Skip to content

Commit

Permalink
Update version
Browse files Browse the repository at this point in the history
  • Loading branch information
fabiofortkamp committed Feb 6, 2024
1 parent e45a068 commit 962c02f
Show file tree
Hide file tree
Showing 2 changed files with 200 additions and 84 deletions.
268 changes: 186 additions & 82 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,183 +1,287 @@
# 2.14.0

- Add keyword argument `folders_to_gitignore` to the `initialize_project` to customize the `.gitignore` file creation during project setup

# 2.13.0

- Add `commmit_message` option in `tag!`, which add an additional `"gitmessage"` field in dictionary `d` and include the git message associated with the commit.

# 2.12.6

- Crucial bugfix to `produce_or_load`. When used with a prefix, it attached double prefix to the file (one coming as a duplicate from `savename`). This is now fixed, but it means that some files produced with prefix and `produce_or_load` in v2.12 may be re-produced after this update.

# 2.12.0

- Arbitrary functions extracting strings from data can be used in `produce_or_load` instead of `savename`. `hash` is the most useful function here. An example in Real World Examples highlights this.
- Additional keywords in `produce_or_load` propagated to `savename` are deprecated.
- The example project created in the Workflow Tutorial no longer has spaces in its name

# 2.11.0

- Now the default project with `initialize_project` will include a documentation and a test folder, and setup CI for both automatically.
- `@produce_or_load` was broken but because CI was disabled this was silently ignored. `produce_or_load` now has call signature `produce_or_load(f::Function, config, path::String = "")` and same signature for the macro with `path` mandatory.

# 2.10.0
* Add `Derived` type and function to allow `dict_list` to compute some parameters from the value of other parameters.
* Add `DRWATSON_UPDATE_MSG` ENV VAR for suppressing update messages

- Add `Derived` type and function to allow `dict_list` to compute some parameters from the value of other parameters.
- Add `DRWATSON_UPDATE_MSG` ENV VAR for suppressing update messages

# 2.9.0
* Add `filename` option in `produce_or_load`, possibly overriding the default filename generated by `savename`

- Add `filename` option in `produce_or_load`, possibly overriding the default filename generated by `savename`

# 2.8.0
* Add filtering of `collect_results` using `rinclude` and `rexclude` keyword arguments.

- Add filtering of `collect_results` using `rinclude` and `rexclude` keyword arguments.

# 2.7.2
* By default `storepatch` keywords are `false`. This means that `gitpatch` is NOT stored by default. This is a BUGFIX, because there is an unknown problem of non-halting when storing the patch.

- By default `storepatch` keywords are `false`. This means that `gitpatch` is NOT stored by default. This is a BUGFIX, because there is an unknown problem of non-halting when storing the patch.

# 2.7.0
* DrWatson-related `ENV`ironment variables are now available to globally set the default values for e.g. story git patches, tagging, or safe-saving in various functions like `tagsave` or `produce_or_load`.

- DrWatson-related `ENV`ironment variables are now available to globally set the default values for e.g. story git patches, tagging, or safe-saving in various functions like `tagsave` or `produce_or_load`.

# 2.6.0
* Use `JLD2`'s jldopen in `collect_results!` to speed up loading of metadata.

- Use `JLD2`'s jldopen in `collect_results!` to speed up loading of metadata.

# 2.5.0
* Add an `update` option of `collect_results!` allowing the updating of an existing results collection if data files were modified or deleted.

- Add an `update` option of `collect_results!` allowing the updating of an existing results collection if data files were modified or deleted.

# 2.4.1
* `savename`'s default options now have `sigdigits = 3` instead of `digits = 3` as stated in the documentation string. This was supposed to happen already since 2.0 but did not because of a bug. (#284)
* Any subtypes of `AbstractDict` now work with DrWatson (#283).

- `savename`'s default options now have `sigdigits = 3` instead of `digits = 3` as stated in the documentation string. This was supposed to happen already since 2.0 but did not because of a bug. (#284)
- Any subtypes of `AbstractDict` now work with DrWatson (#283).

# 2.4.0
* Add the macro version of `produce_or_load` to enable tagging with the calling source file and line.

- Add the macro version of `produce_or_load` to enable tagging with the calling source file and line.

# 2.3.0
* Enable pass through of kwargs to `wsave` in `produce_or_load`, `tagsave` and `safesave` (to e.g. allow compression in JLD2 files).

- Enable pass through of kwargs to `wsave` in `produce_or_load`, `tagsave` and `safesave` (to e.g. allow compression in JLD2 files).

# 2.2.0
* `isdirty(gitpath = projectdir())` function for checking if `gitpath` points to a dirty Git repository. (#263)

- `isdirty(gitpath = projectdir())` function for checking if `gitpath` points to a dirty Git repository. (#263)

# 2.1.0
* `initialize_project` now allows for custom templates.

- `initialize_project` now allows for custom templates.

# 2.0.4
* `produce_or_load` now will not attempt to `tagsave` for inappropriate file formats, like `.csv`.

- `produce_or_load` now will not attempt to `tagsave` for inappropriate file formats, like `.csv`.

# 2.0.3
* Added a kwarg `equals` to `savename` to extend applicability.

- Added a kwarg `equals` to `savename` to extend applicability.

# 2.0

## Breaking
* DrWatson has moved entirely on using JLD2.jl instead of BSON.jl for saving files, and also suggests the same to its users through the documentation.
* `tmpsave` now saves as JLD2.jl and requires loading explicitly the `"params"` field.
* `collect_results` saves as JLD2.jl by default now.
* `produce_or_load` saves as JLD2.jl by default now.
* `savename` will no longer replace a floating point number with its integer version, if they coincide with respect to rounding. Thus, "integer" `AbstractFloat`s will always end with `.0` in `savename`.
* `savename`'s `scientific` keyword has been replaced by `sigdigits`, as this was the only thing it was doing.
* `savename`'s default options now have `sigdigits = 3` instead of `digits = 3`.

- DrWatson has moved entirely on using JLD2.jl instead of BSON.jl for saving files, and also suggests the same to its users through the documentation.
- `tmpsave` now saves as JLD2.jl and requires loading explicitly the `"params"` field.
- `collect_results` saves as JLD2.jl by default now.
- `produce_or_load` saves as JLD2.jl by default now.
- `savename` will no longer replace a floating point number with its integer version, if they coincide with respect to rounding. Thus, "integer" `AbstractFloat`s will always end with `.0` in `savename`.
- `savename`'s `scientific` keyword has been replaced by `sigdigits`, as this was the only thing it was doing.
- `savename`'s default options now have `sigdigits = 3` instead of `digits = 3`.

## Non-breaking
* `savename` now allows a keyword `val_to_string` for customization.

- `savename` now allows a keyword `val_to_string` for customization.

# 1.18.3
* Remove type constraints on `produce_or_load` path argument (#229)

- Remove type constraints on `produce_or_load` path argument (#229)

# 1.18.2
* Keyword `storepatch` now exists in `produce_or_load` as well.

- Keyword `storepatch` now exists in `produce_or_load` as well.

# 1.18.1
* Fix `dict_list_count` returns wrong number of elements with `@onlyif`. (#223)
* Fix `gitpatch` for git versions not supporting submodules. (#224)

- Fix `dict_list_count` returns wrong number of elements with `@onlyif`. (#223)
- Fix `gitpatch` for git versions not supporting submodules. (#224)

# 1.18.0
* Two new functions `tostringdict, tosymboldict` for changing dictionary key types.

- Two new functions `tostringdict, tosymboldict` for changing dictionary key types.

# 1.17.0
* `produce_or_load` now supports passing function as the first argument to support the do-block syntax.

- `produce_or_load` now supports passing function as the first argument to support the do-block syntax.

# 1.16.6
* `@onlyif` now doesn't expand vector arguments if it's placed inside another vector. (#209)
* `@onlyif` now supports chaining. See real world examples: "Defining parameter sets with restrictions". (#210)

- `@onlyif` now doesn't expand vector arguments if it's placed inside another vector. (#209)
- `@onlyif` now supports chaining. See real world examples: "Defining parameter sets with restrictions". (#210)

# 1.16.5
* The patch information for a dirty repository now also contains the diff for submodules.

- The patch information for a dirty repository now also contains the diff for submodules.

# 1.16.1
* `dict_list` now retains the value's type from the passed dictionary.

- `dict_list` now retains the value's type from the passed dictionary.

# 1.16.0
* Add a `sort` option to `savename`.

- Add a `sort` option to `savename`.

# 1.15.0
* Better default readme file for a new project.

- Better default readme file for a new project.

# 1.14.6
* new keyword argument `rpath` for `collect_results!` that allows storing relative paths

- new keyword argument `rpath` for `collect_results!` that allows storing relative paths

# 1.14.0
* New macro `@onlyif` that allows placing restrictions on values in a dictionary used for expansion with `dict_list`

- New macro `@onlyif` that allows placing restrictions on values in a dictionary used for expansion with `dict_list`

# 1.13.1
* `gitpach` now shows a warning if `git` was not found in PATH

- `gitpach` now shows a warning if `git` was not found in PATH

# 1.13.0
* `savename` now includes `TimeType` (dates) in the default allowed types.

- `savename` now includes `TimeType` (dates) in the default allowed types.

# 1.12.0
* `wsave/wload` now support keyword arguments.

- `wsave/wload` now support keyword arguments.

# 1.11.0
* Macros `@pack!, @unpack` are re-exported by DrWatson.

- Macros `@pack!, @unpack` are re-exported by DrWatson.

# 1.10.0
* New function `struct2ntuple` that converts a struct to a NamedTuple (for saving)

- New function `struct2ntuple` that converts a struct to a NamedTuple (for saving)

# 1.9.0
* `savename` now has the `ignore` option.

- `savename` now has the `ignore` option.

# 1.8.0
* `@quickactivate` was enhanced to allow projects that also represent a module.
* `initialize_project` no resolves the folder name for naming the project if the path is given as "." or ".."

- `@quickactivate` was enhanced to allow projects that also represent a module.
- `initialize_project` no resolves the folder name for naming the project if the path is given as "." or ".."

# 1.7.0
* Improve the introductory file created by DrWatson.

- Improve the introductory file created by DrWatson.

# 1.6.2
* `@tag!` and `@tagsave` now support using `;` as keywords separator (#111)

- `@tag!` and `@tagsave` now support using `;` as keywords separator (#111)

# 1.6.0
* `quickactivate` doesn't do anything anymore if you try to activate to already active project.
* New macro `@quickactivate`

- `quickactivate` doesn't do anything anymore if you try to activate to already active project.
- New macro `@quickactivate`

# 1.5.0
* Started to add support for overloading save/load for custom files. See the updated docs around `wsave` and `wload`.

- Started to add support for overloading save/load for custom files. See the updated docs around `wsave` and `wload`.

# 1.4.1
* Fix a bug that created incompatible version strings in generated `Project.toml` files on release candidate versions of Julia.

- Fix a bug that created incompatible version strings in generated `Project.toml` files on release candidate versions of Julia.

# 1.4.0
* `savename` now supports rounding to significant digits with the keyword argument `scientific`, where `scientific` defines the number of significant digits.

- `savename` now supports rounding to significant digits with the keyword argument `scientific`, where `scientific` defines the number of significant digits.

# 1.3.0
* `initialize_project` now adds a Julia version under `compat` in the created `Project.toml` when it is called.
* The functions `tag!, tagsave` and their respective macros now obtain their arguments (besides the first two) as keywords instead of positional arguments. The positional versions are deprecated (#93).
* New keyword `force = false` for `tag!` and co. which replaces the existing `gitcommit` field.

- `initialize_project` now adds a Julia version under `compat` in the created `Project.toml` when it is called.
- The functions `tag!, tagsave` and their respective macros now obtain their arguments (besides the first two) as keywords instead of positional arguments. The positional versions are deprecated (#93).
- New keyword `force = false` for `tag!` and co. which replaces the existing `gitcommit` field.

# 1.2.0
* Improved behavior of `savename` with respect to nested containers. If a nested container is empty, it is not printed instead. For example, `T=100_p=()_x=2` now becomes `T=100_x=2`. (if `p` is not empty then it is expanded as usual)

- Improved behavior of `savename` with respect to nested containers. If a nested container is empty, it is not printed instead. For example, `T=100_p=()_x=2` now becomes `T=100_x=2`. (if `p` is not empty then it is expanded as usual)

# 1.1.0
* `initialize_project` no longer makes a test directory.

- `initialize_project` no longer makes a test directory.

# 1.0.1
* Allow `tag!` and derivatives to handle dictionaries with *key type* `Any`.

- Allow `tag!` and derivatives to handle dictionaries with _key type_ `Any`.

# 1.0.0

First major release (no notable change from 0.8.0).

# 0.8.0
* **[BREAKING]** : The `gitpath` argument used among many functions

- **[BREAKING]** : The `gitpath` argument used among many functions
can now also point to a subdirectory within a git repository.
Previously it had to be the top directory (i.e. the one containing
`.git/`).
* **[BREAKING]** : Slightly changed how `produce_or_load` uses `path` and interacts with `savename`, to better incorporate the changes done in version 0.6.0. `prefix` is now also supported.
* `tag!` and co now also store the git diff patch if the repo is dirty, see `gitpatch` (#80).
* **[BREAKING]** : `tag!` now saves the commit information into a field `gitcommit` instead of just `commit`.
- **[BREAKING]** : Slightly changed how `produce_or_load` uses `path` and interacts with `savename`, to better incorporate the changes done in version 0.6.0. `prefix` is now also supported.
- `tag!` and co now also store the git diff patch if the repo is dirty, see `gitpatch` (#80).
- **[BREAKING]** : `tag!` now saves the commit information into a field `gitcommit` instead of just `commit`.

# 0.7.1
* `projectdir()` now warns if no project (other than the standard one) is

- `projectdir()` now warns if no project (other than the standard one) is
active

# 0.7.0
* New macro `@savename` that is a shortcut for `savename(@dict vars...)`
* New function `gitdescribe` (see below)
* **[DEPRECATED]** `current_commit()` has been deprecated and replaced by

- New macro `@savename` that is a shortcut for `savename(@dict vars...)`
- New function `gitdescribe` (see below)
- **[DEPRECATED]** `current_commit()` has been deprecated and replaced by
`gitdescribe()` which now replaces the output of `git describe` if an
annotated tag exists, otherwise it will return the latest commit hash.

# 0.6.0
* **[BREAKING]** Reworked the way the functions `projectdir` and derivatives work (#47, #64, #66). Now `projectdir(args...)` uses `joinpath` to connect arguments. None of the functions like `projectdir` and derivatives now end in `/` as well, to ensure more stability and motivate users to use `joinpath` or the new functionality of `projectdir(args...)` instead of using string multiplication `*`.
* New function `parse_savename` that attempts to reverse engineer the result of `savename`.

- **[BREAKING]** Reworked the way the functions `projectdir` and derivatives work (#47, #64, #66). Now `projectdir(args...)` uses `joinpath` to connect arguments. None of the functions like `projectdir` and derivatives now end in `/` as well, to ensure more stability and motivate users to use `joinpath` or the new functionality of `projectdir(args...)` instead of using string multiplication `*`.
- New function `parse_savename` that attempts to reverse engineer the result of `savename`.

# 0.5.1
* Improvements to `.gitignore` (#55 , #54)

- Improvements to `.gitignore` (#55 , #54)

# 0.5.0

This release has **breaking changes**.
* Adjusted return value of `produce_or_load` (#52). It now always return the file and the path it is saved. If `loadfile = false` it returns `nothing, path`.
* The functionality of `default_prefix` has been modified (#51). Now there is a nice interplay between defining a `default_prefix` *and* passing a prefix to `savename`. They are merged like `joinpath(prefix, default_prefix)`. This is valid only when `default_prefix` has a value other than `""` (the default).

- Adjusted return value of `produce_or_load` (#52). It now always return the file and the path it is saved. If `loadfile = false` it returns `nothing, path`.
- The functionality of `default_prefix` has been modified (#51). Now there is a nice interplay between defining a `default_prefix` _and_ passing a prefix to `savename`. They are merged like `joinpath(prefix, default_prefix)`. This is valid only when `default_prefix` has a value other than `""` (the default).

# 0.4.0
* Add expand functionality to `savename`, which handles better containers with nested containers (#50)
* `produce_or_load` now allows the possibility of not loading the file
* New function `struct2dict` that converts a struct to a dictionary (for saving)

- Add expand functionality to `savename`, which handles better containers with nested containers (#50)
- `produce_or_load` now allows the possibility of not loading the file
- New function `struct2dict` that converts a struct to a dictionary (for saving)

# 0.3.0
* Added `test` as a directory of the default project (#43)
* Added `tmpsave` functionality: save the result of `dict_list` in temporary files and conveniently work with sequential clusters (#45)
* Now all saving related functions of DrWatson first `mkpath` of the path to save at and then save (#45)

- Added `test` as a directory of the default project (#43)
- Added `tmpsave` functionality: save the result of `dict_list` in temporary files and conveniently work with sequential clusters (#45)
- Now all saving related functions of DrWatson first `mkpath` of the path to save at and then save (#45)

# 0.2.1
* Improve type-stability of return value of `dict_list` (#41)

- Improve type-stability of return value of `dict_list` (#41)

# 0.2.0
* Changed `path` and `projectpath` arguments of various functions (e.g. `tagsave`, `current_commit`) to `gitpath` universally.
* make keyword arguments of `tagsave` positional arguments instead (to work with the macros)
* Added two new macros: `@tag!` and `@tagsave`: these do the same thing as `tag!, tagsave` but in addition are able to record both the script name that called them as well as the line of code that they were called at.

- Changed `path` and `projectpath` arguments of various functions (e.g. `tagsave`, `current_commit`) to `gitpath` universally.
- make keyword arguments of `tagsave` positional arguments instead (to work with the macros)
- Added two new macros: `@tag!` and `@tagsave`: these do the same thing as `tag!, tagsave` but in addition are able to record both the script name that called them as well as the line of code that they were called at.

# 0.1.0

This is the first beta release! Changelog is kept with respect to here!
Loading

0 comments on commit 962c02f

Please sign in to comment.