NAME

Plack::App::dbi2http - Export DBI database as HTTP API (Riap::HTTP)

VERSION

This document describes version 0.07 of Plack::App::dbi2http (from Perl distribution Plack-App-dbi2http), released on 2017-07-11.

SYNOPSIS

Prepare and edit a config file:

% cp share/sample-config/dbi2http.conf.yaml ~
% emacs ~/dbi2http.conf.yaml; # edit log path and supply DBI dsn/user/password

Run service:

% cd share/www
% plackup dbi2http.psgi
HTTP::Server::PSGI: Accepting connections at http://0:5000/

From another console, access HTTP API via, e.g. curl:

% curl http://localhost:5000/list_tables
countries
continents

% curl http://localhost:5000/list_columns?table=countries
id
ind_name
eng_name
tags

% curl 'http://localhost:5000/list_columns?table=countries&detail=1&-riap-fmt=json-pretty'
[
   200,
  "OK",
  [
     {
        "pos" : 1,
        "name" : "id",
        "type" : "text"
     },
     {
        "name" : "ind_name",
        "pos" : 2,
        "type" : "text"
     },
     {
        "type" : "text",
        "name" : "eng_name",
        "pos" : 3
     },
     {
        "pos" : 4,
        "name" : "tags",
        "type" : "text"
     }
  ]
]

% curl 'http://localhost:5000/list_rows?table=countries'
China   cn      Cina    panda
Indonesia       id      Indonesia       bali,tropical
Singapore       sg      Singapura       tropical
United States of America        us      Amerika Serikat

% curl 'http://localhost:5000/list_rows?table=countries&-riap-fmt=text-pretty'
.-----------------------------------------------------------------.
| eng_name                   id   ind_name          tags          |
|                                                                 |
| China                      cn   Cina              panda         |
| Indonesia                  id   Indonesia         bali,tropical |
| Singapore                  sg   Singapura         tropical      |
| United States of America   us   Amerika Serikat                 |
`-----------------------------------------------------------------'

Or use App::riap, a client shell for Riap (with filesystem-like API browsing and shell tab completion):

% riap http://localhost:5000/
riap /> ls
list_columns
list_rows
list_tables

riap /> list_tables
countries
continents

riap /> list_columns --table countries --detail

riap /> list_rows --table countries

DESCRIPTION

This module provides a sample Plack application, which you can customize, to export a DBI database as a HTTP API service.

I was reading Yanick's blog entry today, http://techblog.babyl.ca/entry/waack, titled Instant REST API for Any Databases and I thought I'd quickly cobble up something similar using a different toolbox. Granted, the resulting HTTP API is not REST (read: it's better :-) and at 0.01 the API functions are somewhat limited (DBIx::FunctionalAPI) but this demonstrates how easy it is to create something usable.

The tools and frameworks are: DBIx::FunctionalAPI which provides a set of functions: list_tables, list_columns, list_rows, create_table, create_row, rename_table, etc. These are normal Perl functions that accept dbh, table arguments and so on.

Next we have Perinci::Access::HTTP::Server, a set of Plack middlewares that let you access Perl functions over HTTP using the Riap::HTTP protocol. We compose the middlewares in a PSGI application called dbi2http.psgi.

All you need to do now is just run the PSGI application with Plack, using one of the many available PSGI servers. There is a configuration file required, to be put in the home directory, and can be copied from the provided sample config. All you need to set is basically just path to log file and the DBI connection information (db_dsn, db_user, db_password) and you're good to go.

After the PSGI application is running, you can connect using a plain HTTP client like curl. Riap::HTTP exposes Perl modules and functions directly as URL paths. For example, you can just request the root URL first, and a help message is returned:

% curl http://localhost:5000/
function        list_columns
function        list_rows
function        list_tables

Tips:
* To call a function, try:
    http://localhost:5000/api/list_tables
* Function arguments can be given via GET/POST params or JSON hash in req body
* To find out which arguments a function supports, try:
    http://localhost:5000/api/list_tables?-riap-action=meta
* To find out all available actions on an entity, try:
    http://localhost:5000/api/list_columns?-riap-action=actions
* This server uses Riap protocol for great autodiscoverability, for more info:
    https://metacpan.org/module/Riap

We can see there are 3 top-level functions available. Let's call the list_tables function to, well, list available tables.

% curl http://localhost:5000/list_tables
"main"."continents"
"main"."countries"
"main"."sqlite_master"
"temp"."sqlite_temp_master"
"main"."continents"
"main"."countries"

Functions usually return data structure and the PSGI application formats it as a presentable text to the client. To get the raw data, use the -riap-fmt special argument:

% curl http://localhost:5000/list_tables?-riap-fmt=json-pretty
[
  200,
  "OK",
  [
     "\"main\".\"continents\"",
     "\"main\".\"countries\"",
     "\"main\".\"sqlite_master\"",
     "\"temp\".\"sqlite_temp_master\"",
     "\"main\".\"continents\"",
     "\"main\".\"countries\""
  ]
]

Actually, Riap provides more than just RPC (function call). It can also expose function metadata so the protocol is super-self-discoverable. For example, instead of the default call action, let's call the meta action to request the function metadata:

% curl 'http://localhost:5000/list_tables?-riap-action=meta&-riap-fmt=json-pretty'
[
  200,
  "OK (meta action)",
  {
     "entity_v" : "0.01",
     "result_naked" : "0",
     "entity_date" : "2014-06-15",
     "features" : {},
     "summary": "List available tables",
     "x.perinci.sub.wrapper.logs" : [
        {
           "normalize_schema" : "1",
           "validate_args" : "1",
           "validate_result" : "1"
        }
     ],
     "args_as" : "hash",
     "v" : "1.1",
     "args" : {}
  }
]

Let's check the metadata of another function:

% curl 'http://localhost:5000/list_columns?-riap-action=meta&-riap-fmt=json-pretty'
[
  200,
  "OK (meta action)",
  {
     "summary": "List columns of a table",
     "args" : {
        "detail" : {
           "summary" : "Whether to return detailed records instead of just items/strings",
           "schema" : [
              "bool",
              {},
              {}
           ]
        },
        "table" : {
           "schema" : [
              "str",
              {
                 "req" : 1
              },
              {}
           ],
           "req" : 1,
           "summary" : "Table name"
        }
     },
     "args_as" : "hash",
     "v" : 1.1,
     "result_naked" : 0,
     "x.perinci.sub.wrapper.logs" : [
        {
           "validate_args" : "1",
           "validate_result" : "1",
           "normalize_schema" : "1"
        }
     ],
     "features" : {},
     "entity_date" : "2014-06-15",
     "entity_v" : "0.01"
  }
]

We can see from the above that the list_columns function accepts arguments table (a string, required) and detail (bool).

For the full specification of the metadata format, see Rinci.

Aside from using a low-level HTTP client, we can also use App::riap, a Riap client (just to note that Riap can also be accessed via transport protocol other than HTTP, but it's another subject matter). The client is a command-line shell with some conveniences like filesystem-like browsing of API tree, tab completion, debugging, command history, and others. Let's install the client and try it out:

% cpanm -n App::riap
% riap http://localhost:5000/
riap />

We are first presented with a prompt. Let's try listing what's available in the top-level directory:

riap /> ls -l
.-------------------------.
| type       uri          |
|                         |
| function   list_columns |
| function   list_rows    |
| function   list_tables  |
`-------------------------'

We see that there are three functions available. To call a function, you can run it like a running a program:

riap /> list_tables
.-----------------------------------------------------------.
| "main"."continents"           "temp"."sqlite_temp_master" |
| "main"."countries"            "main"."continents"         |
| "main"."sqlite_master"        "main"."countries"          |
`-----------------------------------------------------------'

Note that you can do completion on the command-line. You can even call with --help option like a normal program:

riap /> list_columns --help
Usage
  /list_columns --help (or -h, -?) [--verbose]
  /list_columns [options]
Options
  --[no]detail
    Whether to return detailed records instead of just items/strings.
  --help, -h, -?
    Display this help message.
  --table=s [str] (required)
    Table name.

This help message is generated from the metadata (so in the background, the client performs a meta request and converts it to a formatted help message).

To add HTTP authentication (or do any other customization), you can just add a Plack middleware to the PSGI application.

Last word, exporting a database as a public API service is usually not a good idea. In case you don't realize that ;-)

HOMEPAGE

Please visit the project's homepage at https://metacpan.org/release/Plack-App-dbi2http.

SOURCE

Source repository is at https://github.com/perlancar/perl-Plack-App-dbi2http.

BUGS

Please report any bugs or feature requests on the bugtracker website https://rt.cpan.org/Public/Dist/Display.html?Name=Plack-App-dbi2http

When submitting a bug or request, please include a test-file or a patch to an existing test-file that illustrates the bug or desired feature.

SEE ALSO

Rinci, Riap, Riap::HTTP, DBIx::FunctionalAPI, App::riap

AUTHOR

perlancar <perlancar@cpan.org>

COPYRIGHT AND LICENSE

This software is copyright (c) 2017, 2015, 2014 by perlancar@cpan.org.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.