+    %% into
+    %%   hello world test 42
+    Json0 =
+        case jsx:decode(Str) of
+            #{RedundantRootKey := Inner} -> Inner;
+            Inner -> Inner
+        end,
+    Json1 = unwrap_redundant_json_nesting(Json0),
+    extract_searchable_text_from_term(Json1).
+
 -spec thrift_term_to_string_(term(), dmt_thrift:thrift_type()) -> binary().
 thrift_term_to_string_(Term, ThriftType) ->
     dmt_json:encode(dmt_json:term_to_json(Term, ThriftType)).
@@ -192,6 +218,12 @@ extract_searchable_text_from_term(Term) ->
     TextList = extract_text(Term, []),
     join_text_list(TextList).
 
+-spec unwrap_redundant_json_nesting(map()) -> map().
+unwrap_redundant_json_nesting(#{~"ref" := _, ~"data" := Data}) ->
+    unwrap_redundant_json_nesting(Data);
+unwrap_redundant_json_nesting(Json0) ->
+    Json0.
+
 -ifdef(TEST).
 
 -include_lib("eunit/include/eunit.hrl").
@@ -229,4 +261,90 @@ stringify_object_test_() ->
         ?_assertEqual(Object, string_to_object(object_to_string(Object)))
     ].
 
+-spec to_search_vector_test_() -> _.
+to_search_vector_test_() ->
+    %% NOTE Since JSON is decoded as map, ordering in output string is not
+    %% guaranteed to be the same w/ different Erlang versions.
+    [
+        ?_assertEqual(
+            "hello world test 42",
+            to_search_vector(
+                ~"my-object",
+                ~"""
+                {
+                  "my-object": {
+                    "ref": {
+                      "id": "my-id"
+                    },
+                    "data": {
+                      "hello": "world",
+                      "test": 42
+                    }
+                  }
+                }
+                """
+            )
+        ),
+        ?_assertEqual(
+            "my-object data hello world test 42 ref id my-id",
+            to_search_vector(
+                ~"other-root-key",
+                ~"""
+                {
+                  "my-object": {
+                    "ref": {
+                      "id": "my-id"
+                    },
+                    "data": {
+                      "hello": "world",
+                      "test": 42
+                    }
+                  }
+                }
+                """
+            )
+        ),
+        ?_assertEqual(
+            "hello world test 42",
+            to_search_vector(
+                ~"my-object",
+                ~"""
+                {
+                  "my-object": {
+                    "hello": "world",
+                    "test": 42
+                  }
+                }
+                """
+            )
+        ),
+        ?_assertEqual(
+            "",
+            to_search_vector(
+                ~"my-object",
+                ~"""
+                {}
+                """
+            )
+        ),
+        ?_assertEqual(
+            "string",
+            to_search_vector(
+                ~"my-object",
+                ~"""
+                "string"
+                """
+            )
+        ),
+        ?_assertEqual(
+            "",
+            to_search_vector(
+                ~"my-object",
+                ~"""
+                null
+                """
+            )
+        )
+    ].
+
 -endif.
diff --git a/apps/dmt/src/dmt_repository.erl b/apps/dmt/src/dmt_repository.erl
index 3de4d62..db896b3 100644
--- a/apps/dmt/src/dmt_repository.erl
+++ b/apps/dmt/src/dmt_repository.erl
@@ -797,18 +797,9 @@ get_object_field({_, _, _, data, _}, Data, _Ref) ->
     Data.
 
 -spec object_to_serialized(domain_object()) -> {ok, JsonBinary :: binary(), SearchString :: string()}.
-object_to_serialized(Data0) ->
+object_to_serialized({Type, _} = Data0) ->
     Data1 = dmt_mapper:object_to_string(Data0),
-    %% NOTE Turns string containing nested json into space-separated string of
-    %% words/lexems.
-    %% As example, this turns
-    %%   {
-    %%     "hello": "world",
-    %%     "test": 42
-    %%   }
-    %% into
-    %%   hello world test 42
-    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data1)),
+    SearchVector = dmt_mapper:to_search_vector(atom_to_binary(Type), Data1),
     {ok, Data1, SearchVector}.
 
 -spec update_object(
diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index 9e93177..dea623f 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -2,6 +2,7 @@
 
 -export([perform/2]).
 
+-define(TMP_TABLE, "tmp_entity_search_vector").
 -define(BATCH_SIZE, 500).
 
 perform(Conn, _MigrationOpts) ->
@@ -22,35 +23,39 @@ perform(Conn, _MigrationOpts) ->
 %%
 
 create_tmp_table(Conn) ->
-    Query = """
-    CREATE TABLE tmp_entity_search_vector (
-        id TEXT NOT NULL,
-        version BIGINT NOT NULL REFERENCES version(version),
-        search_vector tsvector,
-        PRIMARY KEY (id, version)
-    )
-    """,
+    Query = io_lib:format(
+        """
+        CREATE TABLE ~s (
+            id TEXT NOT NULL,
+            version BIGINT NOT NULL REFERENCES version(version),
+            search_vector tsvector,
+            PRIMARY KEY (id, version)
+        )
+        """,
+        [?TMP_TABLE]
+    ),
     case epgsql:squery(Conn, Query) of
         {error, Error} -> erlang:throw(Error);
         _ -> ok
     end.
 
 update_entities(Conn) ->
-    Query = """
-    UPDATE entity
-    SET search_vector = tmp.search_vector
-    FROM (SELECT id, version, search_vector FROM  tmp_entity_search_vector) AS tmp
-    WHERE entity.id = tmp.id AND entity.version = tmp.version
-    """,
+    Query = io_lib:format(
+        """
+        UPDATE entity
+        SET search_vector = tmp.search_vector
+        FROM (SELECT id, version, search_vector FROM ~s) AS tmp
+        WHERE entity.id = tmp.id AND entity.version = tmp.version
+        """,
+        [?TMP_TABLE]
+    ),
     case epgsql:squery(Conn, Query) of
         {error, Error} -> erlang:throw(Error);
         _ -> ok
     end.
 
 drop_tmp_table(Conn) ->
-    Query = """
-    DROP TABLE tmp_entity_search_vector
-    """,
+    Query = io_lib:format("DROP TABLE ~s", [?TMP_TABLE]),
     case epgsql:squery(Conn, Query) of
         {error, Error} -> erlang:throw(Error);
         _ -> ok
@@ -74,18 +79,14 @@ collect_objects(Conn, Offset, Limit) ->
         {error, Error} -> erlang:throw(Error)
     end.
 
-form_search_vector({ID, Version, _Type, Data}) ->
-    SearchVector = dmt_mapper:extract_searchable_text_from_term(jsx:decode(Data)),
+form_search_vector({ID, Version, Type, Data}) ->
+    SearchVector = dmt_mapper:to_search_vector(Type, Data),
     [ID, Version, SearchVector].
 
 insert_objects_into_buffer(_Conn, []) ->
     ok;
 insert_objects_into_buffer(Conn, Objects) ->
-    QueryHead = """
-    INSERT INTO tmp_entity_search_vector (id, version, search_vector)
-    VALUES
-
-    """,
+    QueryHead = io_lib:format("INSERT INTO ~s (id, version, search_vector) VALUES ", [?TMP_TABLE]),
     Values = lists:join(
         $,,
         lists:map(

From 1f71ab880fd02791e247c44962e14cc26bb4366e Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Mon, 6 Jul 2026 22:22:16 +0300
Subject: [PATCH 07/20] Adds `to_text_search_query` string pre-processing func
 for search

---
 apps/dmt/src/dmt_mapper.erl                   | 74 ++++++++++++++++---
 apps/dmt/src/dmt_repository.erl               |  2 +-
 ...2011-fill-search-vector-fill-tmp-table.erl |  2 +-
 3 files changed, 65 insertions(+), 13 deletions(-)

diff --git a/apps/dmt/src/dmt_mapper.erl b/apps/dmt/src/dmt_mapper.erl
index c9ec8e7..7721e2d 100644
--- a/apps/dmt/src/dmt_mapper.erl
+++ b/apps/dmt/src/dmt_mapper.erl
@@ -10,7 +10,8 @@
 -export([string_to_ref/1]).
 -export([object_to_string/1]).
 -export([string_to_object/1]).
--export([to_search_vector/2]).
+-export([to_text_search_vector/2]).
+-export([to_text_search_query/1]).
 -export([to_string/1]).
 -export([from_string/1]).
 
@@ -111,8 +112,8 @@ object_to_string({_Type, _} = Data) ->
 string_to_object(Str) ->
     string_to_thrift_term_(Str, ?OBJECT_TYPE).
 
--spec to_search_vector(binary(), binary()) -> string().
-to_search_vector(RedundantRootKey, Str) ->
+-spec to_text_search_vector(binary(), binary()) -> string().
+to_text_search_vector(RedundantRootKey, Str) ->
     %% NOTE Turns string containing nested json into space-separated string of
     %% words/lexems.
     %% As example, this turns
@@ -137,6 +138,26 @@ to_search_vector(RedundantRootKey, Str) ->
     Json1 = unwrap_redundant_json_nesting(Json0),
     extract_searchable_text_from_term(Json1).
 
+-spec to_text_search_query(binary()) -> binary().
+to_text_search_query(Query0) ->
+    {ok, NoSpecialRE} = re:compile(~"['\"&|!()\\\\]", [unicode, ucp]),
+    {ok, OnlyWordsRE} = re:compile(~"[^\\p{L}\\p{N}\\s\\-]", [unicode, ucp]),
+    F = fun(QueryPart0, Acc) ->
+        QueryPart1 = re:replace(QueryPart0, NoSpecialRE, ~"", [global]),
+        QueryPart2 = re:replace(QueryPart1, OnlyWordsRE, ~" ", [global]),
+        QueryPart3 = iolist_to_binary(QueryPart2),
+        case split_and_trim_into_keywords(QueryPart3) of
+            [] ->
+                Acc;
+            Keywords0 ->
+                Keywords1 = lists:map(fun(KW) -> <<KW/binary, ":*">> end, Keywords0),
+                [binary:join(Keywords1, ~" & ") | Acc]
+        end
+    end,
+    Query1 = genlib_string:to_lower(Query0),
+    Query2 = lists:foldl(F, [], binary:split(Query1, <<$;>>, [global, trim_all])),
+    binary:join(lists:reverse(Query2), ~" | ").
+
 -spec thrift_term_to_string_(term(), dmt_thrift:thrift_type()) -> binary().
 thrift_term_to_string_(Term, ThriftType) ->
     dmt_json:encode(dmt_json:term_to_json(Term, ThriftType)).
@@ -224,6 +245,22 @@ unwrap_redundant_json_nesting(#{~"ref" := _, ~"data" := Data}) ->
 unwrap_redundant_json_nesting(Json0) ->
     Json0.
 
+-spec split_and_trim_into_keywords(binary()) -> [binary()].
+split_and_trim_into_keywords(Bin) ->
+    lists:reverse(split_and_trim_into_keywords(Bin, <<>>, [])).
+
+-spec split_and_trim_into_keywords(binary(), binary(), [binary()]) -> [binary()].
+split_and_trim_into_keywords(<<>>, <<>>, Acc) ->
+    Acc;
+split_and_trim_into_keywords(<<>>, Keyword, Acc) ->
+    [Keyword | Acc];
+split_and_trim_into_keywords(<<$\s, Bin/binary>>, <<>>, Acc) ->
+    split_and_trim_into_keywords(Bin, <<>>, Acc);
+split_and_trim_into_keywords(<<$\s, Bin/binary>>, Keyword, Acc) ->
+    split_and_trim_into_keywords(Bin, <<>>, [Keyword | Acc]);
+split_and_trim_into_keywords(<<H:8, Bin/binary>>, Keyword, Acc) ->
+    split_and_trim_into_keywords(Bin, <<Keyword/binary, H:8>>, Acc).
+
 -ifdef(TEST).
 
 -include_lib("eunit/include/eunit.hrl").
@@ -261,14 +298,14 @@ stringify_object_test_() ->
         ?_assertEqual(Object, string_to_object(object_to_string(Object)))
     ].
 
--spec to_search_vector_test_() -> _.
-to_search_vector_test_() ->
+-spec to_text_search_vector_test_() -> _.
+to_text_search_vector_test_() ->
     %% NOTE Since JSON is decoded as map, ordering in output string is not
     %% guaranteed to be the same w/ different Erlang versions.
     [
         ?_assertEqual(
             "hello world test 42",
-            to_search_vector(
+            to_text_search_vector(
                 ~"my-object",
                 ~"""
                 {
@@ -287,7 +324,7 @@ to_search_vector_test_() ->
         ),
         ?_assertEqual(
             "my-object data hello world test 42 ref id my-id",
-            to_search_vector(
+            to_text_search_vector(
                 ~"other-root-key",
                 ~"""
                 {
@@ -306,7 +343,7 @@ to_search_vector_test_() ->
         ),
         ?_assertEqual(
             "hello world test 42",
-            to_search_vector(
+            to_text_search_vector(
                 ~"my-object",
                 ~"""
                 {
@@ -320,7 +357,7 @@ to_search_vector_test_() ->
         ),
         ?_assertEqual(
             "",
-            to_search_vector(
+            to_text_search_vector(
                 ~"my-object",
                 ~"""
                 {}
@@ -329,7 +366,7 @@ to_search_vector_test_() ->
         ),
         ?_assertEqual(
             "string",
-            to_search_vector(
+            to_text_search_vector(
                 ~"my-object",
                 ~"""
                 "string"
@@ -338,7 +375,7 @@ to_search_vector_test_() ->
         ),
         ?_assertEqual(
             "",
-            to_search_vector(
+            to_text_search_vector(
                 ~"my-object",
                 ~"""
                 null
@@ -347,4 +384,19 @@ to_search_vector_test_() ->
         )
     ].
 
+-spec to_text_search_query_test_() -> _.
+to_text_search_query_test_() ->
+    [
+        ?_assertEqual(~"test:*", to_text_search_query(~"test")),
+        ?_assertEqual(~"hello:* & world:*", to_text_search_query(~"hello world")),
+        ?_assertEqual(
+            ~"this:* & and:* & that:* | something:* & else:*",
+            to_text_search_query(~"this and that; something else")
+        ),
+        ?_assertEqual(
+            ~"weird-request:* & 42:* & special:* & symbols:*",
+            to_text_search_query(~"wEiRd-(rEQuesT)=!%{42[+\\/   spEcIAL@#sym&bols")
+        )
+    ].
+
 -endif.
diff --git a/apps/dmt/src/dmt_repository.erl b/apps/dmt/src/dmt_repository.erl
index db896b3..59ab3f8 100644
--- a/apps/dmt/src/dmt_repository.erl
+++ b/apps/dmt/src/dmt_repository.erl
@@ -799,7 +799,7 @@ get_object_field({_, _, _, data, _}, Data, _Ref) ->
 -spec object_to_serialized(domain_object()) -> {ok, JsonBinary :: binary(), SearchString :: string()}.
 object_to_serialized({Type, _} = Data0) ->
     Data1 = dmt_mapper:object_to_string(Data0),
-    SearchVector = dmt_mapper:to_search_vector(atom_to_binary(Type), Data1),
+    SearchVector = dmt_mapper:to_text_search_vector(atom_to_binary(Type), Data1),
     {ok, Data1, SearchVector}.
 
 -spec update_object(
diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index dea623f..9988e43 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -80,7 +80,7 @@ collect_objects(Conn, Offset, Limit) ->
     end.
 
 form_search_vector({ID, Version, Type, Data}) ->
-    SearchVector = dmt_mapper:to_search_vector(Type, Data),
+    SearchVector = dmt_mapper:to_text_search_vector(Type, Data),
     [ID, Version, SearchVector].
 
 insert_objects_into_buffer(_Conn, []) ->

From 0496bd65d19c6f3e5863c65ffe72f7ee47606360 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Mon, 6 Jul 2026 22:37:05 +0300
Subject: [PATCH 08/20] Adds notes to new func

---
 apps/dmt/src/dmt_mapper.erl | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/apps/dmt/src/dmt_mapper.erl b/apps/dmt/src/dmt_mapper.erl
index 7721e2d..184b661 100644
--- a/apps/dmt/src/dmt_mapper.erl
+++ b/apps/dmt/src/dmt_mapper.erl
@@ -138,6 +138,8 @@ to_text_search_vector(RedundantRootKey, Str) ->
     Json1 = unwrap_redundant_json_nesting(Json0),
     extract_searchable_text_from_term(Json1).
 
+%% @doc Constructs text-search query for full-text search with lexeme's prefix
+%% matching and OR binding via semicolon (";") character.
 -spec to_text_search_query(binary()) -> binary().
 to_text_search_query(Query0) ->
     {ok, NoSpecialRE} = re:compile(~"['\"&|!()\\\\]", [unicode, ucp]),
@@ -156,6 +158,9 @@ to_text_search_query(Query0) ->
     end,
     Query1 = genlib_string:to_lower(Query0),
     Query2 = lists:foldl(F, [], binary:split(Query1, <<$;>>, [global, trim_all])),
+    %% NOTE Parentheses can be used to enforce grouping of these operators. In
+    %% the absence of parentheses, ! (NOT) binds most tightly, <-> (FOLLOWED BY)
+    %% next most tightly, then & (AND), with | (OR) binding the least tightly.
     binary:join(lists:reverse(Query2), ~" | ").
 
 -spec thrift_term_to_string_(term(), dmt_thrift:thrift_type()) -> binary().

From 121ed2eebdb2fb768236ce0b7d2bf787db39bb0c Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 15:04:23 +0300
Subject: [PATCH 09/20] Adds use of pre-processed search query w/ testcases

---
 apps/dmt/src/dmt_database.erl                 |  30 +--
 apps/dmt/src/dmt_db_migration.erl             |   2 +-
 apps/dmt/src/dmt_mapper.erl                   |  62 ++++--
 apps/dmt/src/dmt_repository.erl               |   2 +-
 .../test/dmt_search_objects_tests_SUITE.erl   | 185 +++++++-----------
 ...2011-fill-search-vector-fill-tmp-table.erl |   2 +-
 6 files changed, 136 insertions(+), 147 deletions(-)

diff --git a/apps/dmt/src/dmt_database.erl b/apps/dmt/src/dmt_database.erl
index 27ed3de..384776e 100644
--- a/apps/dmt/src/dmt_database.erl
+++ b/apps/dmt/src/dmt_database.erl
@@ -683,7 +683,7 @@ search_objects(Worker, <<"*">>, Version, Type, Limit, Offset) ->
             _ = logger:error("Error fetching search Params: ~p", [AllParams]),
             {ok, {[], undefined}}
     end;
-search_objects(Worker, Query, Version, Type, Limit, Offset) ->
+search_objects(Worker, Query0, Version, Type, Limit, Offset) ->
     % Use a pattern where the condition is always true when Type is NULL
     TypeValue =
         case Type of
@@ -713,22 +713,23 @@ search_objects(Worker, Query, Version, Type, Limit, Offset) ->
            e.created_at
     FROM entity e
     INNER JOIN ActiveStatusAtRequestedTime las ON e.id = las.id
-    WHERE e.search_vector @@ plainto_tsquery('multilingual', $1)
+    WHERE e.search_vector @@ to_tsquery('multilingual', $1)
     AND e.version <= $2
     AND ($3 = 'NULL' OR e.entity_type = $3)
     AND las.is_active = TRUE
-    ORDER BY e.id, e.version DESC, ts_rank(e.search_vector, plainto_tsquery('multilingual', $1)) DESC
+    ORDER BY e.id, e.version DESC, ts_rank(e.search_vector, to_tsquery('multilingual', $1)) DESC
     LIMIT $4 OFFSET $5
     """,
 
-    AllParams = [Query, Version, TypeValue, Limit, Offset],
+    Query1 = dmt_mapper:to_text_search_query(Query0),
+    AllParams = [Query1, Version, TypeValue, Limit, Offset],
 
     % Execute the query
     case epg_pool:query(Worker, Request, AllParams) of
         {ok, Columns, Rows} ->
             Objects = dmt_mapper:to_marshalled_maps(Columns, Rows),
             NewOffset0 = Offset + Limit,
-            HasMoreResults = has_more_search_results(Worker, Query, Version, TypeValue, NewOffset0),
+            HasMoreResults = has_more_search_results(Worker, Query1, Version, TypeValue, NewOffset0),
             NewOffset1 =
                 case HasMoreResults of
                     true -> NewOffset0;
@@ -795,7 +796,7 @@ has_more_search_results(Worker, Query, Version, TypeValue, Offset) ->
     )
     SELECT 1 FROM entity e
     INNER JOIN ActiveStatusAtRequestedTime las ON e.id = las.id
-    WHERE e.search_vector @@ plainto_tsquery('multilingual', $1)
+    WHERE e.search_vector @@ to_tsquery('multilingual', $1)
     AND e.version <= $2
     AND ($3 = 'NULL' OR e.entity_type = $3)
     AND las.is_active = TRUE
@@ -1209,7 +1210,7 @@ get_multiple_related_graph_edges(Worker, ObjectRefStrings, Version, Depth, Inclu
 ) ->
     {ok, {[dmt_object:object()], [edge()]}} | {error, db_error()}.
 search_related_graph(
-    Worker, Query, SearchedType, Version, Depth, IncludeInbound, IncludeOutbound, ReturnedType
+    Worker, Query0, SearchedType, Version, Depth, IncludeInbound, IncludeOutbound, ReturnedType
 ) ->
     TypeValue =
         case SearchedType of
@@ -1218,7 +1219,7 @@ search_related_graph(
         end,
 
     SearchQuery =
-        case Query of
+        case Query0 of
             ~"*" ->
                 """
                 WITH LatestVersionAtRequestedTime AS (
@@ -1256,18 +1257,21 @@ search_related_graph(
                 SELECT DISTINCT ON (e.id) e.id
                 FROM entity e
                 INNER JOIN ActiveStatusAtRequestedTime las ON e.id = las.id
-                WHERE e.search_vector @@ plainto_tsquery('multilingual', $1)
+                WHERE e.search_vector @@ to_tsquery('multilingual', $1)
                 AND e.version <= $2
                 AND ($3 = 'NULL' OR e.entity_type = $3)
                 AND las.is_active = TRUE
-                ORDER BY e.id, e.version DESC, ts_rank(e.search_vector, plainto_tsquery('multilingual', $1)) DESC
+                ORDER BY e.id, e.version DESC, ts_rank(e.search_vector, to_tsquery('multilingual', $1)) DESC
                 """
         end,
 
     SearchParams =
-        case Query of
-            ~"*" -> [Version, TypeValue];
-            _ -> [Query, Version, TypeValue]
+        case Query0 of
+            ~"*" ->
+                [Version, TypeValue];
+            _ ->
+                Query1 = dmt_mapper:to_text_search_query(Query0),
+                [Query1, Version, TypeValue]
         end,
 
     case epg_pool:query(Worker, SearchQuery, SearchParams) of
diff --git a/apps/dmt/src/dmt_db_migration.erl b/apps/dmt/src/dmt_db_migration.erl
index 066d1d3..f6c6bb1 100644
--- a/apps/dmt/src/dmt_db_migration.erl
+++ b/apps/dmt/src/dmt_db_migration.erl
@@ -32,7 +32,7 @@ run(MigrationsDir) ->
 
 -spec perform(map(), string()) -> ok | {error, term()}.
 perform(DbOpts, MigrationsDir) ->
-    case epg_migrator:perform(?REALM, DbOpts, [], MigrationsDir) of
+    case epg_migrator:perform(?REALM, DbOpts, [reraise], MigrationsDir) of
         {ok, Executed} ->
             _ = logger:info("Applied migrations: ~p", [Executed]),
             ok;
diff --git a/apps/dmt/src/dmt_mapper.erl b/apps/dmt/src/dmt_mapper.erl
index 184b661..95bee59 100644
--- a/apps/dmt/src/dmt_mapper.erl
+++ b/apps/dmt/src/dmt_mapper.erl
@@ -112,8 +112,8 @@ object_to_string({_Type, _} = Data) ->
 string_to_object(Str) ->
     string_to_thrift_term_(Str, ?OBJECT_TYPE).
 
--spec to_text_search_vector(binary(), binary()) -> string().
-to_text_search_vector(RedundantRootKey, Str) ->
+-spec to_text_search_vector(binary(), binary()) -> binary().
+to_text_search_vector(RedundantRootKey, Bin) ->
     %% NOTE Turns string containing nested json into space-separated string of
     %% words/lexems.
     %% As example, this turns
@@ -131,12 +131,13 @@ to_text_search_vector(RedundantRootKey, Str) ->
     %% into
     %%   hello world test 42
     Json0 =
-        case jsx:decode(Str) of
+        case jsx:decode(Bin) of
             #{RedundantRootKey := Inner} -> Inner;
             Inner -> Inner
         end,
     Json1 = unwrap_redundant_json_nesting(Json0),
-    extract_searchable_text_from_term(Json1).
+    Str = extract_searchable_text_from_term(Json1),
+    to_unaccented_lowercase(unicode:characters_to_binary(Str)).
 
 %% @doc Constructs text-search query for full-text search with lexeme's prefix
 %% matching and OR binding via semicolon (";") character.
@@ -144,7 +145,8 @@ to_text_search_vector(RedundantRootKey, Str) ->
 to_text_search_query(Query0) ->
     {ok, NoSpecialRE} = re:compile(~"['\"&|!()\\\\]", [unicode, ucp]),
     {ok, OnlyWordsRE} = re:compile(~"[^\\p{L}\\p{N}\\s\\-]", [unicode, ucp]),
-    F = fun(QueryPart0, Acc) ->
+    KeywordSuffixFun = fun(KW) -> <<KW/binary, ":*">> end,
+    QueryProcessorFun = fun(QueryPart0, Acc) ->
         QueryPart1 = re:replace(QueryPart0, NoSpecialRE, ~"", [global]),
         QueryPart2 = re:replace(QueryPart1, OnlyWordsRE, ~" ", [global]),
         QueryPart3 = iolist_to_binary(QueryPart2),
@@ -152,12 +154,13 @@ to_text_search_query(Query0) ->
             [] ->
                 Acc;
             Keywords0 ->
-                Keywords1 = lists:map(fun(KW) -> <<KW/binary, ":*">> end, Keywords0),
+                Keywords1 = lists:map(KeywordSuffixFun, Keywords0),
                 [binary:join(Keywords1, ~" & ") | Acc]
         end
     end,
-    Query1 = genlib_string:to_lower(Query0),
-    Query2 = lists:foldl(F, [], binary:split(Query1, <<$;>>, [global, trim_all])),
+    %% Query1 = genlib_string:to_lower(Query0),
+    Query1 = to_unaccented_lowercase(Query0),
+    Query2 = lists:foldl(QueryProcessorFun, [], binary:split(Query1, <<$;>>, [global, trim_all])),
     %% NOTE Parentheses can be used to enforce grouping of these operators. In
     %% the absence of parentheses, ! (NOT) binds most tightly, <-> (FOLLOWED BY)
     %% next most tightly, then & (AND), with | (OR) binding the least tightly.
@@ -266,6 +269,19 @@ split_and_trim_into_keywords(<<$\s, Bin/binary>>, Keyword, Acc) ->
 split_and_trim_into_keywords(<<H:8, Bin/binary>>, Keyword, Acc) ->
     split_and_trim_into_keywords(Bin, <<Keyword/binary, H:8>>, Acc).
 
+-spec to_unaccented_lowercase(binary()) -> binary().
+to_unaccented_lowercase(Str0) ->
+    {ok, Re} = re:compile("\\p{Mn}", [unicode]),
+    Str1 = throw_if_bad_binary(unicode:characters_to_nfd_binary(Str0)),
+    Str2 = re:replace(Str1, Re, "", [global]),
+    Str3 = throw_if_bad_binary(unicode:characters_to_nfc_binary(Str2)),
+    throw_if_bad_binary(unicode:characters_to_binary(string:lowercase(Str3))).
+
+-spec throw_if_bad_binary(binary() | {error, _, _}) -> binary() | no_return().
+throw_if_bad_binary({error, _, _}) -> erlang:throw(bad_binary);
+throw_if_bad_binary({incomplete, _, _}) -> erlang:throw(bad_binary);
+throw_if_bad_binary(V) -> V.
+
 -ifdef(TEST).
 
 -include_lib("eunit/include/eunit.hrl").
@@ -309,7 +325,25 @@ to_text_search_vector_test_() ->
     %% guaranteed to be the same w/ different Erlang versions.
     [
         ?_assertEqual(
-            "hello world test 42",
+            ~"key значение с кириллицеи и акцентом е",
+            to_text_search_vector(
+                ~"my-object",
+                ~"""
+                {
+                  "my-object": {
+                    "ref": {
+                      "id": "my-id"
+                    },
+                    "data": {
+                      "KEY": "Значение с кириллицей и акцентом Ё"
+                    }
+                  }
+                }
+                """
+            )
+        ),
+        ?_assertEqual(
+            ~"hello world test 42",
             to_text_search_vector(
                 ~"my-object",
                 ~"""
@@ -328,7 +362,7 @@ to_text_search_vector_test_() ->
             )
         ),
         ?_assertEqual(
-            "my-object data hello world test 42 ref id my-id",
+            ~"my-object data hello world test 42 ref id my-id",
             to_text_search_vector(
                 ~"other-root-key",
                 ~"""
@@ -347,7 +381,7 @@ to_text_search_vector_test_() ->
             )
         ),
         ?_assertEqual(
-            "hello world test 42",
+            ~"hello world test 42",
             to_text_search_vector(
                 ~"my-object",
                 ~"""
@@ -361,7 +395,7 @@ to_text_search_vector_test_() ->
             )
         ),
         ?_assertEqual(
-            "",
+            ~"",
             to_text_search_vector(
                 ~"my-object",
                 ~"""
@@ -370,7 +404,7 @@ to_text_search_vector_test_() ->
             )
         ),
         ?_assertEqual(
-            "string",
+            ~"string",
             to_text_search_vector(
                 ~"my-object",
                 ~"""
@@ -379,7 +413,7 @@ to_text_search_vector_test_() ->
             )
         ),
         ?_assertEqual(
-            "",
+            ~"",
             to_text_search_vector(
                 ~"my-object",
                 ~"""
diff --git a/apps/dmt/src/dmt_repository.erl b/apps/dmt/src/dmt_repository.erl
index 59ab3f8..5688409 100644
--- a/apps/dmt/src/dmt_repository.erl
+++ b/apps/dmt/src/dmt_repository.erl
@@ -796,7 +796,7 @@ get_object_field({_, _, _, ref, _}, _Data, {_Type, Ref}) ->
 get_object_field({_, _, _, data, _}, Data, _Ref) ->
     Data.
 
--spec object_to_serialized(domain_object()) -> {ok, JsonBinary :: binary(), SearchString :: string()}.
+-spec object_to_serialized(domain_object()) -> {ok, JsonBinary :: binary(), SearchString :: binary()}.
 object_to_serialized({Type, _} = Data0) ->
     Data1 = dmt_mapper:object_to_string(Data0),
     SearchVector = dmt_mapper:to_text_search_vector(atom_to_binary(Type), Data1),
diff --git a/apps/dmt/test/dmt_search_objects_tests_SUITE.erl b/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
index d0a50a4..0306c77 100644
--- a/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
+++ b/apps/dmt/test/dmt_search_objects_tests_SUITE.erl
@@ -34,10 +34,10 @@
     search_full_objects_basic_test/1,
     search_full_objects_with_filter_test/1,
     search_full_objects_pagination_test/1,
-    search_objects_without_name_desc_test/1,
     search_deleted_objects_test/1,
     checkout_deleted_version_test/1,
-    search_updated_object_deduplication_test/1
+    search_updated_object_deduplication_test/1,
+    search_prefix_match_test/1
 ]).
 
 %% Initialize per suite
@@ -68,14 +68,14 @@ groups() ->
             search_with_version_filter_test,
             search_multiple_terms_test,
             search_no_results_test,
-            search_objects_without_name_desc_test,
             search_invalid_query_test,
             search_full_objects_basic_test,
             search_full_objects_with_filter_test,
             search_full_objects_pagination_test,
             search_deleted_objects_test,
             checkout_deleted_version_test,
-            search_updated_object_deduplication_test
+            search_updated_object_deduplication_test,
+            search_prefix_match_test
         ]}
     ].
 
@@ -1028,118 +1028,6 @@ search_full_objects_pagination_test(Config) ->
         length(AllRefs), length(UniqueRefs), "Should not have any duplicate results across pages"
     ).
 
-% Test searching for objects without name or description fields
-search_objects_without_name_desc_test(Config) ->
-    Client = dmt_ct_helper:cfg(client, Config),
-
-    % Create author
-    Email = <<"search_objects_without_name_desc_test@test">>,
-    AuthorID = create_author(Email, Client),
-
-    % Ensure the 'dummy' entity type exists in the database, it could be deleted by previous tests
-    {ok, _} = epg_pool:query(
-        default_pool,
-        """
-        INSERT INTO entity_type (name, has_sequence)
-        VALUES ($1, FALSE)
-        ON CONFLICT (name) DO NOTHING;
-        """,
-        [dummy]
-    ),
-
-    % Create a Dummy object which doesn't have name or description fields
-    Revision = 0,
-    DummyRef = #domain_DummyRef{id = <<"search_objects_without_name_desc_test">>},
-    Operations = [
-        {insert, #domain_conf_v2_InsertOp{
-            force_ref = {dummy, DummyRef},
-            object = {dummy, #domain_Dummy{}}
-        }}
-    ],
-
-    {ok, #domain_conf_v2_CommitResponse{
-        version = Version,
-        new_objects = NewObjects
-    }} = dmt_client:commit(Revision, Operations, AuthorID, Client),
-
-    % Extract the created dummy reference
-    [{dummy, #domain_DummyObject{ref = DummyRef}}] = ordsets:to_list(NewObjects),
-
-    % Search for dummy objects (should find by type since there's no name/description)
-    Request = #domain_conf_v2_SearchRequestParams{
-        % This should match the type name
-        query = <<"dummy">>,
-        version = Version,
-        limit = 10
-    },
-
-    {ok, #domain_conf_v2_SearchResponse{
-        result = Results,
-        total_count = Count
-    }} = dmt_client:search_objects(Request, Client),
-
-    % We should still find the object even though it doesn't have name/description
-    ?assertEqual(1, Count, "Should find the dummy object"),
-
-    % Verify the result structure
-    ?assertMatch(
-        [
-            #domain_conf_v2_LimitedVersionedObject{
-                ref = {dummy, DummyRef},
-                info = #domain_conf_v2_VersionedObjectInfo{
-                    version = Version
-                },
-                name = undefined,
-                description = undefined
-            }
-        ],
-        Results,
-        "Should return the dummy object with undefined name and description"
-    ),
-
-    % Also search with explicit type filter
-    TypedRequest = #domain_conf_v2_SearchRequestParams{
-        % Empty query but with type filter
-        query = <<"*">>,
-        version = Version,
-        limit = 10,
-        type = dummy
-    },
-
-    {ok, #domain_conf_v2_SearchResponse{
-        result = TypedResults,
-        total_count = TypedCount
-    }} = dmt_client:search_objects(TypedRequest, Client),
-
-    % Should find by type even with empty query
-    ?assertEqual(1, TypedCount, "Should find the dummy object by type filter"),
-    ?assertEqual(Results, TypedResults, "Same result should be returned for type-based search"),
-
-    % Also test with full objects search
-    {ok, #domain_conf_v2_SearchFullResponse{
-        result = FullResults,
-        total_count = FullCount
-    }} = dmt_client:search_full_objects(Request, Client),
-
-    % Verify full object search works too
-    ?assertEqual(1, FullCount, "Should find the dummy object in full search"),
-    ?assertMatch(
-        [
-            #domain_conf_v2_VersionedObject{
-                info = #domain_conf_v2_VersionedObjectInfo{
-                    version = Version
-                },
-                object =
-                    {dummy, #domain_DummyObject{
-                        ref = DummyRef,
-                        data = #domain_Dummy{}
-                    }}
-            }
-        ],
-        FullResults,
-        "Should return the full dummy object"
-    ).
-
 % Test searching for deleted objects
 search_deleted_objects_test(Config) ->
     Client = dmt_ct_helper:cfg(client, Config),
@@ -1588,7 +1476,59 @@ search_updated_object_deduplication_test(Config) ->
     ?assertEqual(UpdatedName2, FullName, "Full object should have the latest name"),
     ?assertEqual(UpdatedDesc2, FullDesc, "Full object should have the latest description").
 
-%% Helper function
+-define(assertFound(Count, Query, Client),
+    ?assertMatch(
+        #domain_conf_v2_SearchResponse{total_count = Count},
+        search(Query, 10, Client),
+        unicode:characters_to_binary(
+            io_lib:format("Failed to assert that search with '~s' finds ~c results", [Query, Count])
+        )
+    )
+).
+
+search_prefix_match_test(Config) ->
+    Client = dmt_ct_helper:cfg(client, Config),
+    ok = insert(
+        [
+            {category, #domain_Category{
+                name = ~"Main regularly reward",
+                description = ~"""
+                Hard hopéful hope widé kindly angle truth dock quiet two clearly
+                all. Us real airport learn doorway problem most catch because
+                first local sing always simple drink. Join decision push honor
+                hot speed, doctor similar valley read think press builder
+                Wédnésday watch try airport available certainly wire.
+                """
+            }},
+            {category, #domain_Category{
+                name = ~"Йнцидент не исчёрпан: кровь стынЁт в жЙлах",
+                description = ~"""
+                Ясность нашей позиции очевидна: постоянный количественный рост и
+                сфера нашей активности прекрасно подходит для реализации
+                дальнейших направлений развития. Господа, глубокий уровень
+                погружения говорит о возможностях поставленных обществом задач.
+                """
+            }}
+        ],
+        create_author(~"search_prefix_match_test@test", Client),
+        Client
+    ),
+    ?assertFound(1, ~"main reward", Client),
+    ?assertFound(0, ~"ololo not found", Client),
+    ?assertFound(1, ~"hard hop", Client),
+    ?assertFound(1, ~"wednesd", Client),
+    ?assertFound(1, ~"local sing always simple drink", Client),
+    ?assertFound(1, ~"инцидент исчерпан", Client),
+    ?assertFound(1, ~"стын", Client),
+    ?assertFound(1, ~"возможно обще", Client),
+    ?assertFound(0, ~"совершенно другой текст", Client),
+    ?assertFound(1, ~"совершенно другой текст;реал", Client),
+    ?assertFound(1, ~"количество сфера", Client),
+    ?assertFound(2, ~"количество сфера; air wir", Client),
+    ok.
+
+%% Helper functions
+
 create_author(Email, Client) ->
     AuthorParams = #domain_conf_v2_AuthorParams{
         email = Email,
@@ -1596,3 +1536,14 @@ create_author(Email, Client) ->
     },
     {ok, #domain_conf_v2_Author{id = AuthorID}} = dmt_client:create_author(AuthorParams, Client),
     AuthorID.
+
+insert(Objects, AuthorID, Client) ->
+    InsertOperations = [{insert, #domain_conf_v2_InsertOp{object = O}} || O <- Objects],
+    {ok, #domain_conf_v2_CommitResponse{version = _, new_objects = _}} =
+        dmt_client:commit(0, InsertOperations, AuthorID, Client),
+    ok.
+
+search(Query, Limit, Client) ->
+    {ok, #domain_conf_v2_SearchResponse{} = Result} =
+        dmt_client:search_objects(#domain_conf_v2_SearchRequestParams{query = Query, limit = Limit}, Client),
+    Result.
diff --git a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
index 9988e43..9f9eed1 100644
--- a/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
+++ b/migrations/1782722011-fill-search-vector-fill-tmp-table.erl
@@ -90,7 +90,7 @@ insert_objects_into_buffer(Conn, Objects) ->
     Values = lists:join(
         $,,
         lists:map(
-            fun({I, [ID, Version, SearchVector]}) ->
+            fun({I, [_ID, _Version, _SearchVector]}) ->
                 PH = [ph(I, 1), ph(I, 2), ["to_tsvector('multilingual',", ph(I, 3), ")"]],
                 [$(, lists:join($,, PH), $)]
             end,

From 9c3204ffe60f069a3e7b818f20fdbae401ad4358 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 16:43:45 +0300
Subject: [PATCH 10/20] Adds host's uid and gid as dev user inside
 `Dockerfile.dev` image

---
 .github/workflows/erlang-checks.yml | 2 +-
 Dockerfile.dev                      | 8 +++++++-
 Makefile                            | 6 ++++--
 apps/dmt/src/dmt_mapper.erl         | 1 -
 compose.yaml                        | 2 ++
 5 files changed, 14 insertions(+), 5 deletions(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 2e6caee..2fb12cf 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@v1
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@527977799bb5b747c0110d4b47c9b6154f5ba988
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}
diff --git a/Dockerfile.dev b/Dockerfile.dev
index 4c0b0f0..ef6b151 100644
--- a/Dockerfile.dev
+++ b/Dockerfile.dev
@@ -11,9 +11,15 @@ RUN wget -q -O- "https://github.com/valitydev/thrift/releases/download/${THRIFT_
 
 RUN apt-get update && apt-get install -y cmake
 
+# Setup user for matching permissions in mounted volumes
+ARG USER_UID
+ARG USER_GID
+RUN useradd -ms /bin/bash dev
+USER dev
+
 # Set env
 ENV CHARSET=UTF-8
 ENV LANG=C.UTF-8
 
 # Set runtime
-CMD ["/bin/bash"]
\ No newline at end of file
+CMD ["/bin/bash"]
diff --git a/Makefile b/Makefile
index d16562f..a81aee9 100644
--- a/Makefile
+++ b/Makefile
@@ -13,10 +13,12 @@ DOTENV := $(shell grep -v '^\#' .env)
 # Development images
 DEV_IMAGE_TAG = $(TEST_CONTAINER_NAME)-dev
 DEV_IMAGE_ID = $(file < .image.dev)
+USER_UID=$(shell id -u)
+USER_GID=$(shell id -g)
 
 DOCKER ?= docker
 DOCKERCOMPOSE ?= docker compose
-DOCKERCOMPOSE_W_ENV = DEV_IMAGE_TAG=$(DEV_IMAGE_TAG) $(DOCKERCOMPOSE) -f compose.yaml -f compose.tracing.yaml
+DOCKERCOMPOSE_W_ENV = USER_UID=$(USER_UID) USER_GID=$(USER_GID) DEV_IMAGE_TAG=$(DEV_IMAGE_TAG) $(DOCKERCOMPOSE) -f compose.yaml -f compose.tracing.yaml
 REBAR ?= rebar3
 TEST_CONTAINER_NAME ?= testrunner
 
@@ -51,7 +53,7 @@ wc-%: dev-image
 	$(DOCKER_RUN) $(DEV_IMAGE_TAG) make $*
 
 wdeps-shell: dev-image
-	$(DOCKERCOMPOSE_RUN) $(TEST_CONTAINER_NAME) su; \
+	$(DOCKERCOMPOSE_RUN) $(TEST_CONTAINER_NAME) bash; \
 	$(DOCKERCOMPOSE_W_ENV) down
 
 wdeps-%: dev-image
diff --git a/apps/dmt/src/dmt_mapper.erl b/apps/dmt/src/dmt_mapper.erl
index 95bee59..07fc389 100644
--- a/apps/dmt/src/dmt_mapper.erl
+++ b/apps/dmt/src/dmt_mapper.erl
@@ -158,7 +158,6 @@ to_text_search_query(Query0) ->
                 [binary:join(Keywords1, ~" & ") | Acc]
         end
     end,
-    %% Query1 = genlib_string:to_lower(Query0),
     Query1 = to_unaccented_lowercase(Query0),
     Query2 = lists:foldl(QueryProcessorFun, [], binary:split(Query1, <<$;>>, [global, trim_all])),
     %% NOTE Parentheses can be used to enforce grouping of these operators. In
diff --git a/compose.yaml b/compose.yaml
index e573ecd..3128148 100644
--- a/compose.yaml
+++ b/compose.yaml
@@ -14,6 +14,8 @@ services:
       args:
         OTP_VERSION: $OTP_VERSION
         THRIFT_VERSION: $THRIFT_VERSION
+        USER_UID: $USER_UID
+        USER_GID: $USER_GID
     volumes:
       - .:$PWD
     hostname: dmt.default

From 61f13c47443fdfa32ef6173330856a4ff3b6bc48 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 16:53:33 +0300
Subject: [PATCH 11/20] Bumps `cache-version`

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 2fb12cf..fe5bd6a 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -38,4 +38,4 @@ jobs:
       thrift-version: ${{ needs.setup.outputs.thrift-version }}
       run-ct-with-compose: true
       upload-coverage: false
-      cache-version: v4
+      cache-version: v5

From 5b98340ae4166cefd0177ce52f7746bbfb6daa04 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 18:58:04 +0300
Subject: [PATCH 12/20] Debug permissions WF

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index fe5bd6a..8ae1690 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@527977799bb5b747c0110d4b47c9b6154f5ba988
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@ddb84107d75c8286694f2f2b61fcae6efe0fcd88
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From a7d8825bf1d0c405d1f7810882625692e9401047 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 19:09:51 +0300
Subject: [PATCH 13/20] Bumps ci wf action

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 8ae1690..c4631ee 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@ddb84107d75c8286694f2f2b61fcae6efe0fcd88
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@7c5c53ab838da69bc1ac9e0e35b6d78858694cdc
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From 5f092236fe587d846469a6545aaeaa2c1eefbe4a Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 19:16:59 +0300
Subject: [PATCH 14/20] Bumps ci wf action

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index c4631ee..b97d54d 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@7c5c53ab838da69bc1ac9e0e35b6d78858694cdc
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@694c874bc8317d0c8fcd9f7e4def77cdfbad71c3
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From 43e19fdf37645321ad394e8e5bce449b98c6efad Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 19:25:08 +0300
Subject: [PATCH 15/20] Adds user uid/gid to compose

---
 compose.yaml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/compose.yaml b/compose.yaml
index 3128148..d88fa1c 100644
--- a/compose.yaml
+++ b/compose.yaml
@@ -16,6 +16,7 @@ services:
         THRIFT_VERSION: $THRIFT_VERSION
         USER_UID: $USER_UID
         USER_GID: $USER_GID
+    user: "${USER_UID}:${USER_GID}"
     volumes:
       - .:$PWD
     hostname: dmt.default

From da87ff672c8aee7db87078aca6f06db5658ea33e Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 20:11:11 +0300
Subject: [PATCH 16/20] Bumps ci wf action

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index b97d54d..80769ef 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@694c874bc8317d0c8fcd9f7e4def77cdfbad71c3
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@97ea610ac4200a5d865615ce6776686ad30b296d
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From cc7c4dcbf031d7227da6d25b20176039a81af1c9 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 7 Jul 2026 20:27:04 +0300
Subject: [PATCH 17/20] Bumps ci wf action

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 80769ef..5c81868 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@97ea610ac4200a5d865615ce6776686ad30b296d
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@4855ca2f829fd8aa0b3189b8f706f2fb5f428943
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From e23bf23f24a331d5c18a6175a2fc4e43e6977654 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Thu, 9 Jul 2026 11:38:45 +0300
Subject: [PATCH 18/20] Bumps ci wf

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 5c81868..895c007 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@4855ca2f829fd8aa0b3189b8f706f2fb5f428943
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@5b5b09ecfccfdf758d4daecdfa0c31cc9e4ed8bd
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}

From f0ea80413ca38049e8dd5db39648d02e1b31f3d8 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 14 Jul 2026 14:14:16 +0300
Subject: [PATCH 19/20] Makes use of uid/gid args explicit and passes them thru
 w/ defaults

---
 Dockerfile.dev | 7 ++++---
 Makefile       | 4 ++--
 compose.yaml   | 2 +-
 3 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/Dockerfile.dev b/Dockerfile.dev
index ef6b151..2ead6be 100644
--- a/Dockerfile.dev
+++ b/Dockerfile.dev
@@ -12,9 +12,10 @@ RUN wget -q -O- "https://github.com/valitydev/thrift/releases/download/${THRIFT_
 RUN apt-get update && apt-get install -y cmake
 
 # Setup user for matching permissions in mounted volumes
-ARG USER_UID
-ARG USER_GID
-RUN useradd -ms /bin/bash dev
+ARG USER_UID=1000
+ARG USER_GID=1000
+RUN groupadd -g ${USER_GID} dev && \
+    useradd -m -s /bin/bash -u ${USER_UID} -g ${USER_GID} dev
 USER dev
 
 # Set env
diff --git a/Makefile b/Makefile
index a81aee9..1bb4348 100644
--- a/Makefile
+++ b/Makefile
@@ -13,8 +13,8 @@ DOTENV := $(shell grep -v '^\#' .env)
 # Development images
 DEV_IMAGE_TAG = $(TEST_CONTAINER_NAME)-dev
 DEV_IMAGE_ID = $(file < .image.dev)
-USER_UID=$(shell id -u)
-USER_GID=$(shell id -g)
+USER_UID:=$(shell id -u)
+USER_GID:=$(shell id -g)
 
 DOCKER ?= docker
 DOCKERCOMPOSE ?= docker compose
diff --git a/compose.yaml b/compose.yaml
index d88fa1c..505a12d 100644
--- a/compose.yaml
+++ b/compose.yaml
@@ -16,7 +16,7 @@ services:
         THRIFT_VERSION: $THRIFT_VERSION
         USER_UID: $USER_UID
         USER_GID: $USER_GID
-    user: "${USER_UID}:${USER_GID}"
+    user: "${USER_UID:-1000}:${USER_GID:-1000}"
     volumes:
       - .:$PWD
     hostname: dmt.default

From 8c5da58c635cf0d89e4c2cebe59e056f5764a126 Mon Sep 17 00:00:00 2001
From: Aleksey Kashapov <nanodirijabl@gmail.com>
Date: Tue, 14 Jul 2026 18:01:51 +0300
Subject: [PATCH 20/20] Bumps CI action to v2

---
 .github/workflows/erlang-checks.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/erlang-checks.yml b/.github/workflows/erlang-checks.yml
index 895c007..283f4d9 100644
--- a/.github/workflows/erlang-checks.yml
+++ b/.github/workflows/erlang-checks.yml
@@ -30,7 +30,7 @@ jobs:
   run:
     name: Run checks
     needs: setup
-    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@5b5b09ecfccfdf758d4daecdfa0c31cc9e4ed8bd
+    uses: valitydev/erlang-workflows/.github/workflows/erlang-parallel-build.yml@v2
     with:
       otp-version: ${{ needs.setup.outputs.otp-version }}
       rebar-version: ${{ needs.setup.outputs.rebar-version }}