## DataShift Provides tools to shift data between Ruby, external applications, files and ActiveRecord. ### Features Map Active Record models and associations to files, generate sample templates. Import and Export ActiveRecord models through .xls or CSV files, including all associations and with configurable defaults. Create, parse and use Excel/OpenOffice (.xls) documents dynamically from Ruby. High level rake tasks provided for key features. Easily extendable Loader functionality to deal with non trivial import cases, such as complex association lookups. Specific loaders and rake tasks provided out the box for Spree E-Commerce, enabling import/export of all data including Products with complex associations such as Options/Variants and Images. Many example Spreadsheets provided, fully documented with comments for each column. ## Installation Add gem 'datashift' to your Gemfile/bundle, or install the latest gem as usual : `gem install datashift` To use : gem 'datashift' require 'datashift' To pull the tasks in, add this call to your Rakefile : DataShift::load_tasks N.B - To use the Excel loader, OLE and Excel are NOT required, however JRuby is required, since it uses Java's Apache POI under the hood to process .xls files. Guards are provided, and used internally, for mixed Ruby setups. Can be used like : if(DataShift::Guards::jruby? ) ..do something with Excel else ..do something with CSV end ## Active Record - Import/Export Provides high level rake tasks for importing data via ActiveRecord models into a DB, from various sources, currently csv or .xls files (Excel/Open Office) jruby -S rake datashift:import:excel model=BlogPost input=BlogPostImport.xls verbose=true Provides high level rake tasks for exporting data to various sources, currently .xls files (Excel/Open Office) jruby -S rake datashift:export:excel model=BlogPost result=BlogExport.xls The library can be easily extended with Loaders to deal with non trivial cases required to exchange data between such files and any active record supported database. Spree loaders are an example, these illustrate over riding processing for specific columns with complicated lookup requirements. A core feature of DataShift is the MethodDictionary and MethodMapper, which provides features for collecting reflection information from ActiveRecord models (all different associations, including join tables with many-to-many relationships). A full picture of all possible operations on a class can be created very easily, for example ona Blog model : MethodDictionary.find_operators( Blog ) This then allows Import/Export to be achieved, by mapping the file's header and column data to the found operators i.e. the methods to set data on model's attributes and associations. Here we retrieve the method details for a column name from a file, "Blog Date" MethodDictionary.find_method_detail( Blog, "Blog Date" ) Loaders can use this method lookup functionality, to find the correct association for a column heading, and populate AR object with row data. This means data can be mapped to any model without any further coding. Generators are also supplied to export a models attributes and associations to files, thus providing template spreadsheets that any user can fill out. MethodMapper also stores column type information so the raw file data can be provided as is, and whenever possible, under the bonnet the data will be cast to correct DB type. Here we show how a column name from a file, "Blog Date", can be mapped to Assign a stringified date, to the blog_date column, on a new Blog object : MethodDictionary.find_method_detail( Blog, "Blog Date" ).assign( Blog.new, "Sat Jul 23 2011" ) Because it's based on reflection against the model, can build complex relationships between tables during import/export, and extending data files with new columns need not require any additional Ruby coding. New columns can simply be added to the Excel/Open Office spreadsheet, or CSV file, setting the new attribute or association name in the header row. The Loader attempts to handle various human read-able forms of column names. For example, given an association on the model called, product_properties, will successfully load from columns with headings such as 'product_properties', 'Product Properties', 'ProductProperties' 'product properties' etc For has_many associations, either multiple columns can be used, or multiple values can be specified in a single column using suitable delimiters. Modular - so complex associations/mappings that require non generic lookups, can be handled by extending the loader engine. Original focus was on support for the Open Source Spree e-commerce project, so includes specific loaders and rake tasks for loading Spree Products, and associated data such as Product Variants, and Images. ## Template Generation and Export Template generation tasks can be used to export a model's definition as column headings to CSV or .xls. These can be provided to developers or business users, as a template for data collection and then loading. Export tasks can be used to export of a model's definition and any existing data stored in the database. This data can be exported directly to CSV or Excel/OpenOffice spreadsheets. ## Example Spreadsheets A number of example Spreadsheets with headers and comments, can be found in the spec/fixtures directory. Extensive Spree samples - including .xls and csv versions for simple Products or complex Products with multiple taxons, variants properties etc - can be found in the spec/fixtures/spree subdirectory. Column headings contain comments with full descriptions and instructions on syntax. ## Features - *High level wrappers around applications including Excel and Word Quickly and easily access common enterprise applications through Ruby Wrapper around MS Excel File format, acheived via Apache POI under JRuby, so not restricted to Windows and Excel does not need to be installed. The required POI jars are already included. - *Direct Excel export* Excel/OpenOffice spreadsheets are heavily used in many sectors, so direct support makes it easier and quicker to migrate your client's data into a Rails/ActiveRecord project. No need to save to CSV or map to YAML. - *Semi-Smart Name Lookup* Includes helper classes that find and store details of all possible associations on an AR class. Given a user supplied name, attempts to find the requested association. Example usage, load from a file or spreadsheet where the column names are only an approximation of the actual associations, so given 'Product Properties' heading, finds real association 'product_properties' to send or call on the AR object - *Associations* Can handle 'belongs_to, 'has_many' and 'has_one' associations, including assignment of multiple objects via either multiple columns, or via specially delimited entry in a single (column). See Details section. Supports delegated attributes. - *Rake Tasks* High level Rake tasks are provided, only required to supply model class, and file location : jruby -S rake datashift:import:excel model=MusicTrack input=MyTrackListing.xls - *Spree Rake Tasks* Spree's product associations are non trivial so specific Rake tasks are also provided for loading Spree Producta with all associations and Image loading. jruby -S rake datashift:spree:products input=C:\MyProducts.xls - *Seamless Spree Image loading can be achieved by ensuring SKU or class Name features in Image filename. Lookup is performed either via the SKU being prepended to the image name, or by the image name being equal to the **name attribute** of the klass in question. Images can be attached to any class defined with a suitable association. The class to use can be configured in rake task via parameter klass=Xyz. In the Spree tasks, this defaults to Product, so attempts to attach Image to a Product via Product SKU or Name. Image loading **does not** specifically require JRuby A report is generated in the current working directory detailing any Images in the paths that could not be matched with a Product. rake datashift:spree:images input=C:\images\product_images skip_if_no_assoc=true rake datashift:spree:images input=C:\images\taxon_icons skip_if_no_assoc=true klass=Taxon ## Import to Active Record ### Associations To perform a lookup for an associated model, the primary column(s) must be supplied, along with required select values for those columns. A single association column can contain multiple name/value sets, in string form : column:lookup_key_1, lookup_key_2,... So if our Project model has many Categories, we can supply a Category list, which is keyed on the column Category.reference with : |Categories| reference:category_001,category_002 During loading, a call to find_all_by_reference will be made, picking up the 2 categories with matching references, and our Project model will contain those two i.e project.categories = [category_002,category_003] ## Spree Suppprt ### OptionTypes & Variants When loaded with the Spree specific tasks, spree specific over rides are supported, such as direct s support for OptionTypes with values Any 'Option Types' columns can contain the OptionType to associate with the Product, plus a selection of appropriate OptionValues to go with that Type. For example, in a single column/row we could supply 2 OptionTypes (named, size & colour), with a selection values (such as small, medium etc) 'Option Types' size:small,medium,large|colour:red,white If no such OptionType exists, e.g size, then a new one is created with the supplied name. Next the OptionValues are also parsed, again if no such OptionValue exists, e.g small, then a new one is created with the supplied name. Lastly a Variant is created on each OptionValue, with price and availaable dates being copied from Master. Currently a unique SKU is created by adding an index to the master's sku. TODO - Enable a hash of attributes to be supplied in association columns to enable more control over creation of associated objects. ### Properties The properties to associate with this product. Properties are for small snippets of text, shared across many products, and are for display purposes only. An optional display value can be supplied to supplement the displayed text. As for all associations can contain multiple name/value sets in default form : Property:display_value|Property:display_value Example - No values : manufacturer|standard Example - Display values : manufacturer:somebody else plc|standard:ISOBlah21 ## TODO - Add direct Image loading to Spree i.e should be able to specify just path in a column .. "/images/red-tshirt.jpg" - Smart sorting of column processing order .... - Does not currently ensure mandatory columns (for valid?) processed first. - Look at implementing import/export API using something like https://github.com/ianwhite/orm_adapter rather than active record, so we can support additional ORMs - Create separate Spree extension to support import/export via the admin gui ## License Copyright:: (c) Autotelik Media Ltd 2011 Author :: Tom Statter Date :: Dec 2011 The MIT License Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.