All Versions
23
Latest Version
0.3
Avg Release Cycle
184 days
Latest Release
978 days ago

Changelog History
Page 2

  • v0.1.2 Changes

    February 17, 2016
    • โœจ Enhancements

      • New method DataFrame.from_activerecord for importing data sets from ActiveRecord. (by @mrkn)
      • Better importing of data from SQL databases by extracting that functionality into a separate class called Daru::IO::SqlDataSource (by @mrkn).
      • Faster algorithm for performing inner joins by using the bloomfilter-rb gem. Available only for MRI. (by Peter Tung)
      • Added exception SizeError (by Peter Tung).
      • Removed outdated dependencies and build scripts, updated existing dependencies.
      • Ability to sort a Daru::Vector with nils present (by @gnilrets)
    • ๐Ÿ›  Fixes

      • Fix column creation for Dataframe.from_sql (by @dansbits).
      • group_by can now be performed on DataFrames with nils (@gnilrets).
      • Bug fix for DataFrame Vectors not duplicating when calling DataFrame#dup (by @gnilrets).
      • Bug fix when concantenating DataFrames (by @gnilrets)
      • Handling improper arguments to Daru::Vector#[] (by @lokeshh)
      • Resolve narray conflict by using the latest nmatrix require methods (by @lokeshh)
  • v0.1.1 Changes

    August 19, 2015
    • โœจ Enhancements
      • Added a new class Daru::Offsets for providing a uniform API to jump between dates.
      • Added benchmarking scripts
      • Added a new Arel-like querying syntax for Vector and DataFrame. This will allow faster and more intuitive lookup of data than using loops such as filter.
      • Vector
        • #concat now compulsorily requires a second index argument.
        • Added new method #index= to change the index directly.
        • Added basic functions for rolling statistics - mean, std, count, etc.
        • Added cumulative sum function.
        • Added #keep_if.
        • Added #count_values.
      • Indexing
        • Changed Index so that it now accepts all sorts of data (not restricted to only Symbols as it was previously).
        • Re wrote MultiIndex in levels and labels form so that its faster and more accomodative of different kinds of index levels.
        • Changed .new to return appropriate index object based on data passed.
        • Added .from_tuple and .from_array methods to MultiIndex.
        • Added union and intersection behaviour to Index and MultiIndex.
        • Added a new index, DateTimeIndex for indexing with time-based data.
        • Optimized range search for Index.
      • DataFrame
        • Removed the DataFrameByVector class and the #vector function. Now only way to access a Vector in a DF is by using the #[] operator.
        • Added new method #index= and #vectors= for changing row and column indexes directly.
        • Optimized Vector value setting and retreival.
        • Added inner, outer, left outer and right outer joins with the #join method.
        • Added #set_index.
    • ๐Ÿ”„ Changes
      • Removed the + operator overload from Index and replaced in with union.
      • Removed the second 'values' argument from Daru::Index because it's redundant.
      • Changed behaviour of Vector#reindex and DataFrame#reindex and #reindex_vectors to preserve indexing of original data when possible.
    • ๐Ÿ›  Fixes
      • Fixed DataFrame#delete_row and Vector#delete_if.
      • Fixed Vector#rename.
  • v0.1.0 Changes

    June 13, 2015
    • ๐Ÿ›  Fixes
      • Update documentation and fix it in other places.
      • Fix Vector#sum_of_squares and #ranked.
      • Fixed some tests that were giving RSpec warnings
      • Fixed a bug where nyaplot not being present would raise a warning.
      • Fixed a bug in DataFrame row assignment.
    • โœจ Enhancements
      • Wrote a proper .travis.yml
      • Added optional GSL dependency gsl-nmatrix
      • Added Marshalling and unMarshalling capabilities to Vector, Index and DataFrame.
      • Added new method Daru::IO.load for loading data from files by marshalling.
      • Lots of documentation and new notebooks.
      • Added data loading and writing from and to CSV, Excel, plain text and SQL databases.
      • Daru::DataFrame and Vector have now completely replaced Statsample::Dataset and Vector.
      • Vector
        • #center
        • #standardize
        • #vector_percentile
        • Added a new wrapper class Daru::Accessors::GSLWrapper for wrapping around GSL::Vector, which works similarly to NMatrixWrapper or ArrayWrapper.
        • Added a host of statistical methods to GSLWrapper in Daru::Accessors::GSLStatistics that call the relevant GSL::Vector functions for super-fast C level computations.
        • More stats functions - #vector_standardized_compute, #vector_centered_compute, #sample_with_replacement, #sample_without_replacement
        • #only_valid for creating a Vector with only non-nil data.
        • #only_missing for creating a Vector of only missing data.
        • #only_numeric to create Vector of only numerical data.
        • Ported many Statsample::Vector stat methods to Daru::Vector. These are: #percentile, #factors, etc.
        • Added .new_with_size for creating vectors by specifying a size for the vector and a block for generating values.
        • Added Vector#verify, #recode! and #recode.
        • Added #save, #jackknife and #bootstrap.
        • Added #missing_values= that will allow setting values for treating data as 'missing'.
        • Added #split_by_separator, #split_by_separator_freq and #splitted.
        • Added #reset_index!
        • Added #any? and #all?
        • Added #db_type for guessing the type of SQL type contained in the vector.
        • Added and tested plotting support for histogram and box plot.
      • DataFrame
        • #dup_only_valid
        • #clone, #clone_only_valid, #clone_structure
        • #[]= does not clone the vector if it has the same index as the DataFrame.
        • Added a :clone option to initialize that will not clone Daru::Vectors passed into the constructor.
        • Added #save.
        • Added #only_numerics.
        • Added better iterators and changed some behaviour of previous ones to make them more ruby-like. New iterators are #map, #map!, #each, #recode and #collect.
        • Added #vector_sum and #vector_mean.
        • Added #to_gsl to convert to GSL::Matrix.
        • Added #has_missing_data? and #missing_values_rows.
        • Added #compute and #verify.
        • Added .crosstab_by_assignation to generate data frame from row, column and value vectors.
        • Added #filter_vector.
        • Added #standardize and added argument option to #dup.
        • Added #any? and #all? for vector and row axis.
        • Better creation of empty data frames.
        • Added #merge, #one_to_many, #add_vectors_by_split_recode
        • Added constant SPLIT_TOKEN and methods #add_vectors_by_split, .[], #summary.
        • Added #bootstrap.
        • Added a #filter method to wrap around #filter_vectors and #filter_rows.
        • Greatly improved plotting function.
      • Added a lazy update feature that will allow users to delay updating the missing positions index until the last possible moment.
      • Added interoperaility with rserve client which makes it possible to change daru data to R data and perform computation there.
    • ๐Ÿ”„ Changes
      • Changes Vector#nil_positions to Vector#missing_positions so that future changes for accomodating different values for missing data can be made easily.
      • Changed History.txt to History.md
  • v0.0.5 Changes

    February 28, 2015
    • Easy accessors for some methods
    • Faster CSV loading.
    • ๐Ÿ”„ Changed vector #is_valid? to #exists?
    • Revamped dtype specifiers for Vector. Now specify :array/:nmatrix for changing underlying data implementation. Specigfy nm_dtype for specifying the data type of the NMatrix object.
    • #sort for Vector. Quick sort algorithm with preservation of original indexes.
    • Removed #re_index and #to_index from Daru::Index.
    • Ability to change the index of Vector and DataFrame with #reindex/#reindex!.
    • Multi-level #sort! and #sort for DataFrames. Preserves indexing.
    • All vector statistics now work with NMatrix as the underlying data type.
    • Vectors keep a record of all positions with nils with #nil_positions.
    • Know whether a position has nils or not with #is_nil?
    • โž• Added #clone_structure to Vector for cloning only the index and structure or a vector.
    • Figure out the type of data using #type. Running thru the data to determine its type is delayed till the last possible moment.
    • โž• Added arithmetic operations between data frame and scalars or other data frames.
    • โž• Added #map_vectors!.
    • Create a DataFrame from Array of Arrays and Array of Vectors.
    • ๐Ÿ”จ Refactored DataFrame.rows and the DataFrame constructor.
    • โž• Added hierarchial indexing to Vector and DataFrame with MultiIndex.
    • Convert DataFrame to ruby Matrix or NMatrix with #to_matrix and #to_nmatrix.
    • โž• Added #group_by to DataFrame for grouping rows according to elements in a given column. Works similar to SQL GROUP BY, only much simpler.
    • โž• Added new class Daru::Core::GroupBy for supporting various grouping methods like #head, #tail, #get_group, #size, #count, #mean, #std, #min, #max.
    • Tranpose indexed/multi-indexed DataFrame with #transpose.
    • ๐Ÿ’Ž Convert Daru::Vector to horizontal or vertical Ruby Matrix with #to_matrix.
    • Added shortcut to DataFrame to allow access of vectors by using only #[] instead of calling #vector or [vector_names, :vector].
    • โž• Added DSL for Vector and DataFrame plotting with nyaplot. Can now grab the underlying Nyaplot::Plot and Nyaplot::Diagram object for performing different operations. Only need to supply parameters for the initial creation of the diagram.
    • โž• Added #pivot_table to DataFrame for reducing and aggregating data to generate a quick summary.
    • โž• Added #shape to DataFrame for knowing the numbers of rows and columns in a DataFrame.
    • โž• Added statistics methods #mean, #std, #max, #min, #count, #product, #sum to DataFrame.
    • โž• Added #describe to DataFrame for producing multiple statistics data of numerical vectors in one shot.
    • ๐Ÿ’Ž Monkey patched Ruby Matrix to include #elementwise_division.
    • โž• Added #covariance to calculate the covariance between numbers of a DataFrame and #correlation to calculate correlation.
    • Enumerators return Enumerator objects if there is no block.
  • v0.0.4 Changes

    • โž• Added wrappers for Array, NMatrix and MDArray such that the external implementation is completely transparent of the data type being used internally.
    • โž• Added statistics methods for vectors for ArrayWrapper. These are compatible with statsample methods.
    • โž• Added plotting functions for DataFrame and Vector using Nyaplot.
    • Create a DataFrame by specifying the rows with the ".rows" class method.
    • Create a Vector from a Hash.
    • Call a Vector element by specfying the index name as a method call (method_missing logic).
    • Retrive multiple rows of a DataFrame by specfying a Range or an Array with multiple index names.
    • #head and #tail for DataFrame.
    • #uniq for Vector.
    • #max for Vector can return a Vector object with the index set to the index of the max value.
    • ๐Ÿ“š Tonnes of documentation for most methods.
  • v0.0.3 Changes

    • ๐Ÿš€ This release is a complete rewrite of the entire gem to accomodate index values.
  • v0.0.3.1 Changes

    • โž• Added aritmetic methods for vector aritmetic by taking the index of values into account.
  • v0.0.2 Changes

    • โž• Added iterators for dataframe and vector alongwith printing functions (to_html) to interface properly with iRuby notebook.
  • v0.0.2.4 Changes

    • ๐ŸŽ‰ Initialize dataframe from an array which looks like [{a: 10, b: 20}, {a: 11, b: 12}]. Works for parsed JSON.
    • Over-riding vectors in DataFrame will still preserve order.
    • Any re-assignment of rows in #each_row and #each_row_with_index will reflect in the DataFrame.
    • Added #to_a and #to_json to DataFrame.
  • v0.0.2.3 Changes

    • Added #filter_rows and #delete_row to DataFrame and changed #row to return a row containing a Hash of column name and value.
    • Vector objects passed into a DataFrame are now duplicated so that any changes dont affect the original vector.
    • โž• Added an optional opts argument to DataFrame.
    • Sending more fields than vectors in DataFrame will cause addition of nil vectors.
    • Init a DataFrame without having to convert explicitly to vectors.