LWOS contains an application called serve_db which provides database services to clients. serve_db stores databases in folders in the file system. The database tables are stored in files inside the database folders. The table file contains a map with two keys: "columns" and "rows". "columns" is a list of column names of the table and "rows" is a list of the rows of the table. Each row is also a list of column values.

For example, suppose that the following is the current state of LWOS:

• Hostname: "b.com"
• Command: "serve_db"
• Command Context:
  • Home: 1 (/dbs)
  • Database: 2 (/dbs/baseball)
• File System
  • Data:
    • Index: 0
      • Path: '/'
      • Folders: dbs: 1
    • Index: 1
      • Path: '/dbs'
      • Folders: baseball: 2
    • Index: 2
      • Path: '/dbs/baseball'
      • Files:
        • players:
          • 'columns': ['id', 'given_name', 'family_name']
          • 'rows':
            • [1, "Mike", "Trout", ]
            • [2, "Max", "Scherzer", ]
        • stats:
          • 'columns': ['player_id', 'team', 'year', 'home_runs']
          • 'rows':
            • [1, 'Angels', 2015, 40]
            • [1, 'Angels', 2016, 32]
            • [1, 'Angels', 2017, 51]
            • [2, 'Astros', 2015, 3]
            • [2, 'Rangers', 2016, 5]
      • Folders: -

/dbs contains 1 folder (database) called "baseball." And the baseball folder contains 2 files (tables) called players and stats. The table files are formatted in JSON.

The column names of the players table are:

[ "id", "given_name", "family_name" ]

And the table rows are:

[
    [ 1, "Mike", "Trout" ],
    [ 2, "Max", "Scherzer" ]
]

In the first row, the id is 1, the given_name is Mike, and family_name is Trout.

The column names of the stats table are:

[ "player_id", "team", "year", "home_runs" ]

And the table rows are:

[
    [ 1, "Angels", 2015, 40 ]
    [ 1, "Angels", 2016, 32 ]
    [ 1, "Angels", 2017, 51 ]
    [ 2, "Astros", 2015, 3 ]
    [ 2, "Rangers", 2016, 5 ]
]

When serve_db receives a query, the server loads the referenced table file and iterates over the table rows. If the table was updated during query processing, serve_db writes the table file back to the file system.

Querying a Database Inside a Program

Python programs can interact with serve_db by sending requests and receiving responses over the network. As with any network request, Python programs can use the send_message function to send database queries to serve_db. To receive a response from serve_db, Python programs can use the receive_message function and parse the response. The response is in JSON, so we can call the json_loads function to convert the message to a dictionary object. Alternatively, Logicwalk Python provides a function called receive_json_response which receives the message and converts it into a dictionary for you.

Here is an example Python client:

send_message('server.com', 'USE baseball')
response = receive_json_response()
if response['status'] == 'ok':
    send_message('server.com', 'SELECT family_name FROM players WHERE id="2"')
    response = receive_json_response()
    message = response['object']
    if message['status'] == 'ok':
        for row in message['rows']:
            print(row)
else:
    print("use command failed")

We will examine the serve_db command and the receive_json_response function in greater detail later in the chapter.