eplusr v0.13.0

New features

• Idf$add() and Idf$set() have new features:

  # refer to field using '..'
  idf$add(Material = list(..1 = "mat", ..7 = 0.95))
  idf$set(mat = list(..6 = 0.5))
  
  # using vector field values
  idf$add(Material := list(..1 = sprintf("mat%i", 1:10)))
  idf$set(c(sprintf("mat%i", 1:10)) := list(..6 = runif(10)))

• Idf$to_table() gains a new parameter force. The default value is FALSE. If TRUE, you can convert object data from any classes into a wide data.table. This may be useful when you know that target classes have the exact same fields, e.g. Ceiling:Adiabatic and Floor:Adiabatic (#202).

• A new method Idf$purge() has been added (#223). It can be used to delete any resource objects that are not referenced by other objects. Here resource objects indicate all objects that can be referenced by other objects, e.g. all schedules. $purge() will ignore any inputs that are not resources. If inputs contain objects from multiple classes, references among them are also taken into account, which means purging is performed hierarchically. If both materials and constructions are specified, the latter will be purged first, because it is possible that input constructions reference input materials. Idf$purge() makes it quite straightforward to perform IDF cleaning. Actions like removing all materials, constructions and schedules can be easily achieved via

  Idf$purge(class = c("Material", "Construction"), group = "Schedules")

• New methods Idf$duplicatd() and Idf$unique() have been added. They can be used to detect and remove duplicated objects, respectively. Here duplicated objects refer to objects whose field values are the same except the names. Object comments are ignored during comparison. These two methods can be useful when doing model cleaning (#227).

• Now if class is set to NULL in Idf$definition(), the underlying Idd object is returned (#237).

• Internal helper functions with_option(), with_silent(), with_verbose(), with_speed() and without_checking() have been exported. They can be used to evaluate an expression with temporary eplusr options (#240).

• Now Idf$insert() can directly take an Idf object or a list of Idf objects as input. And also Version objects in input will be directly skipped instead of giving an error (#245).

• A new option all has been added in IdfObject$print() with default being FALSE. If TRUE, all fields defined in [Idd] are printed even they do not exist in current object (#247).

• New S3 methods of == for all classes are added, i.e. ==.Idf, ==.IdfObject, ==.Idd, ==.IddObject, ==.Epw, ==.EplusJob, ==.EplusSql, ==.EplusGroupJob, ==.ParametricJob. The negate methods are also added. This makes it quite easy to check the equality of R6 objects of these types (#250).

• A new Generic function reload() is added. eplusr relies heavily on the data.table package. The core data of all main classes in eplusr are saved as data.tables. This introduces a problem when loading saved Idf objects or other class objects via an *.RDS and *.RData file on disk: the stored data.tables lose their column over-allocation. reload() is a helper function that calls data.table::setDT() on all internal data.tables to make sure they are initialized properly. It is recommended to call reload() on each Idd, Idf and other class object in eplusr loaded with readRDS() or load(), to make sure all eplusr's functionaries works properly (#251).

• The implementation of EplusSql$report_data() has been refactored, resulting in a ~200% speed-up (#259).

• Now day_type in EplusSql$report_data() has a few new options (#259): - "Weekday": All working days, i.e. from Monday to Friday - "Weekend": Saturday and Sunday - "DesignDay": Equivalent to "SummerDesignDay" plus "WinterDesignDay" - "CustomDay": CustomDay1 and CustomDay2 - "SpecialDay": Equivalent to "DesignDay" plus "CustomDay" - "NormalDay": Equivalent to "Weekday" and "Weekend" plus "Holiday"

• Some internal functions have been exported. They are mainly useful for developers to handle internal IDD and IDF data more efficiently (#260).

• A new IdfScheduleCompact class is introduced. A constructor function schedule_compact() is added. IdfScheduleCompact class provides more detailed methods to add, modify and extract schedule values. For more details, see vignette("schedule") (#256).

• New IdfGeometry and IdfViewer classes are introduced. IdfGeometry is designed to extract data for all geometry objects and perform geometric operations on them, while IdfViewer is to view IDF geoemtry in 3D using the rgl package in a similar way as OpenStudio SketchUp Plugin. Idf$geometry() and Idf$view() methods are added to directly create an IdfGeometry and IdfViewer object based on current Idf object, respectively (#296).

• A plot.Idf method is added which is basically a wrappper of Idf$view() (#296).

• Now eplusr can utilize the CSV output for report data extraction. Benifiting from the fantastic data.table::fread, this approach can be as 3~10X faster compared to the SQLite approach. eplusr will still use the SQLite if the CSV output is not available.

Major changes

• The algorithm of object/field/value relation extraction has been completed refactored. Now it can correctly detect object recursive-reference and it's faster (#222, #223).

  All relation-related methods now have an unified interface:

  X$method(which, direction, object = NULL, class = NULL, group = NULL, depth = NULL, keep = FALSE)

  Where which is a class index or object ID, direction is the target relation direction to extract. All results can be further constrained via 3 extra arguments, i.e. object, class and object. object only applicable to Idf and IdfObject. The depth argument is used to control the depth for searching recursive relations. Default value is 0L, which means no recursive relations will be detected, while NULL means to search all possible recursive relations.

  A new keep parameter with default value FALSE has been added. If TRUE, all input fields will be returned, even they may not have any relations. This is the default behavior of v0.12.0 and before. In this version, only fields that have relation with other objects will be returned.

  With this update, it is possible, for example, to directly know the structure of an air loop by using idf$object_relation("AnAirLoop", depth = NULL)

  Moreover, a new argument class_ref can be specified in methods of value-relation extraction. It can be used to specify how to handle class-name-references. Class name references refer to references in like field Component 1 Object Type in Branch objects. Their value refers to other many class names of objects, instaed of refering to specific field values. There are 3 options in total, i.e. "none", "both" and "all", with "both" being the default.

  • "none": just ignore class-name-references. It is a reasonable option, as for most cases, class-name-references always come along with field value references. Ignoring class-name-references will not impact the most part of the relation structure.
  • "both": only include class-name-references if this object also reference field values of the same one. For example, if the value of field Component 1 Object Type is Coil:Heating:Water, only the object that is referenced in the next field Component 1 Name is treated as referenced by Component 1 Object Type. This is the default option.
  • "all": include all class-name-references. For example, if the value of field Component 1 Object Type is Coil:Heating:Water, all objects in Coil:Heating:Water will be treated as referenced by that field. This is the most aggressive option.

• read_epw() will proceed parsing for non-standard EPW header format (#236).

• Now EplusSql$report_data() will set the year values of day type SummerDesignDay and WinterDesignDay to current year and the day_type value will be left unchanged (#258).

• Now read_idf() will always make sure all necessary fields are added during parsing (#267).

• [[.IdfObject now only accept standard field names. No underscore-style names are allowed.

• The suffix of automatcially created names in Idf$dup() has been changed from _X to X.

• The warning parameter in read_epw(), Epw$add() and Epw$set() has been deprecated (#298).

Minor changes

• EplusJob, EplusGroupJob and ParametricJob will not parse input EPW files, but only validate their existences and store the paths (#215)
• period parameter in EplusSql$report_data() now works as expected (#259).
• run_idf() and run_multi() now return additional element/column called
• version which contain the versions of EnergyPlus that are called during simulations
• format.Idd() now returns a single line string in format <EnergyPlus IDD v[Version] (Build) with X classes.
• The column datasource returned in Epw$data() has been renamed to data_source.

Bug fixes

• Fix the bug caused by ExpandObjects exectuable that causes run_idf fails when running in parallel (#130)
• Idf$insert() now will remove all duplicated objects in input (#219).
• Fix the bug in install_eplus() on Windows (#230)
• Fix the error in $<-.Idf when input list of IdfObjects are all from the same Idf on the LHS (#238).
• Now Idf$insert() and Idf$load() can now successfully remove duplicated objects by comparing field values case-insensitively (#243)
• Now Epw$save() can work with empty DESIGN CONDITIONS, TYPICAL/EXTREME PERIODS and GROUND TEMPERATURES headers. Thanks @lukas-rokka for the bug report (#263).
• Fix output directory creation error in EplusGroupJob(#270).
• Fix IDF header option parsing (#278).
• Trailing spa...

eplusr v0.12.0

New features

• Now group_job() supports single IDF input with multiple EPW inputs (#185).
• A new method Idf$last_job() has been added to enable getting the last simulation job created using Idf$run() (#187).
• Provide a workaround to fix the issue of EnergyPlus v9.1 and above installation which fails to extract files into correct directory /usr/local/EnergyPlus-X-Y-0, but instead extracting all files directly into /usr/local (NREL/EnergyPlus#7256) (#193).
• A new parameter group_ext has been added in Idf$to_table() and IdfObject$to_table(), with default value being "none". If "group", values from extensible fields will be grouped by the extensible group they belong to. For example, coordinate values of each vertex in class BuildingSurface:Detailed will be put into a list. If "index", values from extensible fields will be grouped by the extensible field indices they belong to. For example, coordinate values of all x coordinates will be put into a list (#74).

Bug fixes

• The algorithm of detecting numeric columns in EplusSql$tabular_data() has been improved (#190).
• Now EplusSql$tabular_data() keeps the original column order when wide is TRUE (#186).
• Fix EnergyPlus installation on macOS (#193).
• Fix parallel simulations on macOS (#194).
• Now eplus_config() will always return the expanded EnergyPlus path (#196).
• Now group_job() will return more informative error messages when input contains Idf objects that havn't been saved (#204).
• Fix error in EplusGroupJob$run() when custom dir is specified (#205).

eplusr v0.11.0

New features

• A new method $models() has been added in ParametricJob class. It returns a
  list of all parametric models generated after a measure has been applied. If
  no measure is applied, NULL is returned (#59). Thanks @yidan1214 for this
  feature request.
• A new method $save() has been added in ParametricJob class. It saves all
  generated parametric models and weather file into specified directory. You can
  use separate argument to determine whether each model is to be saved in a
  separate folder (#58). Thanks @yidan1214 for this feature request.
• New arguments align and all have been added to $to_table() method in
  Idf class. Setting align to TRUE will make sure that all returned object
  data per class have the same field number. The number of fields is the same as
  the object that have the most fields among those you specified. Setting all
  to TRUE will return all available fields in that class definition in IDD.
• Now the weather argument in $run() method in Idf class can be set to
  NULL. If so, design-day-only simulation is performed. Note that this needs
  at least one Sizing:DesignDay object exists in the Idf object (#80).
• Similar as above, the epw argument in eplus_job() and param_job() can
  also be NULL to force a design-day-only simulation.
• Now $status() in ParametricJob class includes a new member named
  job_status which is data.table containing detailed information on each
  simulation job (#70).
• Now $print() in ParametricJob will give you more informative details on
  each simulation job status, especially when wait is set to FALSE in
  $run().
• A new column index is added in the returned RddFile and MddFile. It
  contains index of each variable.
• Two new methods $read_rdd() and $read_mdd() have been added in EplusJob
  class. $read_rdd() and $read_mdd() which parse the simulation RDD and MDD
  file (#84).
• Two new function rdd_to_load() and mdd_to_load() have been added, which
  format RddFile and MddFile into a data.table in acceptable format for
  $load() method in Idf class.
• Similar as Output:SQLite, when $run() in Idf object is called, an object
  in class Output:VariableDictionary is automatically created with Key Field
  being IDF (#85).
• A new argument echo has been added in $run() in Idf, EplusJob and
  ParametricJob class. It is only applicable when wait is TRUE. If FALSE,
  the simulation will be run silently without echoing any message from
  EnergyPlus.
• A new function transition() has been added. Basically it is an
  R implementation of IDFVersionUpdater. Currently the lowest version of IDF is
  v7.2. It should be much faster than IDFVersionUpdater.
• Now the way of find IDD file has been changed to take advantage of IDD files
  distributed along with IDFVersionUpdater. This update makes it possible to
  directly read IDF of most versions without downloading corresponding IDD.
• A new option autocomplete with default value being interactive() has been
  added. It is used to control whether to turn on autocompletion on class and
  field names. Underneath, makeActiveBinding() is used to add or move active
  bindings in Idf and IdfObjects to directly return objects in class or
  field values. This will make it possible to dynamically show current class
  and field names in both RStudio and in the terminal. However, this process
  does come in with a penalty on the performance. It can make adding or
  modifying large mounts of [Idf] and [IdfObject]s extremely slow. Default value
  make sure autocompletion works in interactive mode.
• A new syntax class := list(field = value) in $set() has been added. Note
  the use of := instead of =. The main difference is that, unlike =, the
  left hand side of := should be a valid class name in current Idf object.
  It will set the field of all objects in specified class to specified value.
• A new function dt_to_load() has been added, which formats a data.table from
  Idf$to_table() and IdfObject$to_table() with wide being to TRUE into a
  data.table in acceptable format for Idf$load() method and Idf$update()
  method.
• Now read_epw() support EPW files with non-integer timezones fail to load
  (#113). $location() in Epw class also support setting the timezone to
  non-integer one.
• A new method $parent() has been added in IddObject and IdfObject class
  to get parent Idd and Idf object, respectively (#76).
• Simulation status will be updated in the progress message in run_multi() and
  $run() method in ParametricJob class, instead of only showing COMPLETED
  (#124, #125).
• EplusJob class now will parse and store input IDF and EPW. It will also add
  Output:SQLite and Output:VariableDictionary object if necessary. This
  change makes sure using EplusJob can always have the ability to extract
  simulation results instead of totally relying on the input IDF (#118).
• As a result of previous change, the $run() method in EplusJob now takes
  the same parameters as $run() method in Idf, i.e. you can also change the
  EPW file to use and output directory using epw and dir argument (#118).
• A new class EplusGroupJob is added. It can be created using
  group_job(). EplusGroupJob provides a wrapper of run_multi() to group
  multiple EnergyPlus simulations together for running and collecting different
  EnergyPlus outputs (#117).
• A new parameter named .empty has been added in $set(), $insert(),
  $load(), $update(), $paste() methods in Idf class and $set() method
  in IdfObject class (#133).
• EnergyPlus v9.2 support has been added (#138).
• Now required objects and unique objects can be deleted in $del() when
  .force is set to TRUE (#149).
• Now for Windows and Linux, install_eplus() supports to install EnergyPlus to
  your home directory or your customized directory without requiring
  administrator privileges (#167).

  # install to your home directory
  install_eplus(8.8, local = TRUE)
  # install to custom directory
  install_eplus(8.8, local = TRUE, dir = "~/MyPrograms")

  Please see ?install_eplus for details.
• All documentation in R6 classes have been update thanks to roxyten2 R6 support (#156).
• Deprecated methods in each class have all been remove (#156).
• New parameter case has been added in
  EplusSql$tabular_data(). Similar like case parameter in
  EplusSql$report_data(), it lets you to add a case column to indicate the
  case of simulation. This brings some changes in the returned results of
  EplusSql$tabular_data(). Compared to previous version, there will always be
  a case column, unless case parameter is set to NULL (#182).
• New parameter wide and string_value have been added in
  EplusSql$tabular_data() and EplusGroupJob$tabular_data(). When wide is
  TRUE, each table will be converted into the similar format as it is shown in
  EnergyPlus HTML output file. And when string_value is FALSE, values in
  possible numeric columns are converted into numbers (#182).

Bug fixes

• Fix the error in $set() when $add_unit() is performed in Epw class (#56).
• Now IDF file located in a folder whose name contains spaces can be
  successfully simulated (#60). Thanks @yidan1214 for reporting this bug.
• Now $ref_to_object() and $ref_by_object() can give correct results when
  class argument is specified.
• Now IDD version lower than 8.3 can successfully be downloaded and parsed.
• Now ErrFile objects returned from $errors() in ParametricJob can be
  successfully printed.
• Now one-line empty objects, e.g. "Output:Surfaces:List,,;\n", can be
  successfully parsed (#88).
• Fix the error of $to_table() when no arguments were provided (#103).
• The references of newly added extensible fields are now correctly handled
  (#109).
• $update() now works for class Schedule:Compact with type checking on
  (#111).
• Fix the error in $ref_to_node() when no objects in specified class is found
  (#110).
• Now leading and sep_at argument work as expected in Idf$to_string() (#160).
• Now Idf$to_table() matches object names case-insensitively (#157).
• One-line-formatted Version object with trailing comments can be successfully
  parsed, e.g. Version, 8.6; !- ABC (#170).
• Better support for old IDD (#183).

Minor changes

• Describe on how the arguments are matched in $apply_measure() (#57). Thanks
  @yidan1214 for pointing this out.
• Now the echo argument in run_idf() and run_multi() will only take effect
  when wait is TRUE. This makes it possible to remove the dependency on the
  later package.
• All messages can be suppressed by setting the verbose_info option to
  FALSE.
• $delete() method in Epw class has been deprecatd in flavor of $del() to
  provide a similar API as Idf class.
• When all is TRUE in $report_data() in EplusSql, EplusJob and
  ParametricJob class, an extra column environment_period_index is returned
  which contains the indices of each run period environment. This helps to
  distinguish different run period environment when no name is assigned.
• param_job() now gives more informative error message if input Idf object
  and Epw object is not created from a local file (#112).
• param_job() now preserve parametric model names from argument .names in
  $apply_measure() instead of calling make.name() to convert them into valid
  R names (#115).
• $save() works if weather was not given during initialization for
  ParametricJob (#120).
• Required fields in IdfObject are now marked with * when printing (#135).

eplusr v0.10.4

This is a bug fix release which make sure eplusr is compatible with new version of R6 #164. No new feature is included.

eplusr v0.10.3

New features

• Now you can get autocompletion of class names of Idd and Idf objects, and
  field names of IdfObject class in RStudio.
• [[<-.Idf and $<-.Idf now can work with unique-object classes as well.
• A new method $update() is added in Idf class which makes it possible to
  directly modify existing objects using character vectors and data.frames.
• A new argument wide is added in $to_table() method in Idf class,
  similar to $to_table() method in IdfObject class.
• New function read_rdd() and read_mdd() are added to help to directly parse
  .rdd and .mdd file.
• Now $rename() will not give an error if input name is the same as the old
  one.
• Now new method $version() is added in EplusJob and ParametricJob class,
  which returns the version of IDF it uses.
• A new argument by_group is added in $class_index(), $class_name() in
  Idd class and $class_name() in Idf class. If TRUE, a list is returned
  which separate class indexes/names by the group they belong to.
• Recursive relation support has been added to $object_relation() in Idf
  class, and $value_relation(), $ref_to_object(), $ref_by_object() in
  IdfObject class. All type of relations can be searched recursively by
  setting newly added argument recursive to TRUE. There is also a new
  argument called depth to control how many recursive searching steps should
  be performed.
• Add Node support into $value_possible() in IdfObject class, i.e. if
  specified fields are nodes, all nodes are returned in the source component.
• Now component node references support has been added. A new option "node"
  has been added to the direction argument in $object_relation,
  $objects_in_relation() in Idf class and $value_relation() in IdfObject
  class. New methods $ref_to_node() and $has_ref_node() have been added in
  IdfObject class. Now $has_ref() in IdfObject class will return FALSE
  only when all $has_ref_to(), $has_ref_by() and $has_ref_node() are all
  FALSE.

Bug fixes

• Now all internal data, specifically all data.tables are correctly copied
  when calling $clone() in Idf and Epw class.
• $add() method and other methods that modify field values can correctly
  convert field values to numbers if necessary.
• Fix the error that holiday header got overwritten (#43).
• Fix EPW date parsing error (#42).
• Fix warnings in parsing .err file when there is no warning nor error.
• Fix $errors() error in ParametricJob class.
• The year of returned date time from $data(), $abnormal_data() and
  $redundant_data() now have correct values (#44).
• Reset year after checking datetime in read_epw() (#44).
• Add field name in input before validation in paste() (#45).
• Fix datetime calculation in $report_data() in EplusSql (#46).
• Update doc on EnergyPlus time nodation (#47).
• Fix EPW design condition parsing error when hours_8_to_4_12.8_20.6 is zero.
• Now recurring errors in .err file are parsed correctly in ErrFile.
• Handle large value references in get_value_reference_map() (#50).
• Fix the error when extract numeric default values using $value_possible()
  (#51).
• Fix the error that .ref_to argument in $del() in Idf class did not take
  effect.
• Fix field deletion in $update() in Idf class.
• Fix reference parsing error with mixed source types.
• External files used in Schedule:File class are now only copied once when
  copy_external is TRUE in $save() in Idf class.
• When .ref_to or .ref_by is TRUE in $del() in Idf class, objects
  whose class names are referenced by input will not be deleted, except all
  objects in referenced class are detected by field value reference or there is
  only one object in referenced class.
• Objects detected using .ref_to and .ref_by in $del() in Idf class can
  now be successfully deleted, even if they still have relations with other
  objects.
• Now invalid input object name is kept as it is, instead of converted to lower
  case.
• Now $value_possible() returns correct source values when fields refer to
  class names rather than field values.
• Now $del() method in Idf class works correctly with multiple inputs.
• Now trailing comments in IDF will be removed in read_idf().
• Now $to_table() in Idf class will keep the input object order.
• Now $set() method in IdfObject can successfully delete fields when field
  values are set to NULL.
• Now $run() method in Idf will use the correct model path to run
  simulation.

Minor changes

• Change message types in $ref_to_object() and $ref_by_object() in IdfObject.
• Now ErrFile is stored as a data.table with additional attributes instead
  of a list.
• Now when argument .unqiue is TRUE in $insert(), $load() and $paste()
  in Idf class, object comparison are performed case-insensitively.
• A new default value "." of dir in download_idd() is add.

eplusr v0.10.2

This is a patch update mainly to fix CRAN check errors on a strict Latin locale, with a bug fix.

Bug fixes

$report_data() method in EplusSql class now correctly returns data if even if key_value is not given (#37)

eplusr v0.10.1

This is a patch update mainly to fix CRAN check errors, with few bug fixes.

Bug fixes

• $set() method in Idf class now does not try to convert all input strings
  into numbers (#32)
• $del() method in Idf class now also respects validation settings.
  You should be able to delete any object if current validation level does not
  include reference checking
• $del() method in Idf class now will only print object relation only if
  .ref_to or .ref_by is TRUE
• Now $set() method in Epw class will correctly reset leap year flag in the
  header (#32)
• Now $save() method in Epw class returns the path of saved file invisibly

eplusr v0.10.0

eplusr v0.10.0 introduces dozens of improvements, unfortunately also has some
break changes. Detailed updates per class are as follow:

Idd Class

• $object() now takes a single class name and returns an IddObject.

• New method $objects() is added as the successor of $object() in previous
  versions.

• New method $object_relation() and $objects_in_relation() are added to get
  all objects that have relation (one refer to others or one referred by others).

• $object_in_group() has been deprecated in flavor of $objects_in_group().

• New method $to_table() and $to_string() are added to help easily extract
  internal IDD data in data.table format and string format respectively.

• $clone() method is now deprecated.

• S3 method str.Idd(), format.Idd() are added.

IddObject Class

• A constructor idd_object() has been added to help directly create
  IddObject. It takes a parent IDD version or an Idd object and a valid
  class name or class index, e.g. idd_object(8.8, "Material"),
  idd_object(use_idd(8.8), "Construction")

• New method $version() is added to extract the version of parent IDD.

• New method $field_relation() is added to extract the field relation in
  current class with other class fields.

• New method $has_ref(), $has_ref_to(), and $has_ref_by() are added to
  help to detect field relations

• New method $is_real_field() is added to detect if specified fields should
  have values of real numbers but not integers.

• New method $has_ref(), $has_ref_by(), $has_ref_to() are added to detect
  if specified fields refer to other fields or can be referred by other field in
  other classes.

• New method $to_table() and $to_string() are added to help easily extract
  internal IDD data in data.table format and string format respectively.

• field_possible() now does not need IdfObject data to run.

• Argument lower in $field_name() has been deprecated.

• $field_reference() in IddObject class has been deprecated in flavor of
  $field_relation().

• A brief argument is added in $print() with default value of FALSE. If
  TRUE, only the class name of the IddObject is printed.

• $clone() method is now deprecated.

• S3 method format.IddObject(), as.character.IddObject() and
  str.IddObject() are added which calls $to_string() and $print().

Idf Class

• An argument sorted is added in $class_name() and $group_name()to control
  whether to sort the returned value according to the occurrence in IDD or not.

• $definition() now only accepts one class name and returns a single
  IddObject, not a list of IddObject.

• $object() now takes a single object ID or name and returns an IdfObject.

• New method $objects() is added as the successor of $object() in previous
  versions.

• New method $object_unique() is added which returns an IdfObject in
  unique-object class, e.g. SimulaSimulationContrl. This makes it easy to
  directly extract and modify those objects, e.g.
  model$object_unique("SimulationContrl")$set(...).

• $object_in_class() now is deprecated in flavor of $objects_in_class().

• New method $objects_in_group() is added to get all objects in specified group.

• $search_object(), $search_value() and $replace_value() now can take same
  arguments as base::grepl() to further control the way it searches objects.

• $replace_value() now will perform validations when replacing value.

• Now all methods in Idf class that return IdfObjects will not convert
  object names into valid R names. The former conversion behavior is unnecessary
  and a little bit confusing.

• $add_object(), $set_object(), $del_object(), $ins_object() and
  $del_object() are now all deprecated. Please use $add(), $set(),
  $del() and insert() instead, which provide much cleaner syntax.

• New method $rename() is added which helps to modify only object names.

• New method $paste() is added which will parse contents that copied from IDF
  Editor using the Copy Obj button and directly insert them into current Idf.

• $validate() and $is_valid() now accepts an argument level. Also, a
  helper custom_validate() is added to customize what kind of validations to
  check.

• $string() is deprecated in flavor of $to_string() in Idf class.

• New method $to_table() is added which can extract object data into a
  data.table.

• The default value of deep argument in $clone() method is set to TRUE. As
  in almost all cases, a total cloned object is desired when calling $clone().

• plain argument is deprecated in $print() in IDF class. Because the same
  thing can be achieved using $to_string() and base::cat(). New argument
  zoom and order are added, which give more control on how detailed should
  the model be printed.

• New method $object_relation() and $objects_in_relation() are added to
  extract all objects in relation.

IdfObject Class

• New constructor idf_object() is added.

• New method $version() is added to get the underlying version of IDD.

• $get_comment(), $get_value() now have been deprecated in flavor of
  $comment() and $value()

• $comment() (former $get_comment()) now returns NULL if the object does
  not have any comment.

• $set_value() now has been deprecated in flavor of $set()

• $possible_value() now has been deprecated in flavor of $value_possible().

• $ref_from_object() and $has_ref_from() now have been deprecated in flavor
  of $ref_to_object() and $has_ref_to().

• $has_ref_by(), $has_ref_to() and $has_ref() now return a logical vector
  of the same length as field numbers.

• $validate() and $is_valid() now accepts an argument level. Also, a
  helper custom_validate() is added to customize what kind of validations to
  check.

• $string() and $table() are now deprecated in flavor of $to_string() and
  $to_table().

• $clone() method is now deprecated.

• S3 methods format.IdfObject(), as.character.IdfObject() and
  str.IdfObject() are added, which calls $to_string() and $value().

Epw Class

• Epw class has been totally rewritten which provides much more
  functionalities. Please refer to package documentation for details.
  All headers are parsed and can be retrieve in Epw class.

• The default value of deep in $clone() has been change to TRUE.

EplusSql Class

• New method $path_idf() is added to return the parent IDF file path.

• New arguments period, month, day, hour, minute, interval,
  simulation_days, day_type and environment_name are added in
  $report_data() which provide extra subsetting on the SQL file.

• New arguments report_name, report_for, table_name, column_name and
  row_name are added in $tabular_data() which provide extra subsetting on
  the SQL file.

• $clone() method is now deprecated.

EplusJob Class

• New method $path() is added to return the path of IDF and/or EPW file of
  the job.

• New method $list_table(), $read_table() are added which provide the same
  functionalities as they are in EplusSql class.

• $clone() is now deprecated.

• New arguments period, month, day, hour, minute, interval,
  simulation_days, day_type and environment_name are added in
  $report_data() which provide extra subsetting on the SQL file.

• New arguments report_name, report_for, table_name, column_name and
  row_name are added in $tabular_data() which provide extra subsetting on
  the SQL file.

• $clone() method is now deprecated.

ParametricJob Class

• New arguments force and copy_external are added in $run() which have the
  same effect as in $run() method in Idf class.

• New arguments period, month, day, hour, minute, interval,
  simulation_days, day_type and environment_name are added in
  $report_data() which provide extra subsetting on the SQL file.

• New arguments report_name, report_for, table_name, column_name and
  row_name are added in $tabular_data() which provide extra subsetting on
  the SQL file.

• $clone() method is now deprecated.

Miscellaneous

• Function read_err() is added which takes a path of an .err file, parses it
  and returns an ErrFile object.

• Functions is_eplus_ver(), is_idd_ver(), is_eplus_path(), is_idd(),
  is_idf(), is_iddobject(), is_idfobject() and is_epw() are now exported.

• Function custom_validate() and level_checks() are added to customize
  validation.

• Right now, all returned object and field names will remain what they are as in
  the IDF, instead of converting them into underscore style names in the
  returned lists.

• The num_digits option has been deprecated as formatting of numeric fields
  are now handled by R itself.

• Error will be given instead of warning if there is no build tag or multiple
  build tags found in input IDD.

• eplusr_option() accepts custom validate level using newly-added function
  custom_validate().

v0.9.4

New features

• Now you can directly download EnergyPus Weather File (.epw) and Design Day
  File (.ddy) using new function download_weather(). It takes a regular
  expression as input, searches through EnergyPlus weather data base (stored
  in this package), download corresponding files and return the saved paths.
  Below are some examples:
  • Search locations that contains string "los angeles" and "tmy", return a
    menu to select which one(s) to download. Once selected, download both
    corresponding weather file(s) and design day file(s):

    download_weather("los angeles.*tmy", type = "all", ask = TRUE)

  • Same as above, expect that all files will be renamed to la.epw(ddy),
    la_1.epw(ddy), la_2.epw(ddy):

    download_weather("los angeles.*tmy", filename = "la", type = "all", ask = TRUE)

  • Search locations that contains string "beijing" and "cswd". If no more
    than 3 results found, directly download all weather files and save them to
    temperory directory.

    download_weather("beijing.*cswd", dir = tempdir(), type = "epw", ask = FALSE, max_match = 3)

Minor changes

• Now clean_wd() is run before every call to EnergyPlus.

Bug fixes

• $clone() method has been refactored for Idf and Idd class. This fix the
  issue that internal shared environments were not cloned in version 0.9.3.

• Fix the error that $save() and $string() in Idf class did not respect
  format argument.

• Fix the error that $apply_measure() in ParametricJob class did not
  respect .names argument.

v0.9.3

This is a maintenance release aiming to add support for EnergyPlus v9.0.0.

New features

• Add support for EnergyPlus v9.0.0.

Bug fixes

• Remove duplicates when update value references (#20).