Add fulltext search support

app/models.py

@@ -15,18 +15,29 @@
 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 
-from flask import Flask, url_for
-from flask_sqlalchemy import SQLAlchemy
-from flask_migrate import Migrate
-from urllib.parse import urlparse
-from app import app, gravatar
-from sqlalchemy.orm import validates
-from flask_user import login_required, UserManager, UserMixin, SQLAlchemyAdapter
 import enum, datetime
 
+from app import app, gravatar
+from urllib.parse import urlparse
+
+from flask import Flask, url_for
+from flask_sqlalchemy import SQLAlchemy, BaseQuery
+from flask_migrate import Migrate
+from flask_user import login_required, UserManager, UserMixin, SQLAlchemyAdapter
+from sqlalchemy.orm import validates
+from sqlalchemy_searchable import SearchQueryMixin
+from sqlalchemy_utils.types import TSVectorType
+from sqlalchemy_searchable import make_searchable
+
+
 # Initialise database
 db = SQLAlchemy(app)
 migrate = Migrate(app, db)
+make_searchable(db.metadata)
+
+
+class ArticleQuery(BaseQuery, SearchQueryMixin):
+    pass
 
 
 class UserRank(enum.Enum):
@@ -246,7 +257,7 @@ class PackageType(enum.Enum):
 class PackagePropertyKey(enum.Enum):
 	name          = "Name"
 	title         = "Title"
-	shortDesc     = "Short Description"
+	short_desc     = "Short Description"
 	desc          = "Description"
 	type          = "Type"
 	license       = "License"
@@ -343,19 +354,22 @@ class Dependency(db.Model):
 		return retval
 
 
-
 class Package(db.Model):
+	query_class  = ArticleQuery
+
 	id           = db.Column(db.Integer, primary_key=True)
 
 	# Basic details
 	author_id    = db.Column(db.Integer, db.ForeignKey("user.id"))
 	name         = db.Column(db.String(100), nullable=False)
-	title        = db.Column(db.String(100), nullable=False)
+	title        = db.Column(db.Unicode(100), nullable=False)
-	shortDesc    = db.Column(db.String(200), nullable=False)
+	short_desc   = db.Column(db.Unicode(200), nullable=False)
-	desc         = db.Column(db.Text, nullable=True)
+	desc         = db.Column(db.UnicodeText, nullable=True)
 	type         = db.Column(db.Enum(PackageType))
 	created_at   = db.Column(db.DateTime, nullable=False, default=datetime.datetime.utcnow)
 
+	search_vector = db.Column(TSVectorType("title", "short_desc", "desc"))
+
 	license_id   = db.Column(db.Integer, db.ForeignKey("license.id"), nullable=False, default=1)
 	license      = db.relationship("License", foreign_keys=[license_id])
 	media_license_id = db.Column(db.Integer, db.ForeignKey("license.id"), nullable=False, default=1)
@@ -409,7 +423,7 @@ class Package(db.Model):
 			"name": self.name,
 			"title": self.title,
 			"author": self.author.display_name,
-			"short_description": self.shortDesc,
+			"short_description": self.short_desc,
 			"type": self.type.toName(),
 			"release": self.getDownloadRelease(protonum).id if self.getDownloadRelease(protonum) is not None else None,
 			"thumbnail": (base_url + tnurl) if tnurl is not None else None,
@@ -422,7 +436,7 @@ class Package(db.Model):
 			"author": self.author.display_name,
 			"name": self.name,
 			"title": self.title,
-			"short_description": self.shortDesc,
+			"short_description": self.short_desc,
 			"desc": self.desc,
 			"type": self.type.toName(),
 			"created_at": self.created_at,

app/public/static/package_create.js

@@ -35,10 +35,10 @@ $(function() {
 				setField("#repo", result.repo || repoURL);
 				setField("#issueTracker", result.issueTracker);
 				setField("#desc", result.description);
-				setField("#shortDesc", result.short_description);
+				setField("#short_desc", result.short_description);
 				setField("#harddep_str", result.depends);
 				setField("#softdep_str", result.optional_depends);
-				setField("#shortDesc", result.short_description);
+				setField("#short_desc", result.short_description);
 				setField("#forums", result.forumId);
 				if (result.type && result.type.length > 2) {
 					$("#type").val(result.type);

app/public/static/package_edit.js

@@ -41,7 +41,7 @@ $(function() {
 		It's obvious that this adds something to Minetest,
 		there's no need to use phrases such as \"adds X to the game\".`
 
-	$("#shortDesc").on("change paste keyup", function() {
+	$("#short_desc").on("change paste keyup", function() {
 		var val = $(this).val().toLowerCase();
 		if (val.indexOf("minetest") >= 0 || val.indexOf("mod") >= 0 ||
 				val.indexOf("modpack") >= 0 || val.indexOf("mod pack") >= 0) {

app/querybuilder.py

@@ -40,7 +40,7 @@ class QueryBuilder:
 			query = query.filter(Package.type.in_(self.types))
 
 		if self.search:
-			query = query.filter(Package.title.ilike('%' + self.search + '%'))
+			query = query.search(self.search)
 
 		if self.random:
 			query = query.order_by(func.random())

app/templates/macros/packagegridtile.html

@@ -12,7 +12,7 @@
 			</h3>
 
 			<p>
-				{{ package.shortDesc }}
+				{{ package.short_desc }}
 			</p>
 
 

app/templates/packages/create_edit.html

@@ -49,7 +49,7 @@
 			{{ render_field(form.title, class_="pkg_meta col-sm-7") }}
 			{{ render_field(form.name, class_="pkg_meta col-sm-3") }}
 			</div>
-			{{ render_field(form.shortDesc, class_="pkg_meta") }}
+			{{ render_field(form.short_desc, class_="pkg_meta") }}
 			{{ render_multiselect_field(form.tags, class_="pkg_meta") }}
 			<div class="pkg_meta row">
 				{{ render_field(form.license, class_="not_txp col-sm-6") }}

app/templates/packages/editrequest_create_edit.html

@@ -17,7 +17,7 @@
 		{{ render_field(form.type) }}
 		{{ render_field(form.name) }}
 		{{ render_field(form.title) }}
-		{{ render_field(form.shortDesc) }}
+		{{ render_field(form.short_desc) }}
 		{{ render_field(form.desc) }}
 		{{ render_multiselect_field(form.tags) }}
 

app/templates/packages/view.html

@@ -19,7 +19,7 @@
 			</h1>
 
 			<p class="lead">
-				{{ package.shortDesc }}
+				{{ package.short_desc }}
 			</p>
 
 			<div class="row" style="margin-top: 2rem;">

app/views/packages/packages.py

@@ -171,7 +171,7 @@ def package_download_page(package):
 class PackageForm(FlaskForm):
 	name          = StringField("Name (Technical)", [InputRequired(), Length(1, 20), Regexp("^[a-z0-9_]", 0, "Lower case letters (a-z), digits (0-9), and underscores (_) only")])
 	title         = StringField("Title (Human-readable)", [InputRequired(), Length(3, 50)])
-	shortDesc     = StringField("Short Description (Plaintext)", [InputRequired(), Length(1,200)])
+	short_desc     = StringField("Short Description (Plaintext)", [InputRequired(), Length(1,200)])
 	desc          = TextAreaField("Long Description (Markdown)", [Optional(), Length(0,10000)])
 	type          = SelectField("Type", [InputRequired()], choices=PackageType.choices(), coerce=PackageType.coerce, default=PackageType.MOD)
 	license       = QuerySelectField("License", [InputRequired()], query_factory=lambda: License.query.order_by(db.asc(License.name)), get_pk=lambda a: a.id, get_label=lambda a: a.name)

migrations/versions/2f3c3597c78d_.py

@@ -0,0 +1,36 @@
+"""empty message
+
+Revision ID: 2f3c3597c78d
+Revises: 9ec17b558413
+Create Date: 2019-01-29 02:43:08.865695
+
+"""
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects import postgresql
+from sqlalchemy_utils.types import TSVectorType
+from sqlalchemy_searchable import sync_trigger
+
+# revision identifiers, used by Alembic.
+revision = '2f3c3597c78d'
+down_revision = '9ec17b558413'
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.alter_column('package', 'short_desc', nullable=False, new_column_name='short_desc')
+    op.add_column('package', sa.Column('search_vector', TSVectorType("title", "short_desc", "desc"), nullable=True))
+    op.create_index('ix_package_search_vector', 'package', ['search_vector'], unique=False, postgresql_using='gin')
+
+    conn = op.get_bind()
+    sync_trigger(conn, 'package', 'search_vector', ["title", "short_desc", "desc"])
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_index('ix_package_search_vector', table_name='package')
+    op.drop_column('package', 'search_vector')
+    # ### end Alembic commands ###

migrations/versions/7ff57806ffd5_.py

@@ -0,0 +1,249 @@
+"""empty message
+
+Revision ID: 7ff57806ffd5
+Revises: 2f3c3597c78d
+Create Date: 2019-01-29 02:57:50.279918
+
+"""
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects import postgresql
+
+# revision identifiers, used by Alembic.
+revision = '7ff57806ffd5'
+down_revision = '2f3c3597c78d'
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.execute("""
+    DROP TYPE IF EXISTS tsq_state CASCADE;
+
+CREATE TYPE tsq_state AS (
+    search_query text,
+    parentheses_stack int,
+    skip_for int,
+    current_token text,
+    current_index int,
+    current_char text,
+    previous_char text,
+    tokens text[]
+);
+
+CREATE OR REPLACE FUNCTION tsq_append_current_token(state tsq_state)
+RETURNS tsq_state AS $$
+BEGIN
+    IF state.current_token != '' THEN
+        state.tokens := array_append(state.tokens, state.current_token);
+        state.current_token := '';
+    END IF;
+    RETURN state;
+END;
+$$ LANGUAGE plpgsql IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_tokenize_character(state tsq_state)
+RETURNS tsq_state AS $$
+BEGIN
+    IF state.current_char = '(' THEN
+        state.tokens := array_append(state.tokens, '(');
+        state.parentheses_stack := state.parentheses_stack + 1;
+        state := tsq_append_current_token(state);
+    ELSIF state.current_char = ')' THEN
+        IF (state.parentheses_stack > 0 AND state.current_token != '') THEN
+            state := tsq_append_current_token(state);
+            state.tokens := array_append(state.tokens, ')');
+            state.parentheses_stack := state.parentheses_stack - 1;
+        END IF;
+    ELSIF state.current_char = '"' THEN
+        state.skip_for := position('"' IN substring(
+            state.search_query FROM state.current_index + 1
+        ));
+
+        IF state.skip_for > 1 THEN
+            state.tokens = array_append(
+                state.tokens,
+                substring(
+                    state.search_query
+                    FROM state.current_index FOR state.skip_for + 1
+                )
+            );
+        ELSIF state.skip_for = 0 THEN
+            state.current_token := state.current_token || state.current_char;
+        END IF;
+    ELSIF (
+        state.current_char = '-' AND
+        (state.current_index = 1 OR state.previous_char = ' ')
+    ) THEN
+        state.tokens := array_append(state.tokens, '-');
+    ELSIF state.current_char = ' ' THEN
+        state := tsq_append_current_token(state);
+        IF substring(
+            state.search_query FROM state.current_index FOR 4
+        ) = ' or ' THEN
+            state.skip_for := 2;
+
+            -- remove duplicate OR tokens
+            IF state.tokens[array_length(state.tokens, 1)] != ' | ' THEN
+                state.tokens := array_append(state.tokens, ' | ');
+            END IF;
+        END IF;
+    ELSE
+        state.current_token = state.current_token || state.current_char;
+    END IF;
+    RETURN state;
+END;
+$$ LANGUAGE plpgsql IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_tokenize(search_query text) RETURNS text[] AS $$
+DECLARE
+    state tsq_state;
+BEGIN
+    SELECT
+        search_query::text AS search_query,
+        0::int AS parentheses_stack,
+        0 AS skip_for,
+        ''::text AS current_token,
+        0 AS current_index,
+        ''::text AS current_char,
+        ''::text AS previous_char,
+        '{}'::text[] AS tokens
+    INTO state;
+
+    state.search_query := lower(trim(
+        regexp_replace(search_query, '""+', '""', 'g')
+    ));
+
+    FOR state.current_index IN (
+        SELECT generate_series(1, length(state.search_query))
+    ) LOOP
+        state.current_char := substring(
+            search_query FROM state.current_index FOR 1
+        );
+
+        IF state.skip_for > 0 THEN
+            state.skip_for := state.skip_for - 1;
+            CONTINUE;
+        END IF;
+
+        state := tsq_tokenize_character(state);
+        state.previous_char := state.current_char;
+    END LOOP;
+    state := tsq_append_current_token(state);
+
+    state.tokens := array_nremove(state.tokens, '(', -state.parentheses_stack);
+
+    RETURN state.tokens;
+END;
+$$ LANGUAGE plpgsql IMMUTABLE;
+
+
+-- Processes an array of text search tokens and returns a tsquery
+CREATE OR REPLACE FUNCTION tsq_process_tokens(config regconfig, tokens text[])
+RETURNS tsquery AS $$
+DECLARE
+    result_query text;
+    previous_value text;
+    value text;
+BEGIN
+    result_query := '';
+    FOREACH value IN ARRAY tokens LOOP
+        IF value = '"' THEN
+            CONTINUE;
+        END IF;
+
+        IF left(value, 1) = '"' AND right(value, 1) = '"' THEN
+            value := phraseto_tsquery(config, value);
+        ELSIF value NOT IN ('(', ' | ', ')', '-') THEN
+            value := quote_literal(value) || ':*';
+        END IF;
+
+        IF previous_value = '-' THEN
+            IF value = '(' THEN
+                value := '!' || value;
+            ELSE
+                value := '!(' || value || ')';
+            END IF;
+        END IF;
+
+        SELECT
+            CASE
+                WHEN result_query = '' THEN value
+                WHEN (
+                    previous_value IN ('!(', '(', ' | ') OR
+                    value IN (')', ' | ')
+                ) THEN result_query || value
+                ELSE result_query || ' & ' || value
+            END
+        INTO result_query;
+        previous_value := value;
+    END LOOP;
+
+    RETURN to_tsquery(config, result_query);
+END;
+$$ LANGUAGE plpgsql IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_process_tokens(tokens text[])
+RETURNS tsquery AS $$
+    SELECT tsq_process_tokens(get_current_ts_config(), tokens);
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_parse(config regconfig, search_query text)
+RETURNS tsquery AS $$
+    SELECT tsq_process_tokens(config, tsq_tokenize(search_query));
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_parse(config text, search_query text)
+RETURNS tsquery AS $$
+    SELECT tsq_parse(config::regconfig, search_query);
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION tsq_parse(search_query text) RETURNS tsquery AS $$
+    SELECT tsq_parse(get_current_ts_config(), search_query);
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+-- remove first N elements equal to the given value from the array (array
+-- must be one-dimensional)
+--
+-- If negative value is given as the third argument the removal of elements
+-- starts from the last array element.
+CREATE OR REPLACE FUNCTION array_nremove(anyarray, anyelement, int)
+RETURNS ANYARRAY AS $$
+    WITH replaced_positions AS (
+        SELECT UNNEST(
+            CASE
+            WHEN $2 IS NULL THEN
+                '{}'::int[]
+            WHEN $3 > 0 THEN
+                (array_positions($1, $2))[1:$3]
+            WHEN $3 < 0 THEN
+                (array_positions($1, $2))[
+                    (cardinality(array_positions($1, $2)) + $3 + 1):
+                ]
+            ELSE
+                '{}'::int[]
+            END
+        ) AS position
+    )
+    SELECT COALESCE((
+        SELECT array_agg(value)
+        FROM unnest($1) WITH ORDINALITY AS t(value, index)
+        WHERE index NOT IN (SELECT position FROM replaced_positions)
+    ), $1[1:0]);
+$$ LANGUAGE SQL IMMUTABLE;
+""")
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    pass
+    # ### end Alembic commands ###

migrations/versions/83622276d439_.py

@@ -66,7 +66,7 @@ def upgrade():
     sa.Column('author_id', sa.Integer(), nullable=True),
     sa.Column('name', sa.String(length=100), nullable=False),
     sa.Column('title', sa.String(length=100), nullable=False),
-    sa.Column('shortDesc', sa.String(length=200), nullable=False),
+    sa.Column('short_desc', sa.String(length=200), nullable=False),
     sa.Column('desc', sa.Text(), nullable=True),
     sa.Column('type', sa.Enum('MOD', 'GAME', 'TXP', name='packagetype'), nullable=True),
     sa.Column('license_id', sa.Integer(), nullable=True),
@@ -141,7 +141,7 @@ def upgrade():
     op.create_table('edit_request_change',
     sa.Column('id', sa.Integer(), nullable=False),
     sa.Column('request_id', sa.Integer(), nullable=True),
-    sa.Column('key', sa.Enum('name', 'title', 'shortDesc', 'desc', 'type', 'license', 'tags', 'repo', 'website', 'issueTracker', 'forums', name='packagepropertykey'), nullable=False),
+    sa.Column('key', sa.Enum('name', 'title', 'short_desc', 'desc', 'type', 'license', 'tags', 'repo', 'website', 'issueTracker', 'forums', name='packagepropertykey'), nullable=False),
     sa.Column('oldValue', sa.Text(), nullable=True),
     sa.Column('newValue', sa.Text(), nullable=True),
     sa.ForeignKeyConstraint(['request_id'], ['edit_request.id'], ),

requirements.txt

@@ -8,6 +8,7 @@ Flask-Migrate~=2.3
 Flask-SQLAlchemy~=2.3
 Flask-User~=0.6
 GitHub-Flask~=3.2
+SQLAlchemy-Searchable==1.0.3
 
 beautifulsoup4~=4.6
 celery~=4.2

setup.py

@@ -55,7 +55,7 @@ def defineDummyData(licenses, tags, ruben):
 	mod.repo = "https://github.com/ezhh/other_worlds"
 	mod.issueTracker = "https://github.com/ezhh/other_worlds/issues"
 	mod.forums = 16015
-	mod.shortDesc = "The content library should not be used yet as it is still in alpha"
+	mod.short_desc = "The content library should not be used yet as it is still in alpha"
 	mod.desc = "This is the long desc"
 	db.session.add(mod)
 
@@ -77,7 +77,7 @@ def defineDummyData(licenses, tags, ruben):
 	mod1.repo = "https://github.com/rubenwardy/awards"
 	mod1.issueTracker = "https://github.com/rubenwardy/awards/issues"
 	mod1.forums = 4870
-	mod1.shortDesc = "Adds achievements and an API to register new ones."
+	mod1.short_desc = "Adds achievements and an API to register new ones."
 	mod1.desc = """
 Majority of awards are back ported from Calinou's old fork in Carbone, under same license.
 
@@ -112,7 +112,7 @@ awards.register_achievement("award_mesefind",{
 	mod2.repo = "https://github.com/minetest-mods/mesecons/"
 	mod2.issueTracker = "https://github.com/minetest-mods/mesecons/issues"
 	mod2.forums = 628
-	mod2.shortDesc = "Mesecons adds everything digital, from all kinds of sensors, switches, solar panels, detectors, pistons, lamps, sound blocks to advanced digital circuitry like logic gates and programmable blocks."
+	mod2.short_desc = "Mesecons adds everything digital, from all kinds of sensors, switches, solar panels, detectors, pistons, lamps, sound blocks to advanced digital circuitry like logic gates and programmable blocks."
 	mod2.desc = """
     ########################################################################
     ##  __    __   _____   _____   _____   _____   _____   _   _   _____  ##
@@ -210,7 +210,7 @@ No warranty is provided, express or implied, for any part of the project.
 	mod.repo = "https://github.com/ezhh/handholds"
 	mod.issueTracker = "https://github.com/ezhh/handholds/issues"
 	mod.forums = 17069
-	mod.shortDesc = "Adds hand holds and climbing thingies"
+	mod.short_desc = "Adds hand holds and climbing thingies"
 	mod.desc = "This is the long desc"
 	db.session.add(mod)
 
@@ -233,7 +233,7 @@ No warranty is provided, express or implied, for any part of the project.
 	mod.repo = "https://github.com/ezhh/other_worlds"
 	mod.issueTracker = "https://github.com/ezhh/other_worlds/issues"
 	mod.forums = 16015
-	mod.shortDesc = "Adds space with asteroids and comets"
+	mod.short_desc = "Adds space with asteroids and comets"
 	mod.desc = "This is the long desc"
 	db.session.add(mod)
 
@@ -248,7 +248,7 @@ No warranty is provided, express or implied, for any part of the project.
 	mod.repo = "https://github.com/rubenwardy/food/"
 	mod.issueTracker = "https://github.com/rubenwardy/food/issues/"
 	mod.forums = 2960
-	mod.shortDesc = "Adds lots of food and an API to manage ingredients"
+	mod.short_desc = "Adds lots of food and an API to manage ingredients"
 	mod.desc = "This is the long desc"
 	food = mod
 	db.session.add(mod)
@@ -264,7 +264,7 @@ No warranty is provided, express or implied, for any part of the project.
 	mod.repo = "https://github.com/rubenwardy/food_sweet/"
 	mod.issueTracker = "https://github.com/rubenwardy/food_sweet/issues/"
 	mod.forums = 9039
-	mod.shortDesc = "Adds sweet food"
+	mod.short_desc = "Adds sweet food"
 	mod.desc = "This is the long desc"
 	food_sweet = mod
 	db.session.add(mod)
@@ -282,7 +282,7 @@ No warranty is provided, express or implied, for any part of the project.
 	game1.repo = "https://github.com/rubenwardy/capturetheflag"
 	game1.issueTracker = "https://github.com/rubenwardy/capturetheflag/issues"
 	game1.forums = 12835
-	game1.shortDesc = "Two teams battle to snatch and return the enemy's flag, before the enemy takes their own!"
+	game1.short_desc = "Two teams battle to snatch and return the enemy's flag, before the enemy takes their own!"
 	game1.desc = """
 As seen on the Capture the Flag server (minetest.rubenwardy.com:30000)
 
@@ -307,7 +307,7 @@ Uses the CTF PvP Engine.
 	mod.type = PackageType.TXP
 	mod.author = ruben
 	mod.forums = 14132
-	mod.shortDesc = "This is an update of the original PixelBOX texture pack by the brillant artist Gambit"
+	mod.short_desc = "This is an update of the original PixelBOX texture pack by the brillant artist Gambit"
 	mod.desc = "This is the long desc"
 	db.session.add(mod)