rroonga 12.0.2 → 12.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9e1afd17d5a479b6e5c472729d3548f3551b15cec0db30bbf46f8a43266d96a1
-  data.tar.gz: 05b218925dfbf29f4deb9d34c5ab6408cbcecdf11de588d2c8d45efaf0e28417
+  metadata.gz: a85769bf192cecc677daaaddb349b4728bdeb7d240d1ecb4fc9660495c5fa2de
+  data.tar.gz: 05375332c1e8f93a9e1fb1b584a9e92d8923af6848ffdef223702a17d32852a8
 SHA512:
-  metadata.gz: ce775711f0f90338164a3f5c3f5c7731b9f8d06e8563836f0adb7de325f471d43d9be5bb09a428f2598e801dca502564fbea3b9cb285bb38d8f7310cd22a7251
-  data.tar.gz: fe208b1bf0ee5e04fbccbe6220b2b0b44e80e9825c8b9aaab86017d7d87de8185a7a51eea026532b6d13fdcd310cab7e8484702514e5b256d6957f984107f2c2
+  metadata.gz: 3ec5e421bb75cdf1c944fcf36f70778f59113da9a8c4c20df7318fd56e21a1f00856fe9a9f308cab584183c32cbf055758b3e6bd52d415771c3f0c28080588bf
+  data.tar.gz: d7d630a97a5831323c70ececc208016d4a5a02b4ddd476e043038eca71e8b6d37410d01c7eac8d5eb71b0788d7fbcc915d76737dfa38c0f3db02f1cc92b9884f

data/doc/text/news.md

@@ -1,5 +1,21 @@
 # NEWS
 
+## 12.1.0: 2022-12-05 {#version-12-1-0}
+
+### Fixes
+
+  * [{Groonga::PatriciaTrie}] Fixed a bug that `Groonga::PatriciaTrie#scan` returned wrong offsets if there are many hit words. [GitHub#207][Patch by Naoya Murakami]
+
+### Thanks
+
+  * Naoya Murakami
+
+## 12.0.8: 2022-09-28 {#version-12-0-8}
+
+### Improvements
+
+  * Add support for creating a {Groonga::PatriciaTrie} without database.
+
 ## 12.0.2: 2022-04-04 {#version-12-0-2}
 
 ### Improvements

data/ext/groonga/rb-grn-patricia-trie.c

@@ -1,7 +1,7 @@
 /* -*- coding: utf-8; mode: C; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
 /* vim: set sts=4 sw=4 ts=8 noet: */
 /*
-  Copyright (C) 2009-2021  Sutou Kouhei <kou@clear-code.com>
+  Copyright (C) 2009-2022  Sutou Kouhei <kou@clear-code.com>
 
   This library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
@@ -129,11 +129,45 @@ VALUE rb_cGrnPatriciaTrie;
  *
  *       省略した場合はShortText型をキーとして使用する。この場合、
  *       4096バイトまで使用可能である。
+ *
+ *     @option options :key_size (nil) The key size in bytes or the
+ *       max key size in bytes. If `key_variable_size: true` is also
+ *       specified, the size is the max key size not the key size.
+ *
+ *       This is used only when the `:context` isn't associated with a
+ *       database. If `:context` is associated with a database, this
+ *       value is ignored.
+ *
+ *       @see :key_type
+ *       @see :key_variable_size
+ *       @since 12.0.8
+ *
+ *     @option options :key_variable_size (false) Whether the key is
+ *       variable size or not.
+ *
+ *       This is used only when the `:context` isn't associated with a
+ *       database. If `:context` is associated with a database, this
+ *       value is ignored.
+ *
+ *       @see :key_type
+ *       @see :key_size
+ *       @since 12.0.8
+ *
  *     @option options :value_type
  *       値の型を指定する。省略すると値のための領域を確保しない。
  *       値を保存したい場合は必ず指定すること。
  *
  *       参考: {Groonga::Type.new}
+ *
+ *     @option options :value_size (nil) The value size in bytes.
+ *
+ *       This is used only when the `:context` isn't associated with a
+ *       database. If `:context` is associated with a database, this
+ *       value is ignored.
+ *
+ *       @see :value_type
+ *       @since 12.0.8
+ *
  *     @option options :default_tokenizer
  *       {Groonga::IndexColumn} で使用するトークナイザを指定する。
  *       デフォルトでは何も設定されていないので、テーブルに
@@ -170,7 +204,10 @@ rb_grn_patricia_trie_s_create (int argc, VALUE *argv, VALUE klass)
     VALUE rb_table;
     VALUE options, rb_context, rb_name, rb_path, rb_persistent;
     VALUE rb_key_normalize, rb_key_with_sis, rb_key_type;
+    VALUE rb_key_size;
+    VALUE rb_key_variable_size;
     VALUE rb_value_type;
+    VALUE rb_value_size;
     VALUE rb_default_tokenizer;
     VALUE rb_token_filters;
     VALUE rb_sub_records;
@@ -186,7 +223,10 @@ rb_grn_patricia_trie_s_create (int argc, VALUE *argv, VALUE klass)
                         "key_normalize", &rb_key_normalize,
                         "key_with_sis", &rb_key_with_sis,
                         "key_type", &rb_key_type,
+                        "key_size", &rb_key_size,
+                        "key_variable_size", &rb_key_variable_size,
                         "value_type", &rb_value_type,
+                        "value_size", &rb_value_size,
                         "default_tokenizer", &rb_default_tokenizer,
                         "token_filters", &rb_token_filters,
                         "sub_records", &rb_sub_records,
@@ -227,10 +267,26 @@ rb_grn_patricia_trie_s_create (int argc, VALUE *argv, VALUE klass)
     if (RVAL2CBOOL(rb_sub_records))
         flags |= GRN_OBJ_WITH_SUBREC;
 
-    table = grn_table_create(context, name, name_size, path,
-                             flags, key_type, value_type);
-    if (!table)
-        rb_grn_context_check(context, rb_ary_new_from_values(argc, argv));
+    if (grn_ctx_db(context)) {
+        table = grn_table_create(context, name, name_size, path,
+                                 flags, key_type, value_type);
+        if (!table)
+            rb_grn_context_check(context, rb_ary_new_from_values(argc, argv));
+    } else {
+        unsigned int key_size = NUM2UINT(rb_key_size);
+        unsigned int value_size =
+            RB_NIL_P(rb_value_size) ? 0 : NUM2UINT(rb_value_size);
+        if (RVAL2CBOOL(rb_key_variable_size)) {
+            flags |= GRN_OBJ_KEY_VAR_SIZE;
+        }
+        table = (grn_obj *)grn_pat_create(context,
+                                          path,
+                                          key_size,
+                                          value_size,
+                                          flags);
+        if (!table)
+            rb_grn_context_check(context, rb_ary_new_from_values(argc, argv));
+    }
     rb_table = GRNOBJECT2RVAL(klass, context, table, GRN_TRUE);
 
     if (!NIL_P(rb_default_tokenizer))
@@ -389,10 +445,14 @@ rb_grn_patricia_trie_scan (VALUE self, VALUE rb_string)
     grn_pat_scan_hit hits[1024];
     const char *string;
     long string_length;
+    const char *original_string;
+    long rest_offset;
     grn_bool block_given;
 
     string = StringValuePtr(rb_string);
     string_length = RSTRING_LEN(rb_string);
+    original_string = string;
+    rest_offset = 0;
 
     rb_grn_table_key_support_deconstruct(SELF(self), &table, &context,
                                          NULL, NULL, NULL,
@@ -425,7 +485,7 @@ rb_grn_patricia_trie_scan (VALUE self, VALUE rb_string)
             matched_info = rb_ary_new_from_args(4,
                                                 record,
                                                 term,
-                                                UINT2NUM(hits[i].offset),
+                                                UINT2NUM(hits[i].offset + rest_offset),
                                                 UINT2NUM(hits[i].length));
             if (block_given) {
                 rb_yield(matched_info);
@@ -434,6 +494,7 @@ rb_grn_patricia_trie_scan (VALUE self, VALUE rb_string)
             }
             previous_offset = hits[i].offset;
         }
+        rest_offset = rest - original_string;
         string_length -= rest - string;
         string = rest;
     }

data/ext/groonga/rb-grn-table.c

@@ -712,15 +712,11 @@ rb_grn_table_get_columns (int argc, VALUE *argv, VALUE self)
 {
     grn_ctx *context = NULL;
     grn_obj *table;
-    grn_obj *columns;
-    grn_obj *key_type;
-    grn_rc rc;
+    grn_hash *columns;
     int n;
-    grn_table_cursor *cursor;
     VALUE rb_prefix, rb_columns;
     char *prefix = NULL;
     unsigned prefix_size = 0;
-    VALUE exception;
 
     rb_grn_table_deconstruct(SELF(self), &table, &context,
                              NULL, NULL,
@@ -734,23 +730,27 @@ rb_grn_table_get_columns (int argc, VALUE *argv, VALUE self)
         prefix_size = RSTRING_LEN(rb_prefix);
     }
 
-    key_type = grn_ctx_at(context, GRN_DB_SHORT_TEXT);
-    columns = grn_table_create(context, NULL, 0, NULL, GRN_TABLE_HASH_KEY,
-                               key_type, 0);
+    columns = grn_hash_create(context,
+                              NULL,
+                              sizeof(grn_id),
+                              0,
+                              GRN_OBJ_TABLE_HASH_KEY|GRN_HASH_TINY);
     rb_grn_context_check(context, self);
-    n = grn_table_columns(context, table, prefix, prefix_size, columns);
+    n = grn_table_columns(context,
+                          table,
+                          prefix,
+                          prefix_size,
+                          (grn_obj *)columns);
     rb_grn_context_check(context, self);
 
     rb_columns = rb_ary_new2(n);
     if (n == 0) {
-        grn_obj_unlink(context, columns);
+        grn_hash_close(context, columns);
         return rb_columns;
     }
 
-    cursor = grn_table_cursor_open(context, columns, NULL, 0, NULL, 0,
-                                   0, -1, GRN_CURSOR_ASCENDING);
-    rb_grn_context_check(context, self);
-    while (grn_table_cursor_next(context, cursor) != GRN_ID_NIL) {
+    VALUE exception = RUBY_Qnil;
+    GRN_HASH_EACH_BEGIN(context, columns, cursor, id) {
         void *key;
         grn_id *column_id;
         grn_obj *column;
@@ -758,14 +758,12 @@ rb_grn_table_get_columns (int argc, VALUE *argv, VALUE self)
         grn_user_data *user_data;
         grn_bool need_to_set_name = GRN_FALSE;
 
-        grn_table_cursor_get_key(context, cursor, &key);
+        grn_hash_cursor_get_key(context, cursor, &key);
         column_id = key;
         column = grn_ctx_at(context, *column_id);
         exception = rb_grn_context_to_exception(context, self);
-        if (!NIL_P(exception)) {
-            grn_table_cursor_close(context, cursor);
-            grn_obj_unlink(context, columns);
-            rb_exc_raise(exception);
+        if (!RB_NIL_P(exception)) {
+            break;
         }
 
         user_data = grn_obj_user_data(context, column);
@@ -784,13 +782,12 @@ rb_grn_table_get_columns (int argc, VALUE *argv, VALUE self)
         }
 
         rb_ary_push(rb_columns, rb_column);
+    } GRN_HASH_EACH_END(context, cursor);
+    grn_hash_close(context, columns);
+    if (!RB_NIL_P(exception)) {
+        rb_exc_raise(exception);
     }
-    rc = grn_table_cursor_close(context, cursor);
-    grn_obj_unlink(context, columns);
-    if (rc != GRN_SUCCESS) {
-        rb_grn_context_check(context, self);
-        rb_grn_rc_check(rc, self);
-    }
+    rb_grn_context_check(context, self);
 
     return rb_columns;
 }

data/ext/groonga/rb-grn.h

@@ -92,8 +92,8 @@ RB_GRN_BEGIN_DECLS
 #define RB_GRN_HAVE_FLOAT32 GRN_VERSION_OR_LATER(10, 0, 2)
 
 #define RB_GRN_MAJOR_VERSION 12
-#define RB_GRN_MINOR_VERSION 0
-#define RB_GRN_MICRO_VERSION 2
+#define RB_GRN_MINOR_VERSION 1
+#define RB_GRN_MICRO_VERSION 0
 
 #define RB_GRN_OBJECT(object) ((RbGrnObject *)(object))
 #define RB_GRN_NAMED_OBJECT(object) ((RbGrnNamedObject *)(object))

data/test/test-patricia-trie.rb

@@ -1,6 +1,4 @@
-# -*- coding: utf-8 -*-
-#
-# Copyright (C) 2009-2014  Kouhei Sutou <kou@clear-code.com>
+# Copyright (C) 2009-2022  Sutou Kouhei <kou@clear-code.com>
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Lesser General Public
@@ -156,6 +154,44 @@ class PatriciaTrieTest < Test::Unit::TestCase
                  words.scan('muTEki リンクの冒険 ミリバール アルパカ ガッ'))
   end
 
+  def test_scan_for_many_words
+    Groonga::Context.default_options = {:encoding => "utf-8"}
+    words = Groonga::PatriciaTrie.create(:key_type => "ShortText",
+                                         :key_normalize => true)
+    words.add("x")
+    dot = words.add(".")
+    longtext = ""
+    scanned = []
+
+    1025.times.each do |i|
+      longtext += "."
+      scanned.push([dot, ".", i, 1])
+    end
+    assert_equal(scanned,
+                 words.scan(longtext))
+  end
+
+  def test_scan_no_database
+    Groonga::Context.open(encoding: "utf-8") do |context|
+      Groonga::PatriciaTrie.create(context: context,
+                                   key_size: 4096,
+                                   key_variable_size: true) do |words|
+        words.add("リンク")
+        arupaka = words.add("アルパカ")
+        words.add("アルパカ(生物)")
+        adventure_of_link = words.add('リンクの冒険')
+        words.add('冒険')
+        words.add('ｶﾞｯ')
+        words.add('ＭＵＴＥＫＩ')
+        assert_equal([
+                       [adventure_of_link, "リンクの冒険", 7, 18],
+                       [arupaka, "アルパカ", 42, 12],
+                     ],
+                     words.scan('muTEki リンクの冒険 ミリバール アルパカ ガッ'))
+      end
+    end
+  end
+
   def test_tag_keys
     Groonga::Context.default_options = {:encoding => "utf-8"}
     words = Groonga::PatriciaTrie.create(:key_type => "ShortText",

data/test/test-remote.rb

@@ -1,4 +1,4 @@
-# Copyright (C) 2009-2019  Kouhei Sutou <kou@clear-code.com>
+# Copyright (C) 2009-2022  Sutou Kouhei <kou@clear-code.com>
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Lesser General Public
@@ -100,6 +100,7 @@ class RemoteTest < Test::Unit::TestCase
                    "command_version",
                    "default_command_version",
                    "max_command_version",
+                   "memory_map_size",
                    "n_jobs",
                    "n_queries",
                    "start_time",

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rroonga
 version: !ruby/object:Gem::Version
-  version: 12.0.2
+  version: 12.1.0
 platform: ruby
 authors:
 - Kouhei Sutou
@@ -9,10 +9,10 @@ authors:
 - daijiro
 - Yuto HAYAMIZU
 - SHIDARA Yoji
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-04-04 00:00:00.000000000 Z
+date: 2022-12-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: groonga-client
@@ -371,7 +371,7 @@ licenses:
 - LGPL-2.1
 metadata:
   msys2_mingw_dependencies: groonga>=12.0.2
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -387,8 +387,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.0.dev
-signing_key: 
+rubygems_version: 3.3.24
+signing_key:
 specification_version: 4
 summary: Ruby bindings for Groonga that provide full text search and column store
   features.