Librarian

圖書館員

我會發現您腦海中的內容，但還沒有在那裡。

概述

簡單的CLI命令在您的文本文件上命令，您想以智能的方式搜索，並在沒有麻煩的軟件的情況下應用全文搜索技術。

作為CLI命令，可將其導入的抽象旨在重複使用您的需求。您可以將它們插入服務。

單文件實用程序。僅Python和Sqlite。

要求

1. Python$ ./configure --enable-loadable-sqlite-extensions -Python符合的SQLite二進製文件，並啟用了加載fts5擴展名。代碼中無需進一步的設置。
2. fts5stemmer.so編譯的共享庫，用於用您的優選語言索引目標文件。

用法

常見選擇

$ python librarian.py -h
usage: librarian.py [-h] [--db DB] [--table TABLE] [--debug] [--sql-trace] {index,match,update} ...

optional arguments:
  -h, --help            show this help message and exit
  --db DB               DB file path. (default: librarian.db)
  --table TABLE         Table name to store files content. (default: documents)
  --debug               Flag of print additional events. (default: False)
  --sql-trace           Flag of print sqlite statements. (default: False)

available commands:
  Use -h with each of them to get help.

  {index,match,update}
    index               Command to build a db and index. Have to be run once.
    match               Command to run query on indexed files.
    update              Command to check if content is changed and update in the database.

索引您的文本文件

$ python librarian.py index -h
usage: librarian.py index [-h] [--file-extensions string] [--language LANGUAGE] target

positional arguments:
  target                Directory or a file to build an index on.

optional arguments:
  -h, --help            show this help message and exit
  --file-extensions string
                        List of file extensions separated by space which to scan only. (default: frozenset({ ' .md ' }))
  --language LANGUAGE   list of available languages https://snowballstem.org/algorithms/ (default: russian)

搜尋

$ python librarian.py match -h
usage: librarian.py match [-h] [--limit LIMIT] [--fields field,...] [--format {raw,csv}] [--snippet 1, ' ' , ' ' , ' ' , ' ' ,10] query

positional arguments:
  query                 Sqlite query term executed by " MATCH " statement. Syntax can be found on https://sqlite.org/fts5.html#full_text_query_syntax.

optional arguments:
  -h, --help            show this help message and exit
  --limit LIMIT         Max count of results. (default: 5)
  --fields field,...    List of document fields to retrieve separated by comma, order is preserved. Choices: ( ' path ' , ' extension ' , ' size ' , ' created ' , ' modified ' , ' hash ' , ' rank ' , ' snippet ' , ' rowid ' ). (default:
                        ( ' path ' , ' snippet ' ))
  --format {raw,csv}    Choose a results output format. (default: csv)
  --snippet 1, ' ' , ' ' , ' ' , ' ' ,10
                        Snippet properties settings https://sqlite.org/fts5.html#the_snippet_function (default: None)

更新

$  python librarian.py update -h
usage: librarian.py update [-h] [--clean]

optional arguments:
  -h, --help  show this help message and exit
  --clean     Delete missing file records (default: False)

例子

讓我們建立一個DB。

$ python librarian.py --debug --sql-trace index ~ /Text-files-documents
SELECT fts5(NULL)
CREATE REAL TABLE IF NOT EXISTS documents 
                              USING FTS5(path, content, extension, sizee, created, modified, tokenize= ' snowball russian ' ) ;
Excluded: /home/user/Text-files-documents/Today
Excluded: /home/user/Text-files-documents/Yesterday
Indexed: /home/user/Text-files-documents/Project/README.md
...

並嘗試一下。

$ python librarian.py match --fields created,path тесты
[(1620583081.3579128,
 ' /home/user/Text-files-documents/Project/README.md ' ,
 ' Для запуска тестов нужно поднимать докер-сервисыn   nn### Описание запуска ' ),
 ...