numere - v1.1.7 "Chandrasekhar"

Highlighted changes in this version

UI changes

• The default UI style has been changed to a more bluish theme. Click on "RESET" for the "UI THEME" within the "Style" tab in the settings dialog to load the new default theme (only if you want to change, of course).
• The default icon for files used within the file tree has been changed to appear more neutral

New and improved functionalities

• The function zip() combines multiple arrays elementwise into embedded clusters in the returned cluster
• The function getuserinfo() will return a key-value list containing user information
• The command mail can be used to send e-mails via a SMTP server
• The date() function will now also accept strings as its "formatting type" parameter
• getfileparts() will now return the host as drive for UNC paths, i.e. "\\MY.SERVER.TDL"
• String functions, which convert one input into an array, do now return arrays of clusters (one for each input element). Notable candidates: textparse(), to_value(), split() and all character classifier functions like is_alnum()
• Functions for base-n encoding encode_base_n() and decoding decode_base_n() have been added

Experimental features

• Structures were added: dictstruct (a structure/dictionary mixture), file (for arbitrary file accesses), path (for all type of paths), queue (FIFO storage) and stack (LIFO storage)
• readxml() can import XML files and readjson() can import JSON files (both are imported as dictstruct instances)
• The command obj can be used to declare procedure-local objects and structures (including categories)
• The index operator VAR[IDX] can now be used on usual vectors. It also accepts the dimension variable nlen

General changes

• A duration type is now available to increase the date-time logic
• The performance of the new parser implementation was greatly improved and is now at about 70-80% of the previous approach for scalar operations
• Vectorial variables are now supported in matop, but not fully compatible to inline created vectors using the brace syntax {x,y,z,...}, because they will not be auto-expanded to matrices if necessary. To resolve this, either wrap the vectorial variable into additional braces or enclose it into an explicit repmat() call. This is a known issue and will be resolved in a future version together with the complete rework of matop.
• Variable method resolution is much more precise now
• It is now possible to use SQL placeholders in SQL statements and supply the parameters via params=[PARAMLIST]
• HTML exports of tables will now also parse a simplified markdown syntax, so that headlines, code segments, bold and italic text are possible
• Many improvements and fixes

The complete list of changes can be found in the ChangesLog.

Remarks

• This is a stable release and builds ontop of the changes from the last release candidate.
• If you experience troubles with downloads from SourceForge (esp. via MS Edge for Business), consider downloading from Github directly (see assets section at the end of the release notes).

- C++
Published by numeredev 4 months ago

numere - v1.1.7 Release Candidate

Highlighted changes in this version

UI changes

• The starting page was greatly reworked and provides more options.
• Double-clicking on any embedded cluster in the table viewer window will open the embedded cluster in a new table viewer window. Standard variables can now also be viewed in a table viewer. This is especially useful for vectorial variables.
• The table viewer can now sort and filter the displayed table according a selected column.
• You can now change the file filters for each of the default folders in the left tree. The option can be found beneath the path settings in the settings dialog.
• Detection of readable and loadable files has been improved, so that you can now also read files without known extension.
• Table autosizing has been sped up and vertical alignment of cells with empty lines has been improved.
• The method autocompletion recommendation has been improved and should now return better fitting candidates. The autocompletion in the terminal is much faster now.
• Window layouts are now also part of the dependency analysis and will correctly return their event procedures.
• The command install will now automatically search through the package repository for a package with matching filename.
• Large tables won't calculate the complete stats anymore, if opened in the table viewer (this is for performance reasons).

New and improved functionalities

• A possibility to define a post-install script for a package was added to the package creator. The "required version" of packages now uses the correct variant of the current version for checking, e.g. "v.1.1.7.2412". The package creator will now log the last modification time of each bundled procedure in its project file. Once the project is loaded again, it is checked, whether there have been any file modifications in the mean time.
• It is now possible to set a custom HTTP header and a payload if you want to interact with REST APIs using url.
• The command print can now also display a table's contents (the first and last 5 rows) in the terminal.
• The function convertunit() allows for unit conversions towards SI units (like the corresponding table methods).
• The function getoverlap() returns the overlap of all to be passed intervals.
• The functions complement(), intersection() and union() provide means to apply set theory operations.
• The new table method TAB().replacevals() can be used to replace every occurence of a set of values with new values.
• The option nobars can be used together with hist to switch the output graph to a dashed line plot with distinct marks at the bars' positions. This can be handy if you try to compare multiple histograms in a single diagram.
• Most table methods accept now table column names in addition to column IDs. Note: If there are multiple columns with the same name, in general, all will be used for the table method.
• The target folders used with move or copy can now contain wildcards within their path as well and will be replaced by their counterpart of the copied/moved filename, if it exists.
• Improved and cleaned the stats command: some field were renamed or replaced with more important values.
• The automatic type conversion now avoids automatic conversions from floating point to int.
• The function cnt() will now return 0 if void values are inserted as its argument (note that this is not true for an array of void values). Completely empty and uninitialized columns and empty tables will now also return void instead of nan or an empty string.
• The function student_t will no longer subtract 1 from the passed DOF internally.
• The interface for odesolve and fit have been adapted to follow the current syntax approach more.
• The return value of the table method TAB().indexof() has been changed to a cluster, where consecutive index sets are stored as embedded clusters.
• The table method TAB().anovaof() will now return the results of multi-dimensional ANOVAs in distinct embedded clusters.
• Important: The table methods TAB().categoriesof() and TAB().categorize() will now return an array of categories (see function category() for reference) instead of a key-value list. This might break some code.

Experimental features

No experimental change in this release

General changes

• The kernel was greatly rewritten and supports much more different types and functionalities. Strange typing problems shouldn't appear any more. Overall performance should be improved (although at the cost that some segments might run slower; we're still investigating)
• We now only provide a 64 bit version due to limitations of external libraries
• We'll now provide an up-to-date PDF documentation with every new release (including this one)
• A Report an issue feature has been added, which you can use even without an GitHub account. You can also check directly from NumeRe, whether a new release is available.
• Many improvements and fixes

The complete list of changes can be found in the ChangesLog.

Remarks

• This is a release candidate. Some features are subject to change. If you experience problems, please let us know
• If you experience troubles with downloads from SourceForge (esp. via MS Edge for Business), consider downloading from Github directly (see assets section at the end of the release notes).

- C++
Published by numeredev 8 months ago

numere - v1.1.7 Beta

Highlighted changes in this version

NOTE: This release is an actual beta release. Be carefule, if you use this version. This contains a hotfix for a problem with English localizaton.

UI changes

• The table viewer was improved and does now provide a categoricaly conditional cell colouring scheme
• The terminal history has now a search bar
• It is now possible to hide the main window using set windowshown=false and unhide it again. Using set appautoclose=true, the app will automatically close itself, once the last open window has been closed. Note that those settings are not saved and only affect NumeRe for the current running session.

New and improved functionalities

• remove may now also remove folders (and folder structures recursively)
• k means clustering is now available as table method TAB().kmeansof(...)
• New functions: uuid(), strjoin(), getdisplayscale(), is_void(), is_equal(), is_ordered(), is_unique()
• New statistical functions: rms(), skew(), exc(), stderr(), inv_pct()
• New time-specific functions: today(), get_utc_offset(), is_daylightsavingtime(), is_leapyear()
• New noise functions: ridgedmulti(), billownoise(), voronoinoise()
• Functions for converting between different numerical data types and for creating categories were added.
• NumeRe can now interact with databases using the command database. Available interfaces for now are: MySQL/MariaDB, SQLite, ODBC. PostGreSQL is NOT supported right now due to dependency issues.
• Table method updates: Table method modifiers .scwin() and .srwin() have been added and can be used to define a moving window for multi argument functions. Table method handling has been improved and chaining with the new internal method system is now possible. The new table method modifier .cells() together with .rows or .cols allows for selection of table cells along the direction of application.
• Table columns can now have (physical) units to represent data even better. The units will be detected automatically (if possible), or you can use the dedicated methods to modify the unit of each column.
• Tables may now be exported in MarkDown and HTML format and it's now also possible to copy only a part of a table to another table at a desired location using TAB(i1:i2,j1:j2) = TAB2(i1:i2,j1:j1) syntax.
• Local variables may now also be declared within control flow blocks. Note that they are only declared for the first time per execution the control reaches their command.
• Index-based for loops can now have an additional condition as their second argument (i.e. for (i = 1:0, a !=b)), which is checked before each loop iteration.
• The interaction between custom windows and the window command has been cleaned and improved. Some interactions will now return more fitting data types. It might be possible that this change will break some existing solutions. Furthermore, the textfield can now be configured as type=markup to support simplified markdown styling.
• The tabs created with group -style=tabs ... in window layouts can now have an ID and an onchange event handler firing when the user changes the tab. It is also possible to access the properties of the tab group using window ... -get and window ... -set.
• The window command option options={KEY-VALUE-LIST} can be used to modify additional window element properties. This option is currently only supported for tablegrid elements, where one can change the minimal number of rows and cols.

Experimental features

• Clusters can now be created as hierarchical data structures, i.e. clusters can now contain clusters in their fields. You can use the .sel(ID) method for navigating and .unwrap to unwrap the whole structure in a single-level cluster. Writing a set of values into a single cluster element will embed them as a sub-structure. Reading this single element, will return the embedded structure.

General changes

• The kernel was greatly rewritten and supports much more different types and functionalities. Strange typing problems shouldn't appear any more. Overall performance should be improved (although at the cost that some segments might run slower)
• We now only provide a 64 bit version due to limitations of external libraries
• We'll now provide an up-to-date PDF documentation with every new release (including this one)
• A Report an issue feature has been added, which you can use even without an GitHub account
• Many improvements and fixes

The complete list of changes can be found in the ChangesLog.

Remarks

• This is a beta version. Some features are subject to change. If you experience problems, please let us know
• If you experience troubles with downloads from SourceForge (esp. via MS Edge for Business), consider downloading from Github directly (see assets section at the end of the release notes).

- C++
Published by numeredev about 1 year ago

numere - v1.1.6 "Casimir"

Highlighted changes in this version

UI changes

• An optional feature to extract the argument types of functions, methods and procedures based upon their names, was added.
• Usability of mouse-over tooltips was improved. It should now be possible to show the fulltips of arguments inside of function parentheses again.
• The auto-indentation functionality does not interfere with CTRL-Z any more.

New and improved functionalities

• The new plotting option stacked enables the stacking of bars and hbars on each other.
• Bar charts now always include the value y=0. In some situations it is reasonable to enable origin=sliding as an additional plotting option.
• One can now pass custom axis ticks using a list of strings instead of the newline-separated strings.

Experimental features

• It is now possible to create new columns in tables if values are assigned to an unknown column using string indexing with the column headlines, i.e. TABLE(:, "MY-COL-HEADLINE"). The new column will be appended at the end of the table.
• matop functions now also accept matrix results as their scalar inputs. For consecutive scalar inputs, even a single vector can be passed, e.g. matop zero({1,3})
• We added support for coloured window themes. If you want to test it, the setting can be found at the end of the syntax colour list. Recommended colour sets are the blue set (foreground: rgb(159,213,244) and background: rgb(234,245,253)), the light grey set (foreground: rgb(82,82,82), background: rgb(181,183,185)) and the splash-image set (foreground: rgb(214,177,182), background: rgb(222,237,234))

General changes

• Increased performance of stfa, audioread and some statistics functions.
• write now uses noquotes as default. The parameter was removed and has no effect any more.
• If the values nan and inf are used within an index vector, then they are treated as invalid values and will always return an invalid value.
• Many improvements and fixes

The complete list of changes can be found in the ChangesLog.

Remarks

• This is a stable release and builds ontop of the changes from the last release candidate.
• If you experience troubles with downloads from SourceForge (esp. via MS Edge for Business), consider downloading from Github directly (see assets section at the end of the release notes).

- C++
Published by numeredev over 1 year ago

numere - v1.1.6 Release Candidate

Highlighted changes in this version

UI changes

• Added feature of autowrapping for comments if line indicator is active.
• Using the new file dialog and selecting "App", one can create a custom app from some predefined templates
• The terminal rendering speed was increased, it provides global variables for autocompletion, autocompletion was improved generally
• The table viewer will now group common table header rows together, creating a virtual hierarchy. Strings are also automatically wrapped around and the grid lines can be hidden
• Folding has been enabled for LaTeX files and a simple lexer for INI files was added

New and improved functionalities

• getfilelist() and getfolderlist() may now follow shell links (*.LNK files)
• getuilang() returns the user's current language, getfilediff() returns the unified diff of two files, the functions to_utf8() and to_ansi() can be used to convert to and from UTF-8 encoded strings
• vectshift() and circshift() perform element shifts in matrices
• Weibull probability functions and probability density functions for all distributions
• Multiple new string functions with corresponding methods: firstch(), lastch(), startswith(), endswith()
• New table methods: TAB().insertXYZ(), TAB().removeXYZ(), TAB().reorderXYZ() (XYZ means rows, cols or cells) and TAB().shrink()
• TAB().convert() now provides an "auto" mode and may now convert more different number formats (notably EU and US separators for decimals and thousands). TAB().fndcols() can now be used with regular expressions
• The "value" column type of tables was extended to provide far more different fundamental data types, e.g. "value.f32" or "value.ui8". Some file types will now be loaded with the best fitting data type
• Loading and saving of regular file types with custom file extensions is now possible
• Procedures can now be called with relative namespaces
• Conditional expressions in if and while, which return multiple values, have now all return values being considered (implictly using and())
• leave extends the functionality of break by jumping out of all nested control flow statements (with exception of procedure ... endprocedure)
• One can create QR codes with qrcode
• More interaction possibilities to interact with and create custom windows, layouts can now be defined dynamically via strings and custom windows may now be used as dialogs.

Experimental features

• Conditional breakpoints (available via right-click in the editor)
• Column headlines are now usable for column indices, i.e. TABLE(:, "MY-COL-HEADLINE")

General changes

• We now provide 32 bit and 64 bit versions. The 64 bit version is likely to be faster and more versatile so switching is greatly encouraged
• We'll now provide an up-to-date PDF documentation with every new release (including this one)
• File loading was improved in many scenarios
• Many improvements and fixes

The complete list of changes can be found in the ChangesLog.

Remarks

• This is a release candidate. Some features are subject to change. If you experience problems, please let us know
• If you experience troubles with downloads from SourceForge (esp. via MS Edge for Business), consider downloading from Github directly (see assets section at the end of the release notes).

- C++
Published by numeredev almost 2 years ago

numere - v1.1.5 "Cherenkov"

Highlighted changes in this version

• It's now possible to define an enumeration using the syntax declare enum -> {VAL1, VAL2, ...}
• The documentation browser can now be opened again using help window and similar. The necessary time for opening up with this method was greatly reduced.
• Empty cells are now preserved, if one converts string columns to logical columns.
• Multiple head rows in tables and data files are now much better supported for XLS, XLSX, ODS and CSV
• Copying and moving (using copy and move) of mixed-value tables is now possible again.
• It's now possible to deactivate the table auto-save by setting the interval to zero
• Downloading of zip files now works as expected.
• The calculation of needed memory per table and cluster does now reflect better the actually needed memory.
• TAB().num now also considers strings as valid values for the num() operation and empty columns are now correctly calculated as empty if used with TAB().cnt.cols
• Improved performance when copying a large table out of the table viewer.
• Procedures execute now always the latest changes without the need of waiting a few seconds
• Plain app names were removed from code and moved to the langauge files (allows also translations of the app name).
• Many improvements and fixes

The complete list of changes can be found in the attached ChangesLog

- C++
Published by numeredev over 2 years ago

numere - v1.1.5 Release Candidate

Highlighted changes in this version

• A large number of table methods have been added: TAB().pcorrof(), TAB().scorrof(), TAB().covarof(), TAB().*rankof(), TAB().zscoreof(), TAB().anovaof(), TAB().fndcols(), TAB().indexof(), TAB().countif(), TAB().binsof(), TAB().categoriesof(), TAB().categorize(), TAB().rows, TAB().cols
• A filter() function was added to matop to filter a matrix with an arbitrary kernel
• A large number of random distributions (Laplace, Cauchy, Rayleigh, Landau, Levy (alpha stable), Fisher-F, Student-T) were added as functions
• Functions to work with time values were added: as_date(), as_time(), weekday(), weeknum()
• sha256() function calculates SHA-256 of strings and files
• to_time() and timeformat() now use automatic pattern matching if no pattern is given
• The range-based for loop has been added as for (ITER -> RANGE) and can also have multiple iterators at once
• archive files can be packed with pack and unpacked with unpack
• diff can now differentiate numerically up to the third order
• hist2d now also supports time axes (experimental)
• vect can now also generate vector plots from complex functions and complex data grids
• With the new plot options ohlc ("Open-High-Low-Close") and candlestick, plots can be generated for the display of price developments (stocks)
• Tables can now have columns with logical values and columns with categories. In addition, the Table Viewer has received further functionalities that further simplify the interaction with tables
• External files are now under automatic version control. Furthermore, restoring a previous version has been improved so that the current file is now automatically overwritten. In addition, a menu option has been added to compare two different files with each other
• It is now possible to disable the "floating" of most windows on the main window
• Procedures in the file tree now show their signature and documentation (if available) as tooltip
• It is now possible to copy text with styling information from the editor to MS-Word or MS-PowerPoint with an appropriate menu option
• It is now possible to drag code with comments into the terminal. The code is executed normally
• When creating a new file, a helpful dialog with more information about the file types now appears
• The support of Excel files and JCAMP-DX files has been improved
• The help command and the documentation browser have been improved and now also show function and procedure documentation
• Toolbar icons have been updated and modernized. Three selectable styles are available in the options (Purist, Focused (default) and Colourful). It is also possible to extend the toolbar over the entire window width
• Many improvements and fixes

The complete list of changes can be found in the attached ChangesLog

- C++
Published by numeredev almost 3 years ago