naoa

最終課題レポート Naoya Murakami

最終提出モデル

https://huggingface.co/naoa/gemma-2-9b-base71_lora

自己紹介

Patentfield という特許検索・分析サービス を開発しています。 情報検索、自然言語処理に興味関心を持っており、特許情報をより活用できるようにするため、

naoa

apt-get update -qq && apt-get install -y build-essential

apt-get -y install software-properties-common
add-apt-repository -y universe
add-apt-repository -y ppa:groonga/ppa
apt-get update
apt-get install -y mecab libmecab-dev mecab-naist-jdic
apt-get install -y debhelper autotools-dev pkg-config liblz4-dev zlib1g-dev libmsgpack-dev libzmq3-dev libevent-dev libpcre3-dev libstemmer-dev libjemalloc-dev libssl-dev

apt-get install -y git

naoa

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <groonga.h>
#include <time.h>
#include <limits.h>

grn_obj *
create_lexicon(grn_ctx *ctx, grn_obj *target_table, grn_obj *target_column,
               char *lexicon_name)

naoa

#include <stdio.h>
#include <string.h>
#include <groonga.h>

int main()
{
  grn_ctx ctx;
  grn_obj *db, *table;
  grn_id id;
  const char *path = "score.grn";

naoa

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <groonga.h>
#include <time.h>
#include <limits.h>

typedef struct {
  double score;
  int n_subrecs;

naoa

plugin_register functions/vector
[[0,0.0,0.0],true]
table_create Users TABLE_NO_KEY
[[0,0.0,0.0],true]
column_create Users names COLUMN_VECTOR ShortText
[[0,0.0,0.0],true]
table_create Words TABLE_PAT_KEY ShortText
[[0,0.0,0.0],true]
column_create Words users COLUMN_INDEX Users names
[[0,0.0,0.0],true]

naoa

table_create Tags TABLE_PAT_KEY ShortText
[[0,0.0,0.0],true]
table_create Memos TABLE_HASH_KEY ShortText
[[0,0.0,0.0],true]
column_create Memos tags COLUMN_VECTOR Tags
[[0,0.0,0.0],true]
load --table Memos
[
{"_key": "Rroonga", "tags": ["Groonga", "Ruby"]},
{"_key": "Groonga", "tags": ["Groonga"]},

naoa

% du -hc /share/all/post/base/36915/ #PGroonga 実際利用しているサイズ
518M    /share/all/post/base/36915/
518M    合計

% ls -sl /share/all/post/base/36915/ |awk '{i+=$1} END{print i/1024}' #スパースを考慮して実際利用しているサイズ
517.461

% ls -sl /share/all/post/base/36915/pgrn* |awk '{i+=$1} END{print i/1024}' # Groongaのみが実際に利用しているサイズ
425.984

naoa

<?php
  $db = "localhost";
  $db_name = $argv[1];
  $table = "text";

  $article = $argv[2];

  if ($handle = opendir($article)) {
    while (false !== ($file = readdir($handle))) {
      echo "-------$file------\n";

naoa

require "groonga"

database_path = ARGV[0]
table_name = ARGV[1]
index_column_name = ARGV[2]

Groonga::Database.open(database_path)
index = Groonga["#{table_name}.#{index_column_name}"]
terms = Groonga::Hash.create(:key_type => Groonga::Type::SHORT_TEXT,
                             :value_type => Groonga::Type::UINT32)