/android-contentprovider-generator

A small tool to generate an Android ContentProvider.

Primary LanguageJavaGNU General Public License v3.0GPL-3.0

Android ContentProvider Generator

A tool to generate an Android ContentProvider. It takes a set of entity (a.k.a "table") definitions as the input, and generates:

  • a ContentProvider class
  • a SQLiteOpenHelper class
  • a SQLiteOpenHelperCallbacks class
  • one Columns class per entity
  • one Cursor class per entity
  • one ContentValues class per entity
  • one Selection class per entity
  • one Model interface per entity

How to use

The _config.json file

This is where you declare a few parameters that will be used to generate the code.

These are self-explanatory so here is an example:

{
	"syntaxVersion": 3,
	"projectPackageId": "com.example.app",
	"authority": "com.example.app.provider",
	"providerJavaPackage": "com.example.app.provider",
	"providerClassName": "ExampleProvider",
	"sqliteOpenHelperClassName": "ExampleSQLiteOpenHelper",
	"sqliteOpenHelperCallbacksClassName": "ExampleSQLiteOpenHelperCallbacks",
	"databaseFileName": "example.db",
	"databaseVersion": 1,
	"enableForeignKeys": true,
	"useAnnotations": true
}

Entity files

Create one file per entity, naming it <entity_name>.json. Inside each file, declare your fields (a.k.a "columns") with a name and a type. You can also optionally declare a default value, an index flag, a documentation and a nullable flag.

Currently the type can be:

  • String (SQLite type: TEXT)
  • Integer (INTEGER)
  • Long (INTEGER)
  • Float (REAL)
  • Double (REAL)
  • Boolean (INTEGER)
  • Date (INTEGER)
  • byte[] (BLOB)
  • enum (INTEGER).

You can also optionally declare table constraints.

Here is a person.json file as an example:

{
	"documentation": "A human being which is part of a team.",
	"fields": [
		{
			"documentation": "First name of this person. For instance, John.",
			"name": "first_name",
			"type": "String",
			"defaultValue": "John"
		},
		{
			"documentation": "Last name (a.k.a. Given name) of this person. For instance, Smith.",
			"name": "last_name",
			"type": "String",
			"nullable": true,
			"defaultValue": "Doe"
		},
		{
			"name": "age",
			"type": "Integer",
			"index": true
		},
		{
			"name": "gender",
			"type": "enum",
			"enumName": "Gender",
			"enumValues": [
				"MALE",
				"FEMALE",
				{"OTHER": "Value to use when neither male nor female"}
			],
			"nullable": false
		}
	],

	"constraints": [
		{
			"name": "unique_name",
			"definition": "UNIQUE (first_name, last_name) ON CONFLICT REPLACE"
		}
	]
}

Notes:

  • An _id primary key field is automatically (implicitly) declared for all entities. It must not be declared in the json file.
  • nullable is optional (true by default).
  • if documentation is present the value will be copied in Javadoc blocks in the generated code.

A more comprehensive sample is available in the etc/sample folder.

You can also have a look at the corresponding generated code in the etc/sample/app folder.

By convention, you should name your entities and fields in lower case with words separated by '_', like in the example above.

The header.txt file (optional)

If a header.txt file is present, its contents will be inserted at the top of every generated file.

Get the tool

Download the jar from here: https://github.com/BoD/android-contentprovider-generator/releases/latest

Run the tool

java -jar android_contentprovider_generator-1.9.3-bundle.jar -i <input folder> -o <output folder>

  • Input folder: where to find _config.json and your entity json files
  • Output folder: where the resulting files will be generated

Use the generated files

  • When querying a table, use the corresponding Selection class as shown in this example:
PersonSelection where = new PersonSelection();
where.firstName("John").or().age(42);
Cursor c = context.getContentResolver().query(PersonColumns.CONTENT_URI, projection,
        where.sel(), where.args(), null);
  • When using the results of a query, wrap the resulting Cursor in the corresponding wrapper class. You can then use the generated getters directly as shown in this example:
PersonCursor person = new PersonCursor(c);
String lastName = person.getLastName();
Long age = person.getAge();
  • You can also conveniently combine these two facilities by using the query (or delete) method:
PersonSelection where = new PersonSelection();
where.firstName("John").or().age(42).orderByFirstName();
PersonCursor person = where.query(getContentResolver());
person.moveToNext();
String lastName = person.getLastName();
Long age = person.getAge();
  • When updating or inserting into a table, use the corresponding ContentValues class as shown in this example:
PersonContentValues values = new PersonContentValues();
values.putFirstName("John").putAge(42);
context.getContentResolver().update(personUri, values.values(), null, null);

Advanced usage

Foreign key / joins

There is limited support for foreign keys and joins. Here is an example of the syntax:

{
	"fields": [
		{
			"name": "main_team_id",
			"type": "Long",
			"nullable": false,
			"foreignKey": {
				"table": "team",
				"onDelete": "CASCADE"
			}
		},
		{
			"name": "first_name",
			"type": "String",
			"nullable": false
		},

		(...)
}

In this example, the field main_team_id is a foreign key referencing the primary key of the team table.

  • The appropriate FOREIGN KEY SQL constraint is generated (if enableForeignKeys is set to true in _config.json).
  • The team table will be automatically joined when querying the person table (only if any team columns are included in the projection).
  • Getters for team columns are generated in the PersonCursor wrapper.
  • Of course if team has foreign keys they will also be handled (and recursively).

Limitations

  • Foreign keys always reference the _id column (the implicit primary key of all tables) and thus must always be of type Long - by design.
  • Only one foreign key to a particular table is allowed per table. In the example above only one column in person can point to team.
  • Loops (i.e. A has a foreign key to B and B has a foreign key to A) aren't detected. The generator will infinitely loop if they exist.
  • Cases such as "A has a FK to B, B has a FK to C, A has a FK to C" generate ambiguities in the queries, because C columns appear twice. In the sample app you can see an example of how to deal with this case, using prefixes and aliases (SQL's AS keyword).

Sample

A sample is available in the etc/sample folder.

You can have a look at the corresponding generated code in the etc/sample/app folder.

Here is the table shema of the sample: Table shema of the sample

Building

You need maven to build this tool.

mvn package

This will produce android_contentprovider_generator-1.9.3-bundle.jar in the target folder.

Similar tools

Here is a list of other tools that try to tackle the same problem.

I did not have the chance to try them out.

Licence

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.

Just to be absolutely clear, this license applies to this program itself, not to the source it will generate!