README

A general tool for extracting and cleaning selected tables from a database for use as fixtures. Copies a subset of tables from one database to another under the control of a json configuration file. The latter database can then be dumped to SQL for use as a fixture file.

Usage

Create the target database(s) you wish to fill. The tool will only output to existing databases. The content of the Destination Databases will be trashed.
Create config/credentials.json with the DB server configuration.
Create config/*.db.json files to define mappings for each required database
Run bin/dbsampler.php

Configuration formats

All config files live in the config subdirectory. Files cannot contain comments as the JSON format does not support this, but as a convention, fields called "comment" will be ignored where possible.

credentials.json

"driver": "pdo_mysql" is currently assumed but this may change in future.

MySQL

See config/credentials.dist.json:

{
  "driver": "pdo_mysql",
  "dbUser": "root",
  "dbPassword": "SOMEPASSWORD",
  "dbHost": "127.0.0.1"
}

If you need different source and dest servers, this becomes:

{
  "source": {
    "driver": "pdo_mysql",
    "dbUser": "root",
    "dbPassword": "SOMEPASSWORD",
    "dbHost": "sourceDB.example.com"
  },

  "dest" : {
    "driver": "pdo_mysql",
    "dbUser": "root",
    "dbPassword": "SOMEPASSWORD",
    "dbHost": "127.0.0.1"
  }
}

If you need to prepare the connections, add an initialSql stanza:

{
  "source": {
    "driver": "pdo_mysql",
    "dbUser": "root",
    "dbPassword": "SOMEPASSWORD",
    "dbHost": "sourceDB.example.com",
    "initialSql": [
      "SET NAMES UTF8"
    ]
  },
  "dest": {
    "driver": "pdo_mysql",
    "dbUser": "root",
    "dbPassword": "SOMEPASSWORD",
    "dbHost": "127.0.0.1",
    "initialSql": [
      "SET NAMES UTF8",
      "SET foreign_key_checks = 0"
    ]
  }
}

Sqlite

See config/credentials.dist.json:

{
    "driver": "pdo_sqlite",
    "directory": "..\/path\/to\/sqlite-dbs"
}

Paths are assumed to be relative to the config file unless they start with a '/'. Sqlite databases to be migrated are assumed to be *.sqlite files in this directory

dbname.db.json

{
  "name": "small-sqlite-test",          # Configuration name
  "sourceDb": "small-source",           # Name of the source DB
  "destDb": "small-dest",               # Name of the destination DB. This DB will get trashed
  "tables": {                           # A set of tables to be copied over. Each table is defined as "table": config
                                        # Every config stanza requires a sampler field. For now, look these up in 
                                        # \Quidco\DbSampler\MigrationConfigProcessor::$samplerMap
                                        # All other fields depend on the specific sampler being used; these should 
                                        # all be documented in their own class files in src/Sample
    "fruits": {
      "sampler": "matched",
      "constraints": {
        "name": [
          "apple",
          "pear"
        ]
      },
      "remember": {
        "id": "fruit_ids"               # Cross-referencing is supported by "remember" stanzas
                                        # These take the field name of which the values are to be remembered
                                        # matched to a variable name in which the values will be stored
                                        # Note: Variable declarations do not include a '$' symbol 
                                        # References MUST be 'remember'ed before being used, there is no
      }                                 # dependency resolution here, so order your config appropriately
    },
    "vegetables": {
      "sampler": "NewestById",
      "idField": "id",
      "quantity": 2
    },
    "fruit_x_basket": {
      "sampler": "matched",
      "constraints": {
        "fruit_id": "$fruit_ids"        # Remembered variables, with $ sign, can be used as cross-references
                                        # This will expand to all ids of the fruits table matched above
      },
      "where" : [
        "basket_id > 1"                 # The matched sampler can also accept a list of arbitrary WHERE clauses
      ],
      "remember": {
        "basket_id": "basket_ids"
      }
    },
    "baskets": {
      "sampler": "cleanMatched",        # some samplers support field cleaners that are defined in
                                        # \Quidco\DbSampler\FieldCleanerProvider::getCleanerByName
                                        # They modify or replace the content of the field that they are keyed to
      "constraints": {
        "id": "$basket_ids"
      },
      "cleanFields": {
        "name": "fakefullname"
      }
    }
  },
  "views": [                            # view support is experimental
    "some_view"                         # views are specified as name only but format may change
  ]                                     # The destination's CURRENT_USER() is used as the DEFINER for MySQL DBs
}

"Faker" cleaners

Any 'faker' (fzaninotto/faker) generator that does not require parameters can be used directly in the cleanFields stanza by using "name": "faker:GENERATOR", eg:

 "cleanFields": {
   "ip": "faker:ipv4"
 },

Extending the project

The tool is designed to be extended primarily by adding custom Samplers (which must implement \Quidco\DbSampler\SamplerInterface) and cleaners (documented in \Quidco\DbSampler\FieldCleanerProvider::getCleanerByName).

It is likely that a mechanism to register external cleaners and samplers will be provided.

Currently, only mysql and sqlite databases are supported, but this could also be extended.

maple-syrup-group / dbsampler

Maintainers

Details