server/search: refactor, extract parsing

22 files changed, 308 insertions, 243 deletions

diff --git a/server/szurubooru/api/comment_api.py b/server/szurubooru/api/comment_api.py
index e25ce60..6f5ee00 100644
--- a/server/szurubooru/api/comment_api.py
+++ b/server/szurubooru/api/comment_api.py
@@ -12,8 +12,8 @@ def _serialize(ctx, comment, **kwargs):
 class CommentListApi(BaseApi):
     def __init__(self):
         super().__init__()
-        self._search_executor = search.SearchExecutor(
-            search.CommentSearchConfig())
+        self._search_executor = search.Executor(
+            search.configs.CommentSearchConfig())
 
     def get(self, ctx):
         auth.verify_privilege(ctx.user, 'comments:list')
diff --git a/server/szurubooru/api/post_api.py b/server/szurubooru/api/post_api.py
index 1496952..c6a2713 100644
--- a/server/szurubooru/api/post_api.py
+++ b/server/szurubooru/api/post_api.py
@@ -12,7 +12,8 @@ def _serialize_post(ctx, post):
 class PostListApi(BaseApi):
     def __init__(self):
         super().__init__()
-        self._search_executor = search.SearchExecutor(search.PostSearchConfig())
+        self._search_executor = search.Executor(
+            search.configs.PostSearchConfig())
 
     def get(self, ctx):
         auth.verify_privilege(ctx.user, 'posts:list')
diff --git a/server/szurubooru/api/snapshot_api.py b/server/szurubooru/api/snapshot_api.py
index 9f16e48..3f830f9 100644
--- a/server/szurubooru/api/snapshot_api.py
+++ b/server/szurubooru/api/snapshot_api.py
@@ -5,7 +5,8 @@ from szurubooru.func import auth, snapshots
 class SnapshotListApi(BaseApi):
     def __init__(self):
         super().__init__()
-        self._search_executor = search.SearchExecutor(search.SnapshotSearchConfig())
+        self._search_executor = search.Executor(
+            search.configs.SnapshotSearchConfig())
 
     def get(self, ctx):
         auth.verify_privilege(ctx.user, 'snapshots:list')
diff --git a/server/szurubooru/api/tag_api.py b/server/szurubooru/api/tag_api.py
index 372db1a..c49ec1b 100644
--- a/server/szurubooru/api/tag_api.py
+++ b/server/szurubooru/api/tag_api.py
@@ -19,7 +19,8 @@ def _create_if_needed(tag_names, user):
 class TagListApi(BaseApi):
     def __init__(self):
         super().__init__()
-        self._search_executor = search.SearchExecutor(search.TagSearchConfig())
+        self._search_executor = search.Executor(
+            search.configs.TagSearchConfig())
 
     def get(self, ctx):
         auth.verify_privilege(ctx.user, 'tags:list')
diff --git a/server/szurubooru/api/user_api.py b/server/szurubooru/api/user_api.py
index b7d5bd0..6125fc5 100644
--- a/server/szurubooru/api/user_api.py
+++ b/server/szurubooru/api/user_api.py
@@ -12,7 +12,8 @@ def _serialize(ctx, user, **kwargs):
 class UserListApi(BaseApi):
     def __init__(self):
         super().__init__()
-        self._search_executor = search.SearchExecutor(search.UserSearchConfig())
+        self._search_executor = search.Executor(
+            search.configs.UserSearchConfig())
 
     def get(self, ctx):
         auth.verify_privilege(ctx.user, 'users:list')
diff --git a/server/szurubooru/search/__init__.py b/server/szurubooru/search/__init__.py
index d16e987..919475f 100644
--- a/server/szurubooru/search/__init__.py
+++ b/server/szurubooru/search/__init__.py
@@ -1,8 +1,2 @@
-''' Search parsers and services. '''
-
-from szurubooru.search.search_executor import SearchExecutor
-from szurubooru.search.user_search_config import UserSearchConfig
-from szurubooru.search.snapshot_search_config import SnapshotSearchConfig
-from szurubooru.search.tag_search_config import TagSearchConfig
-from szurubooru.search.comment_search_config import CommentSearchConfig
-from szurubooru.search.post_search_config import PostSearchConfig
+from szurubooru.search.executor import Executor
+import szurubooru.search.configs
diff --git a/server/szurubooru/search/configs/__init__.py b/server/szurubooru/search/configs/__init__.py
new file mode 100644
index 0000000..9f48e14
--- /dev/null
+++ b/server/szurubooru/search/configs/__init__.py
@@ -0,0 +1,5 @@
+from szurubooru.search.configs.user_search_config import UserSearchConfig
+from szurubooru.search.configs.snapshot_search_config import SnapshotSearchConfig
+from szurubooru.search.configs.tag_search_config import TagSearchConfig
+from szurubooru.search.configs.comment_search_config import CommentSearchConfig
+from szurubooru.search.configs.post_search_config import PostSearchConfig
diff --git a/server/szurubooru/search/base_search_config.py b/server/szurubooru/search/configs/base_search_config.py
index 3830768..2dda98f 100644
--- a/server/szurubooru/search/base_search_config.py
+++ b/server/szurubooru/search/configs/base_search_config.py
@@ -2,13 +2,14 @@ import sqlalchemy
 from szurubooru import db, errors
 from szurubooru.func import util
 from szurubooru.search import criteria
+from szurubooru.search import tokens
 
 def wildcard_transformer(value):
     return value.replace('*', '%')
 
 class BaseSearchConfig(object):
-    SORT_DESC = -1
-    SORT_ASC = 1
+    SORT_ASC = tokens.SortToken.SORT_ASC
+    SORT_DESC = tokens.SortToken.SORT_DESC
 
     def create_filter_query(self):
         raise NotImplementedError()
@@ -38,11 +39,11 @@ class BaseSearchConfig(object):
         Decorate SQLAlchemy filter on given column using supplied criterion.
         '''
         try:
-            if isinstance(criterion, criteria.PlainSearchCriterion):
+            if isinstance(criterion, criteria.PlainCriterion):
                 expr = column == int(criterion.value)
-            elif isinstance(criterion, criteria.ArraySearchCriterion):
+            elif isinstance(criterion, criteria.ArrayCriterion):
                 expr = column.in_(int(value) for value in criterion.values)
-            elif isinstance(criterion, criteria.RangedSearchCriterion):
+            elif isinstance(criterion, criteria.RangedCriterion):
                 assert criterion.min_value != '' \
                     or criterion.max_value != ''
                 if criterion.min_value != '' and criterion.max_value != '':
@@ -57,40 +58,45 @@ class BaseSearchConfig(object):
         except ValueError:
             raise errors.SearchError(
                 'Criterion value %r must be a number.' % (criterion,))
-        if criterion.negated:
-            expr = ~expr
         return expr
 
     @staticmethod
     def _create_num_filter(column):
-        return lambda query, criterion: query.filter(
-            BaseSearchConfig._apply_num_criterion_to_column(column, criterion))
+        def wrapper(query, criterion, negated):
+            expr = BaseSearchConfig._apply_num_criterion_to_column(
+                column, criterion)
+            if negated:
+                expr = ~expr
+            return query.filter(expr)
+        return wrapper
 
     @staticmethod
     def _apply_str_criterion_to_column(column, criterion, transformer):
         '''
         Decorate SQLAlchemy filter on given column using supplied criterion.
         '''
-        if isinstance(criterion, criteria.PlainSearchCriterion):
+        if isinstance(criterion, criteria.PlainCriterion):
             expr = column.ilike(transformer(criterion.value))
-        elif isinstance(criterion, criteria.ArraySearchCriterion):
+        elif isinstance(criterion, criteria.ArrayCriterion):
             expr = sqlalchemy.sql.false()
             for value in criterion.values:
                 expr = expr | column.ilike(transformer(value))
-        elif isinstance(criterion, criteria.RangedSearchCriterion):
+        elif isinstance(criterion, criteria.RangedCriterion):
             raise errors.SearchError(
                 'Composite token %r is invalid in this context.' % (criterion,))
         else:
             assert False
-        if criterion.negated:
-            expr = ~expr
         return expr
 
     @staticmethod
     def _create_str_filter(column, transformer=wildcard_transformer):
-        return lambda query, criterion: query.filter(
-            BaseSearchConfig._apply_str_criterion_to_column(
-                column, criterion, transformer))
+        def wrapper(query, criterion, negated):
+            expr = BaseSearchConfig._apply_str_criterion_to_column(
+                column, criterion, transformer)
+            if negated:
+                expr = ~expr
+            return query.filter(expr)
+        return wrapper
 
     @staticmethod
     def _apply_date_criterion_to_column(column, criterion):
@@ -98,15 +104,15 @@ class BaseSearchConfig(object):
         Decorate SQLAlchemy filter on given column using supplied criterion.
         Parse the datetime inside the criterion.
         '''
-        if isinstance(criterion, criteria.PlainSearchCriterion):
+        if isinstance(criterion, criteria.PlainCriterion):
             min_date, max_date = util.parse_time_range(criterion.value)
             expr = column.between(min_date, max_date)
-        elif isinstance(criterion, criteria.ArraySearchCriterion):
+        elif isinstance(criterion, criteria.ArrayCriterion):
             expr = sqlalchemy.sql.false()
             for value in criterion.values:
                 min_date, max_date = util.parse_time_range(value)
                 expr = expr | column.between(min_date, max_date)
-        elif isinstance(criterion, criteria.RangedSearchCriterion):
+        elif isinstance(criterion, criteria.RangedCriterion):
             assert criterion.min_value or criterion.max_value
             if criterion.min_value and criterion.max_value:
                 min_date = util.parse_time_range(criterion.min_value)[0]
@@ -120,14 +126,17 @@ class BaseSearchConfig(object):
                 expr = column <= max_date
         else:
             assert False
-        if criterion.negated:
-            expr = ~expr
         return expr
 
     @staticmethod
     def _create_date_filter(column):
-        return lambda query, criterion: query.filter(
-            BaseSearchConfig._apply_date_criterion_to_column(column, criterion))
+        def wrapper(query, criterion, negated):
+            expr = BaseSearchConfig._apply_date_criterion_to_column(
+                column, criterion)
+            if negated:
+                expr = ~expr
+            return query.filter(expr)
+        return wrapper
 
     @staticmethod
     def _create_subquery_filter(
@@ -137,12 +146,12 @@ class BaseSearchConfig(object):
             filter_factory,
             subquery_decorator=None):
         filter_func = filter_factory(filter_column)
-        def func(query, criterion):
+        def wrapper(query, criterion, negated):
             subquery = db.session.query(right_id_column.label('foreign_id'))
             if subquery_decorator:
                 subquery = subquery_decorator(subquery)
             subquery = subquery.options(sqlalchemy.orm.lazyload('*'))
-            subquery = filter_func(subquery, criterion)
+            subquery = filter_func(subquery, criterion, negated)
             subquery = subquery.subquery('t')
             return query.filter(left_id_column.in_(subquery))
-        return func
+        return wrapper
diff --git a/server/szurubooru/search/comment_search_config.py b/server/szurubooru/search/configs/comment_search_config.py
index 0322426..bc6eea7 100644
--- a/server/szurubooru/search/comment_search_config.py
+++ b/server/szurubooru/search/configs/comment_search_config.py
@@ -1,6 +1,6 @@
 from sqlalchemy.sql.expression import func
 from szurubooru import db
-from szurubooru.search.base_search_config import BaseSearchConfig
+from szurubooru.search.configs.base_search_config import BaseSearchConfig
 
 class CommentSearchConfig(BaseSearchConfig):
     def create_filter_query(self):
diff --git a/server/szurubooru/search/post_search_config.py b/server/szurubooru/search/configs/post_search_config.py
index 0797491..75ab71a 100644
--- a/server/szurubooru/search/post_search_config.py
+++ b/server/szurubooru/search/configs/post_search_config.py
@@ -2,14 +2,15 @@ from sqlalchemy.orm import subqueryload, lazyload, defer
 from sqlalchemy.sql.expression import func
 from szurubooru import db, errors
 from szurubooru.func import util
-from szurubooru.search.base_search_config import BaseSearchConfig
+from szurubooru.search.configs.base_search_config import BaseSearchConfig
 
 def _enum_transformer(available_values, value):
     try:
         return available_values[value.lower()]
     except KeyError:
-        raise errors.SearchError('Invalid value: %r. Possible values: %r.' % (
-            value, list(sorted(available_values.keys()))))
+        raise errors.SearchError(
+            'Invalid value: %r. Possible values: %r.' % (
+                value, list(sorted(available_values.keys()))))
 
 def _type_transformer(value):
     available_values = {
diff --git a/server/szurubooru/search/snapshot_search_config.py b/server/szurubooru/search/configs/snapshot_search_config.py
index 971863f..a99aee7 100644
--- a/server/szurubooru/search/snapshot_search_config.py
+++ b/server/szurubooru/search/configs/snapshot_search_config.py
@@ -1,5 +1,5 @@
 from szurubooru import db
-from szurubooru.search.base_search_config import BaseSearchConfig
+from szurubooru.search.configs.base_search_config import BaseSearchConfig
 
 class SnapshotSearchConfig(BaseSearchConfig):
     def create_filter_query(self):
diff --git a/server/szurubooru/search/tag_search_config.py b/server/szurubooru/search/configs/tag_search_config.py
index d9ed898..4422d87 100644
--- a/server/szurubooru/search/tag_search_config.py
+++ b/server/szurubooru/search/configs/tag_search_config.py
@@ -2,7 +2,7 @@ from sqlalchemy.orm import subqueryload
 from sqlalchemy.sql.expression import func
 from szurubooru import db
 from szurubooru.func import util
-from szurubooru.search.base_search_config import BaseSearchConfig
+from szurubooru.search.configs.base_search_config import BaseSearchConfig
 
 class TagSearchConfig(BaseSearchConfig):
     def create_filter_query(self):
diff --git a/server/szurubooru/search/user_search_config.py b/server/szurubooru/search/configs/user_search_config.py
index bc6927e..5d50be1 100644
--- a/server/szurubooru/search/user_search_config.py
+++ b/server/szurubooru/search/configs/user_search_config.py
@@ -1,6 +1,6 @@
 from sqlalchemy.sql.expression import func
 from szurubooru import db
-from szurubooru.search.base_search_config import BaseSearchConfig
+from szurubooru.search.configs.base_search_config import BaseSearchConfig
 
 class UserSearchConfig(BaseSearchConfig):
     ''' Executes searches related to the users. '''
diff --git a/server/szurubooru/search/criteria.py b/server/szurubooru/search/criteria.py
index 422966b..ac0e95c 100644
--- a/server/szurubooru/search/criteria.py
+++ b/server/szurubooru/search/criteria.py
@@ -1,23 +1,22 @@
-class _BaseSearchCriterion(object):
-    def __init__(self, original_text, negated):
+class _BaseCriterion(object):
+    def __init__(self, original_text):
         self.original_text = original_text
-        self.negated = negated
 
     def __repr__(self):
         return self.original_text
 
-class RangedSearchCriterion(_BaseSearchCriterion):
-    def __init__(self, original_text, negated, min_value, max_value):
-        super().__init__(original_text, negated)
+class RangedCriterion(_BaseCriterion):
+    def __init__(self, original_text, min_value, max_value):
+        super().__init__(original_text)
         self.min_value = min_value
         self.max_value = max_value
 
-class PlainSearchCriterion(_BaseSearchCriterion):
-    def __init__(self, original_text, negated, value):
-        super().__init__(original_text, negated)
+class PlainCriterion(_BaseCriterion):
+    def __init__(self, original_text, value):
+        super().__init__(original_text)
         self.value = value
 
-class ArraySearchCriterion(_BaseSearchCriterion):
-    def __init__(self, original_text, negated, values):
-        super().__init__(original_text, negated)
+class ArrayCriterion(_BaseCriterion):
+    def __init__(self, original_text, values):
+        super().__init__(original_text)
         self.values = values
diff --git a/server/szurubooru/search/executor.py b/server/szurubooru/search/executor.py
new file mode 100644
index 0000000..d4113be
--- /dev/null
+++ b/server/szurubooru/search/executor.py
@@ -0,0 +1,118 @@
+import sqlalchemy
+from szurubooru import db, errors
+from szurubooru.func import cache
+from szurubooru.search import tokens, parser
+
+def _format_dict_keys(source):
+    return list(sorted(source.keys()))
+
+def _get_direction(direction, default_direction):
+    if direction == tokens.SortToken.SORT_DEFAULT:
+        return default_direction
+    if direction == tokens.SortToken.SORT_NEGATED_DEFAULT:
+        if default_direction == tokens.SortToken.SORT_ASC:
+            return tokens.SortToken.SORT_DESC
+        elif default_direction == tokens.SortToken.SORT_DESC:
+            return tokens.SortToken.SORT_ASC
+        assert False
+    return direction
+
+class Executor(object):
+    '''
+    Class for search parsing and execution. Handles plaintext parsing and
+    delegates sqlalchemy filter decoration to SearchConfig instances.
+    '''
+
+    def __init__(self, search_config):
+        self.config = search_config
+        self.parser = parser.Parser()
+
+    def execute(self, query_text, page, page_size):
+        '''
+        Parse input and return tuple containing total record count and filtered
+        entities.
+        '''
+        key = (id(self.config), query_text, page, page_size)
+        if cache.has(key):
+            return cache.get(key)
+
+        search_query = self.parser.parse(query_text)
+
+        filter_query = self.config.create_filter_query()
+        filter_query = filter_query.options(sqlalchemy.orm.lazyload('*'))
+        filter_query = self._prepare_db_query(filter_query, search_query, True)
+        entities = filter_query \
+            .offset((page - 1) * page_size) \
+            .limit(page_size) \
+            .all()
+
+        count_query = self.config.create_count_query()
+        count_query = count_query.options(sqlalchemy.orm.lazyload('*'))
+        count_query = self._prepare_db_query(count_query, search_query, False)
+        count_statement = count_query \
+            .statement \
+            .with_only_columns([sqlalchemy.func.count()]) \
+            .order_by(None)
+        count = db.session.execute(count_statement).scalar()
+
+        ret = (count, entities)
+        cache.put(key, ret)
+        return ret
+
+    def execute_and_serialize(self, ctx, serializer):
+        query = ctx.get_param_as_string('query')
+        page = ctx.get_param_as_int('page', default=1, min=1)
+        page_size = ctx.get_param_as_int('pageSize', default=100, min=1, max=100)
+        count, entities = self.execute(query, page, page_size)
+        return {
+            'query': query,
+            'page': page,
+            'pageSize': page_size,
+            'total': count,
+            'results': [serializer(entity) for entity in entities],
+        }
+
+    def _prepare_db_query(self, db_query, search_query, use_sort):
+        ''' Parse input and return SQLAlchemy query. '''
+
+        for token in search_query.anonymous_tokens:
+            if not self.config.anonymous_filter:
+                raise errors.SearchError(
+                    'Anonymous tokens are not valid in this context.')
+            db_query = self.config.anonymous_filter(
+                db_query, token.criterion, token.negated)
+
+        for token in search_query.named_tokens:
+            if token.name not in self.config.named_filters:
+                raise errors.SearchError(
+                    'Unknown named token: %r. Available named tokens: %r.' % (
+                        token.name,
+                        _format_dict_keys(self.config.named_filters)))
+            db_query = self.config.named_filters[token.name](
+                db_query, token.criterion, token.negated)
+
+        for token in search_query.special_tokens:
+            if token.value not in self.config.special_filters:
+                raise errors.SearchError(
+                    'Unknown special token: %r. Available special tokens: %r.' % (
+                        token.value,
+                        _format_dict_keys(self.config.special_filters)))
+            db_query = self.config.special_filters[token.value](
+                db_query, token.negated)
+
+        if use_sort:
+            for token in search_query.sort_tokens:
+                if token.name not in self.config.sort_columns:
+                    raise errors.SearchError(
+                        'Unknown sort token: %r. Available sort tokens: %r.' % (
+                            token.name,
+                            _format_dict_keys(self.config.sort_columns)))
+                column, default_direction = self.config.sort_columns[token.name]
+                direction = _get_direction(token.direction, default_direction)
+                if direction == token.SORT_ASC:
+                    db_query = db_query.order_by(column.asc())
+                elif direction == token.SORT_DESC:
+                    db_query = db_query.order_by(column.desc())
+
+        db_query = self.config.finalize_query(db_query)
+        return db_query
diff --git a/server/szurubooru/search/parser.py b/server/szurubooru/search/parser.py
new file mode 100644
index 0000000..432f05c
--- /dev/null
+++ b/server/szurubooru/search/parser.py
@@ -0,0 +1,90 @@
+import re
+from szurubooru import errors
+from szurubooru.search import criteria, tokens
+
+def _create_criterion(original_value, value):
+    if '..' in value:
+        low, high = value.split('..', 1)
+        if not low and not high:
+            raise errors.SearchError('Empty ranged value')
+        return criteria.RangedCriterion(original_value, low, high)
+    if ',' in value:
+        return criteria.ArrayCriterion(
+            original_value, value.split(','))
+    return criteria.PlainCriterion(original_value, value)
+
+def _parse_anonymous(value, negated):
+    criterion = _create_criterion(value, value)
+    return tokens.AnonymousToken(criterion, negated)
+
+def _parse_named(key, value, negated):
+    original_value = value
+    if key.endswith('-min'):
+        key = key[:-4]
+        value += '..'
+    elif key.endswith('-max'):
+        key = key[:-4]
+        value = '..' + value
+    criterion = _create_criterion(original_value, value)
+    return tokens.NamedToken(key, criterion, negated)
+
+def _parse_special(value, negated):
+    return tokens.SpecialToken(value, negated)
+
+def _parse_sort(value, negated):
+    if value.count(',') == 0:
+        direction_str = None
+    elif value.count(',') == 1:
+        value, direction_str = value.split(',')
+    else:
+        raise errors.SearchError('Too many commas in sort style token.')
+    try:
+        direction = {
+            'asc': tokens.SortToken.SORT_ASC,
+            'desc': tokens.SortToken.SORT_DESC,
+            '': tokens.SortToken.SORT_DEFAULT,
+            None: tokens.SortToken.SORT_DEFAULT,
+        }[direction_str]
+    except KeyError:
+        raise errors.SearchError(
+            'Unknown search direction: %r.' % direction_str)
+    if negated:
+        direction = {
+            tokens.SortToken.SORT_ASC: tokens.SortToken.SORT_DESC,
+            tokens.SortToken.SORT_DESC: tokens.SortToken.SORT_ASC,
+            tokens.SortToken.SORT_DEFAULT: tokens.SortToken.SORT_NEGATED_DEFAULT,
+            tokens.SortToken.SORT_NEGATED_DEFAULT: tokens.SortToken.SORT_DEFAULT,
+        }[direction]
+    return tokens.SortToken(value, direction)
+
+class SearchQuery():
+    def __init__(self):
+        self.anonymous_tokens = []
+        self.named_tokens = []
+        self.special_tokens = []
+        self.sort_tokens = []
+
+class Parser(object):
+    def parse(self, query_text):
+        query = SearchQuery()
+        for chunk in re.split(r'\s+', (query_text or '').lower()):
+            if not chunk:
+                continue
+            negated = False
+            while chunk[0] == '-':
+                chunk = chunk[1:]
+                negated = not negated
+            if ':' in chunk and chunk[0] != ':':
+                key, value = chunk.split(':', 2)
+                if key == 'sort':
+                    query.sort_tokens.append(
+                        _parse_sort(value, negated))
+                elif key == 'special':
+                    query.special_tokens.append(
+                        _parse_special(value, negated))
+                else:
+                    query.named_tokens.append(
+                        _parse_named(key, value, negated))
+            else:
+                query.anonymous_tokens.append(_parse_anonymous(chunk, negated))
+        return query
diff --git a/server/szurubooru/search/search_executor.py b/server/szurubooru/search/search_executor.py
deleted file mode 100644
index d23b288..0000000
--- a/server/szurubooru/search/search_executor.py
+++ /dev/null
@@ -1,161 +0,0 @@
-import re
-import sqlalchemy
-from szurubooru import db, errors
-from szurubooru.func import cache
-from szurubooru.search import criteria
-
-class SearchExecutor(object):
-    '''
-    Class for search parsing and execution. Handles plaintext parsing and
-    delegates sqlalchemy filter decoration to SearchConfig instances.
-    '''
-
-    def __init__(self, search_config):
-        self.config = search_config
-
-    def execute(self, query_text, page, page_size):
-        '''
-        Parse input and return tuple containing total record count and filtered
-        entities.
-        '''
-        key = (id(self.config), query_text, page, page_size)
-        if cache.has(key):
-            return cache.get(key)
-        filter_query = self.config.create_filter_query()
-        filter_query = filter_query.options(sqlalchemy.orm.lazyload('*'))
-        filter_query = self._prepare(filter_query, query_text)
-        entities = filter_query \
-            .offset((page - 1) * page_size) \
-            .limit(page_size) \
-            .all()
-
-        count_query = self.config.create_count_query()
-        count_query = count_query.options(sqlalchemy.orm.lazyload('*'))
-        count_query = self._prepare(count_query, query_text)
-        count_statement = count_query \
-            .statement \
-            .with_only_columns([sqlalchemy.func.count()]) \
-            .order_by(None)
-        count = db.session.execute(count_statement).scalar()
-        ret = (count, entities)
-        cache.put(key, ret)
-        return ret
-
-    def execute_and_serialize(self, ctx, serializer):
-        query = ctx.get_param_as_string('query')
-        page = ctx.get_param_as_int('page', default=1, min=1)
-        page_size = ctx.get_param_as_int('pageSize', default=100, min=1, max=100)
-        count, entities = self.execute(query, page, page_size)
-        return {
-            'query': query,
-            'page': page,
-            'pageSize': page_size,
-            'total': count,
-            'results': [serializer(entity) for entity in entities],
-        }
-
-    def _prepare(self, query, query_text):
-        ''' Parse input and return SQLAlchemy query. '''
-        for token in re.split(r'\s+', (query_text or '').lower()):
-            if not token:
-                continue
-            negated = False
-            while token[0] == '-':
-                token = token[1:]
-                negated = not negated
-
-            if ':' in token and token[0] != ':':
-                key, value = token.split(':', 2)
-                query = self._handle_key_value(query, key, value, negated)
-            else:
-                query = self._handle_anonymous(
-                    query, self._create_criterion(token, negated))
-
-        query = self.config.finalize_query(query)
-        return query
-
-    def _handle_key_value(self, query, key, value, negated):
-        if key == 'sort':
-            return self._handle_sort(query, value, negated)
-        elif key == 'special':
-            return self._handle_special(query, value, negated)
-        else:
-            return self._handle_named(query, key, value, negated)
-
-    def _handle_anonymous(self, query, criterion):
-        if not self.config.anonymous_filter:
-            raise errors.SearchError(
-                'Anonymous tokens are not valid in this context.')
-        return self.config.anonymous_filter(query, criterion)
-
-    def _handle_named(self, query, key, value, negated):
-        if key.endswith('-min'):
-            key = key[:-4]
-            value += '..'
-        elif key.endswith('-max'):
-            key = key[:-4]
-            value = '..' + value
-        criterion = self._create_criterion(value, negated)
-        if key in self.config.named_filters:
-            return self.config.named_filters[key](query, criterion)
-        raise errors.SearchError(
-            'Unknown named token: %r. Available named tokens: %r.' % (
-                key, list(self.config.named_filters.keys())))
-
-    def _handle_special(self, query, value, negated):
-        if value in self.config.special_filters:
-            return self.config.special_filters[value](query, negated)
-        raise errors.SearchError(
-            'Unknown special token: %r. Available special tokens: %r.' % (
-                value, list(self.config.special_filters.keys())))
-
-    def _handle_sort(self, query, value, negated):
-        if value.count(',') == 0:
-            dir_str = None
-        elif value.count(',') == 1:
-            value, dir_str = value.split(',')
-        else:
-            raise errors.SearchError('Too many commas in sort style token.')
-
-        try:
-            column, default_sort = self.config.sort_columns[value]
-        except KeyError:
-            raise errors.SearchError(
-                'Unknown sort style: %r. Available sort styles: %r.' % (
-                    value, list(self.config.sort_columns.keys())))
-
-        sort_asc = self.config.SORT_ASC
-        sort_desc = self.config.SORT_DESC
-
-        try:
-            sort_map = {
-                'asc': sort_asc,
-                'desc': sort_desc,
-                '': default_sort,
-                None: default_sort,
-            }
-            sort = sort_map[dir_str]
-        except KeyError:
-            raise errors.SearchError('Unknown search direction: %r.' % dir_str)
-
-        if negated and sort:
-            sort = -sort
-
-        transform_map = {
-            sort_asc: lambda input: input.asc(),
-            sort_desc: lambda input: input.desc(),
-            None: lambda input: input,
-        }
-        transform = transform_map[sort]
-        return query.order_by(transform(column))
-
-    def _create_criterion(self, value, negated):
-        if '..' in value:
-            low, high = value.split('..', 1)
-            if not low and not high:
-                raise errors.SearchError('Empty ranged value')
-            return criteria.RangedSearchCriterion(value, negated, low, high)
-        if ',' in value:
-            return criteria.ArraySearchCriterion(
-                value, negated, value.split(','))
-        return criteria.PlainSearchCriterion(value, negated, value)
diff --git a/server/szurubooru/search/tokens.py b/server/szurubooru/search/tokens.py
new file mode 100644
index 0000000..1e1a803
--- /dev/null
+++ b/server/szurubooru/search/tokens.py
@@ -0,0 +1,24 @@
+class AnonymousToken(object):
+    def __init__(self, criterion, negated):
+        self.criterion = criterion
+        self.negated = negated
+
+class NamedToken(AnonymousToken):
+    def __init__(self, name, criterion, negated):
+        super().__init__(criterion, negated)
+        self.name = name
+
+class SortToken(object):
+    SORT_DESC = 'desc'
+    SORT_ASC = 'asc'
+    SORT_DEFAULT = 'default'
+    SORT_NEGATED_DEFAULT = 'negated default'
+
+    def __init__(self, name, direction):
+        self.name = name
+        self.direction = direction
+
+class SpecialToken(object):
+    def __init__(self, value, negated):
+        self.value = value
+        self.negated = negated
diff --git a/server/szurubooru/tests/search/test_comment_search_config.py b/server/szurubooru/tests/search/test_comment_search_config.py
index 235fcea..5a88a71 100644
--- a/server/szurubooru/tests/search/test_comment_search_config.py
+++ b/server/szurubooru/tests/search/test_comment_search_config.py
@@ -4,8 +4,7 @@ from szurubooru import db, errors, search
 
 @pytest.fixture
 def executor():
-    search_config = search.CommentSearchConfig()
-    return search.SearchExecutor(search_config)
+    return search.Executor(search.configs.CommentSearchConfig())
 
 @pytest.fixture
 def verify_unpaged(executor):
diff --git a/server/szurubooru/tests/search/test_post_search_config.py b/server/szurubooru/tests/search/test_post_search_config.py
index a3d1749..59a0530 100644
--- a/server/szurubooru/tests/search/test_post_search_config.py
+++ b/server/szurubooru/tests/search/test_post_search_config.py
@@ -38,7 +38,7 @@ def feature_factory(user_factory):
 
 @pytest.fixture
 def executor(user_factory):
-    return search.SearchExecutor(search.PostSearchConfig())
+    return search.Executor(search.configs.PostSearchConfig())
 
 @pytest.fixture
 def auth_executor(executor, user_factory):
diff --git a/server/szurubooru/tests/search/test_tag_search_config.py b/server/szurubooru/tests/search/test_tag_search_config.py
index 7582a57..5b3002f 100644
--- a/server/szurubooru/tests/search/test_tag_search_config.py
+++ b/server/szurubooru/tests/search/test_tag_search_config.py
@@ -4,8 +4,7 @@ from szurubooru import db, errors, search
 
 @pytest.fixture
 def executor():
-    search_config = search.TagSearchConfig()
-    return search.SearchExecutor(search_config)
+    return search.Executor(search.configs.TagSearchConfig())
 
 @pytest.fixture
 def verify_unpaged(executor):
@@ -321,19 +320,4 @@ def test_sort_by_category(
     tag2 = tag_factory(names=['t2'], category=cat2)
     tag3 = tag_factory(names=['t3'], category=cat1)
     db.session.add_all([tag1, tag2, tag3])
-    import sqlalchemy
-    from sqlalchemy.orm import joinedload
-    print('test', [tag.first_name for tag in db.session.query(db.Tag)
-        .join(db.TagCategory).options(
-            joinedload(db.Tag.names),
-            joinedload(db.Tag.category),
-            joinedload(db.Tag.suggestions).joinedload(db.Tag.names),
-            joinedload(db.Tag.implications).joinedload(db.Tag.names)
-        )
-        .options(sqlalchemy.orm.lazyload('*'))
-        .order_by(db.TagCategory.name.asc())
-        .order_by(db.Tag.first_name.asc())
-        .offset(0)
-        .limit(100)
-        .all()])
     verify_unpaged(input, expected_tag_names)
diff --git a/server/szurubooru/tests/search/test_user_search_config.py b/server/szurubooru/tests/search/test_user_search_config.py
index dfaecac..7a2c450 100644
--- a/server/szurubooru/tests/search/test_user_search_config.py
+++ b/server/szurubooru/tests/search/test_user_search_config.py
@@ -4,8 +4,7 @@ from szurubooru import db, errors, search
 
 @pytest.fixture
 def executor():
-    search_config = search.UserSearchConfig()
-    return search.SearchExecutor(search_config)
+    return search.Executor(search.configs.UserSearchConfig())
 
 @pytest.fixture
 def verify_unpaged(executor):