Minor changes to help documentation.
Only a month after 2.0.0, this is a quick update that fixes a few bugs, adds a couple of enhancements, and has quite a large revision to the newest function subset_rate.
==================================== New =======================================
calc_rate function can now be used to calculate background rates for use in adjust_rate as the by adjustment input. In other words, the by input is no longer limited to using only calc_rate.bg objects (or numerics). While using a combination of subset_data and calc_rate.bg is the recommended way of determining background rates, there are circumstances when getting a rate via calc_rate might be easier (of course, these rates could always previously have been entered manually as a numeric value). See here for further information about this.convert_rate and convert_rate.ft summary table elements in the output now contain all rate regression parameters, data locations, adjustments (if applied), units, and more. This makes it much easier to save and keep track of all relevant data regarding rate results. They can be saved by extracting $summary directly from the output, or using summary(object, export = TRUE). Note, summary() only prints a condensed version to the console as the full table is too large to print.=============================== subset_rate() ==================================
subset_rate can now reorder auto_rate results in various ways. This can be useful in several situations. See help docs and vignette on website for more information and examples."rank" method for subsetting based on the $rank column of the summary table. The rank always refers to the original rate ranking or ordering as determined in the original auto_rate call, which is retained unchanged regardless of how the results are subsequently subset or reordered."oxygen" method for subsetting based on oxygen values in the raw data. This can be used to constrain results to particular oxygen ranges. Conversely, to exclude particular oxygen values or ranges you can use the oxygen_omit method. See help("subset_rate") for specific details."row_omit" and "time_omit" methods, the n input can now be a numeric vector of any length. Regressions (i.e. rates) fit across any time or row value in n will be omitted. A continuous range of rows or time can still be entered by using regular R syntax for creating vectors such as n = 10:20 or seq().rate, rsq, row, time, and density methods the n input of two values can now be entered in any order.plot input default has been changed to FALSE, and the internal plotting for this function has been revised to be much quicker and show more plots (up to a max of 20 from 9). See help file for more info but briefly, this plotting functionality is intended to provide a quick view of how many rates remain after subsetting and where they occur in the data. In many cases it does not need to be run in every subset_rate call.subset_rate and auto_rate now behave better with objects which contain zero results, and allow piping operations to continue even if an empty object is encountered somewhere in the pipe. This can occur if subsetting criteria excludes every rate. These auto_rate_subset objects with no results now work with print, summary, and mean giving a message that they contain no rates but still printing to the console. subset_rate will now not stop if an empty object is input as x or piped from a previous subset_rate operation. Trying to plot these empty objects in plot or plot_ar and also will not stop any pipes, and will result in a console message but no plot.=============================== General Changes ================================
auto_rate output objects have been rearranged slightly to be more consistent.method = "mean" has not been user-specified, adjust_rate now issues a message if multiple rates have been entered in by confirming the mean value will be used for adjustments.assertthat, ggplot2, cowplot, broom, and gridExtra are no longer dependencies.==================================== Fixes =====================================
adjust_rate - Fix for warnings when using auto_rate objects which have been passed through subset_rate with paired or dynamic adjustment methods.plot_ar - Fix for incorrect subset number appearing in plot titles.plot_ar - Fix for the pos and highlight inputs sometimes failing to produce expected behaviour.===============================================================================
It’s been a long time - over two and a half years (!) since the last update. We are happy to see respR being used by the respirometry community in that time. Amazingly we have picked up over 30 citations, so we are very grateful and happy that so many scientists are finding the package of use.
We are glad to announce that respR has reached version 2.0 and is finally available on CRAN.
v2.0 has been a massive and thorough update where almost every corner of the package has been revised, rewritten, streamlined and tested. In addition, we have added a lot of new functionality which you can read about below.
We have also updated and added even more vignettes, function guides and more on our brand new website detailing all the old and new functionality.
Unfortunately, a major version update inevitably comes with some code breaking changes, and this one is no exception. We thought long and hard about introducing changes that might break code written for v1.1, but made the decision to fix as many as possible of the inconsistencies, bugs, and poor design decisions we made while we had the chance with this major version update. The upside of this is that the package will be much easier to update in the future without breaking existing code.
We have created a page on the new website to detail how v2.0 code has changed and how you may go about revising your earlier code to update it. The good news is most of these changes will be very easy to make. If you have submitted code or are in the process of submitting it as part of a publication, we want this code to remain reproducible long into the future. See here for how you can link to resources that will allow anyone to run v1.1 code and keep these analyses reproducible.
We are happy to say the package has never been more stable and even easier to use, and it is now on a great base upon which to build future functionality. Here are the major changes in this version:
================================== HIGHLIGHTS ==================================
respR is now available on CRAN. It can be installed just like any other package via the RStudio package manager tab or by running install.packages("respR"). If you want to be on the bleeding edge, for example test out the latest dev version, you can use install.github() and the ref input to select a different branch.subset_rate function. This allows auto_rate results to be explored and filtered according to various criteria.plot_ar function. This plots auto_rate objects in a way that visualises how results are distributed within the dataset. Pairs very nicely with subset_rate above.adjust_rate contains several new methods of adjusting for background respiration, including the ability to perform adjustments from paired or concurrent blank chambers, and to perform dynamic adjustments for background rates which change over the course of an experiment.inspect and inspect.ft performs additional data checks: that columns contain numeric data and a check for infinite valuesarea input in convert_rate (accepts "mm2", "cm2", "m2", "km2"). You can also use day as the time metric in output rates (e.g. "mgO2/day/m2")auto_rate "rolling" method. Performs fixed-width rolling regressions across the entire dataset with no ordering of results. This pairs nicely with the new subset_rate function to allows users full control of selection criteria (albeit with fixed width regressions)auto_rate "highest" and "lowest" methods for absolute minimum and maximum of rates regardless of signconvert_DO and convert_rate including percent oxygen saturation, moles of oxygen, and more.convert_val function. A simple conversion function to help with inputting experimental parameters in the correct units.citation("respR") returns the citation to the respR journal publication, including a BibTeX entry for LaTeX users or other applications which accept this.R 4.1 and the new native pipes (|>)R=============================== General Changes ================================
print(), summary(), plot()) revised to produce cleaner, consistent outputs.$rank column to all $summary tables as first column to helps with consistent S3 summary outputs and generally keeping track of summary rows$dataframe for consistency (was previously $df, $data, $dataframe, etc.)df to x where appropriate (auto_rate, inspect, inspect.ft, subsample). This was inconsistent and often confusing, in that df often did not have to be a data frame, but could be a different class of object.inspect_data and pcrit have been removed$call element$inputs elementpar() restoration behaviour.======================== Function specific changes =============================
================================== inspect() ===================================
NEW: Checks that each column contains numeric data. This is the most important check, and if any column fails, the function skips the remaining checks for that column and returns a NULL result
NEW: Checks for infinite values (Inf/-Inf) in all inspected columns
NEW: If uneven time intervals are detected, it now prints the minimum and maximum intervals found in the time data which can help indicate if there is a large gap in the data
NEW: Added a width input to adjust rolling rate plot (default is 0.1)
NEW: An additional data type (for example temperature, salinity, etc.) can be plotted alongside the oxygen timeseries by using the add.data input. This should be a column number from the same input data frame and share the same time data. Data checks are not performed on this column (unless all columns are specifically inspected), it is simply a plotting aid to understand where the parameter may affect rates in different regions of the data. There is similar functionality in inspect.ft.
NEW: Can pass rate.rev = FALSE to plot rates not numerically reversed (for oxygen production respirometry)
NEW: Can pass legend = FALSE to suppress legends
NEW: summary() now works on output (although it’s simply a wrapper for print())
CHANGE: Multiple rate plot improved. Now all plot on same axis range, better colours and spacing.
CHANGE: Prints NA locations for all oxygen columns they are found in (previously only printed for first column)
CHANGE: If multiple columns are inspected, message that columns 1 and 2 are used by default by subsequent functions, although all are inspected.
CHANGE: Stops if oxygen or time columns are not present in the input, or conflict with each other
CHANGE: Prints on assigning
CHANGE: Output has additional elements $call, $inputs and $add.data
CHANGE: Output $list and $list_raw renamed $locs and $locs_raw
FIX: Rolling rate plot failed with even a single NA in time or oxygen data.
FIX: Fix for points sometimes being plotted twice slightly offset in large datasets.
FIX: Any time error locations are now printed when multiple oxygen columns are inspected
================================ calc_rate() ===================================
NEW: Plot updated with title showing the rank of the current plot. It also now plots oxygen and rate against both time and row index
NEW: Now calculates from start of the data if from = NULL, and to the end of the data if to = NULL
NEW: S3 Generics: plot() has the additional inputs panel to plot selected panels individually, and legend to suppress the legend and other labels.
CHANGE: Detects and warns about multi-column inputs
CHANGE: Output elements have been renamed
CHANGE: Most $summary columns have been renamed and reordered to be consistent with other functions
CHANGE: Plots full 4 panel plot by default
FIX: by = "proportion" and by = "oxygen" now work correctly with oxygen production data. Previously these methods didn’t work with rising oxygen as the code for finding the from-to values assumed it was decreasing. Now it identifies all data within the range of from-to. Essentially this makes from-to interchangeable. Note this change means with some data it may give slightly different results than v1.1 because of the way the from-to region is identified internally. With most data any difference in rates should be very minor.
FIX: Fix for failure to plot with very small datasets (2-5 values)
FIX: Fix for failing to work when there were NA in the time data
================================ auto_rate() ===================================
NEW - "rolling" method. Performs simple rolling regression of specified width, with no ordering of results. This pairs nicely with the new subset_rate function to allows users full control of selection criteria (albeit with fixed width regressions)
NEW "highest" and "lowest" methods for absolute minimum and maximum rate values regardless of sign. These can only be used when rates all have the same sign.
NEW "maximum" and "minimum" methods. These work the opposite way to "min" and "max" from v1.1. These are strictly numerical, and order by value taking account of the sign.
NEW: Added the Kernel Density metric $density to the $summary table for the "linear" method. This allows you to see how the results are ranked in regards to KDE (also present for other methods, but as an NA column).
NEW: Added $oxy and $endoxy to the $summary table, so you can easily see roughly how much of an oxygen change each regression is using to calculate a rate. This matches the summary table in calc_rate and calc_rate.ft
NEW: Plot updated with title showing the rank of the current plot. It also now plots oxygen and rate against both time and row index.
NEW: Can pass rate.rev = FALSE to plot rates not numerically reversed (for oxygen production respirometry)
NEW: S3 Generics: plot() has the additional inputs panel to plot selected panels individually, and legend to suppress the legend and other labels.
CHANGE: The "min"and "max" methods have been deprecated. They still work and produce the same results as in v1.1, but have a prominent warning. These will be removed entirely in a later version of respR.
CHANGE: Under the "linear" method, occasionally linear regions which were identical but had different density scores were identified. The lower ranked duplicate results are now removed.
CHANGE: The width input behaviour has changed slightly. Previously if it was a value between 0 and 1 it represented a proportional width of the whole dataset regardless of the by input. This now applies only to by = "row". If by = "time" any value between 0 and 1 now represents (as with values above 1) a time window in the units of the time data.
CHANGE: Reordered inputs to put method second
CHANGE: Summary table reordered to be consistent with other functions
CHANGE: Tidied and reordered summary S3 printout. Summary table comes first, density result printed more compactly for "linear" method
CHANGE: The rolling regression plot (panel 3) is now plotted on the same x-axes of time and row as the full timeseries plot, and has better spacing.
CHANGE: Prints all results of "interval" method regardless of pos or width inputs
FIX: plot() now works correctly on auto_rate objects
FIX: Fix for stopping with an obscure message when input datasets are small
FIX: Removed unused method "default"
FIX: Fix for plots failing when the result contained an NA
FIX: method = "max" was missing $total_regs element from metadata
FIX: Fixed an issue whereby under by = "time" and method = "linear" a non-proportional width (i.e. greater than 1) would get incorrectly passed as a row width rather than a time width. This was only an issue in data where time and row spacing were mismatched, for instance where time was in minutes or hours, or in seconds but not at one second intervals. This would not have affected returned rate values, but instead led to poor ranking of the returned rates.
FIX: Fixed an issue where the function would fail when assembling final output
FIX: Fixed an issue where rates would not be returned over regions which contained NA values. The function is now generally much more robust to missing values (though input data should wherever possible not contain NA values).
=============================== adjust_rate() ==================================
NEW: Completely rewritten to allow several new methods, including dynamic background correction. It has the same original inputs (x, by) plus five additional ones. The method defaults to "mean" so previous code will output the same results as it did in v1.1, but there is a great deal of new functionality.
CHANGE: In the output, $corrected changed to $rate.adjusted, and $input.rate changed to $rate.input
FIX: The adjustment column was missing from $summary table
FIX: summary() now works correctly on objects
FIX: print() prints adjustment correctly if pos > 1
=============================== convert_rate() =================================
NEW: Can now convert to surface area-specific rates via the area input (accepts "mm2", "cm2", "m2", "km2")
NEW: Can now import data and output rates using "day" as the time unit e.g. "mgO2/day/m2"
NEW: Can now output oxygen used in moles e.g. "molO2/day/kg"
NEW: Can now accept input oxygen concentration units in moles e.g. "mol/L", "mol/kg" and percent oxygen saturation ("%Oxy"). Because of this % air saturation is now "%Air" and the previous "%" operator for this has been deprecated.
CHANGE: The o2.unit input has been renamed to oxy.unit
CHANGE: oxy.unit = NULL and time.unit = NULL now stop the function instead of applying a default unit
CHANGE: Now applies a default output.unit for mass- and area specific rates as well as absolute rates
CHANGE: The output.unit oxygen amount component now has “O2” appended to ensure it is not confused with the mass-specific component. For example, "mgO2/h/mg", "umolO2/day/kg". These variations are also accepted as inputs.
CHANGE: output elements have been renamed (e.g. $output to $rate.output, $absolute to $rate.absolute etc.).
FIX: summary() prints summary table as columns correctly
================================== oxy_crit() ==================================
NEW: This replaces the calc_pcrit function from v1.1
NEW: method input. This selects the method to use to determine the critical oxygen value. At present either "bsr" for Broken-Stick or "segmented" for the Segmented method. Previously both were performed.
NEW: thin input. Defaults to 5000. This applies only to the Broken-Stick regression analysis. If the dataset is longer than this, this determines the number of rows it is uniformly subsampled to before running the analysis. The BSR method is quite computationally intensive, so this speeds it up. In testing, values above this have little effect on the result, but this may vary with different data. To prevent any subsampling and use the entire dataset enter this as NULL.
NEW: S3 Generics: plot() has the additional inputs legend, quiet, rate.rev and panel.
NEW: S3 Generics: summary() allows results to be exported via export = TRUE.
CHANGE: parallel = FALSE is now the default. Use parallel = TRUE to use parallel processing if your datasets are particularly large and the function is taking too long to process.
CHANGE: Uptake rates are now plotted as negative in the rolling rate plot on a reversed y-axis. Plots look the same, but y-axis values are now negative. There is a rate.rev input to control this, for instance if you are using the function with existing positive rate values.
CHANGE: An increase to the internal thin value (see above) from 1000 to 5000 may mean results will differ slightly from the previous version, but should be more accurate.
CHANGE: Output has been completely restructured and elements renamed. Previous code used to extract results will likely now not work.
FIX: Fixed issue which caused a failure if the package wasn’t loaded but function was called, i.e. using respR::oxy_crit()
FIX: Fix for plot failure when results were very close together.
================================ import_file() =================================
NEW: Imports European style formatted numeric data which uses commas instead of points for decimals.
NEW: Support for Presens Datamanager files
CHANGE: Importing now generally does not remove any columns (e.g. empty or NA columns). Column naming is now much improved (e.g. unique names, data type linked to channel IDs, etc.).
FIX: Fix for PreSens Oxyvew csv and txt files not being imported as numeric. Now separate parsing for each.
FIX: Fix for MiniDot files not being imported as numeric.
FIX: Fix for Pyro/Firesting numeric columns being imported as character when Firesting software replaces missing data with “—”
================================ format_time() =================================
CHANGE: Numeric time column name changed from elapsed to time.num
FIX: Fix for not appending new numeric time column onto a data.table
FIX: Fix for data frame columns being renamed on return
FIX: Fix for not working with multiple date-time columns if one of them wasn’t column 1
FIX: Fix for failing to format date-times which were already in POSIX/POSIX.ct format
================================ convert_val() =================================
convert_DO and convert_rate. See Examples.=============================== calc_rate.bg() =================================
NEW: The oxygen timeseries is now plotted against both time and row index
NEW: Plots have legend input to suppress equation text boxes
CHANGE: In the output $bgrate changed to $rate.bg, and $mean changed to $rate.bg.mean
CHANGE: Stops if oxygen or time columns not present in the input, or conflict with each other
CHANGE: Removed “input detected” message at start
================================ convert_DO() ==================================
NEW: Added support for % oxygen saturation ("%O2") units. % air saturation is now "%Air" and the previous "%" operator for this has been deprecated
NEW: Added support for oxygen concentration units of mol/L and mol/kg
NEW: Added support for oxygen concentration units of cm3/L. This is a scarce unit, only used in older publications and is equivalent to ml/L.
NEW: Added simplify input. The default is simplify = TRUE and means the converted values are output as a numeric vector, rather than within a list object. If FALSE the output is a list object of class convert_DO which works with print(), summary(), and mean()
CHANGE: Removed ability to accept objects of class calc_rate, auto_rate or adjust_rate. This didn’t make any sense in this function as rates are not a DO measure!
CHANGE: Added warning for P atmospheric pressure values being outside a realistic range.
FIX: Fix for error with non-numeric x inputs
================================ unit_args() ===================================
================================= subsample() ==================================
length.out input which uniformly subsamples to an exact number of rows (data frames) or length (vectors)=============================== subset_data() ==================================
NEW: Support for inspect.ft objects
NEW: Added quiet input to allow console output to be suppressed
NEW: Now subsets from start of the data if from = NULL, and to the end of the data if to = NULL
CHANGE: Updated help documentation, particularly to specify that by = "oxygen" or by = "proportion" applies to first column only in multi-column dataframes
CHANGE: Warns if output subset is empty
FIX: Fix for failure with inspect() objects
FIX: by = "proportion" and by = "oxygen" now work correctly with oxygen production data. Note this now generally results in longer subsets than v1.1 code because of a change in identifying where to is within the data.
============================== flowthrough data ================================
NEW: inspect.ft
NEW: calc_rate.ft
NEW: adjust_rate.ft
NEW: convert_rate.ft
This is a completely new workflow specific to flowthough respirometry data. See help documentation and vignettes for full details.
==================================== data ======================================
data.tableflowthrough.rd amended to make delta column (i.e. unitless rates) negative values to be consistent with how uptake rates are always negative in the packageflowthrough.rd column names changed to time, oxy.out, oxy.in and oxy.deltazeb_intermittent.rd column names changed to Time and Oxygensardine.rd has had a Temperature column added (to demonstrate the new inspect() functionality to plot an additional data type)NEW: New example datasets
algae.rd: oxygen production dataflowthrough_mult.rd: multiple column flowthrough dataflowthrough_sim.rd: flowthrough data with increasing background recordingbackground_con.rd: constant level background, for testing new background correction functionalitybackground_lin.rd: linear increasing background, for testing new background correction functionalitybackground_exp.rd: exponential increasing background, for testing new background correction functionality================================= S3 Methods ===================================
mean.adjust_rate
mean.adjust_rate.ft
mean.auto_rate
mean.calc_rate
mean.calc_rate.bg
mean.calc_rate.ft
mean.convert_rate
mean.convert_rate.ft
NEW: These are new generic S3 functions for calculating mean rates, and work on the relevant function output rate. None alter the output object, but have an export input so the mean values can be exported as a numeric value if export = TRUE. They also have a pos input (where relevant) to select a range of rates to average. Most useful is mean.convert_rate which can be used to output a final mean rate at the end of an analysis.
mean.convert_DO
NEW: This works as the above but averages converted oxygen values from convert_DO. The output must have been saved as a convert_DO object for this to work by using simplify = FALSE (otherwise convert_DO by default outputs a numeric vector upon which mean() works anyway and gives the same result).
NEW: All plot() S3 have a quiet input to suppress console output
NEW: All summary() S3 (where appropriate) have export = TRUE for export of summary table or value
NEW: All S3 (where appropriate) have pos input to select positions or ranks of results to plot, print or average, and stop with an error if pos is too high
NEW: Plotting functions accept the generic par() inputs oma, mai, tck, mgp, las, and pch via ... to allow default parameters to be changed. Particularly useful are las = 1 to make axis labels horizontal and adjusting the second (left side) of the four mai (inner margins) input to make y-axis labels more readable.
CHANGE: Consistently named inputs (for summary() the first input should be object, but in all others it is x)
CHANGE: All console output prints more consistently with nicer spacing, especially in pipes
FIX: In auto_rate S3 methods stop with a message if no rates found in object (this can happen with over-enthusiastic subsetting in the new subset_rate function)
FIX: All revised to work correctly with |> or %>% pipes
This is the final release of v1 of respR.
There is only one change, the addition of function called installation_help. This function simply opens its own help file. It contains resources which may be helpful in running v1 code in the future, such as a list of dependencies and the latest versions known to work with v1.1.1, and how to install them. This is intended for users who have recently submitted v1.1 code as part of a publication, and investigators who may wish to reproduce their analyses in the future.
The reason this may be necessary is because version 2.0 of respR will be released shortly and contains several code-breaking changes which means code written using v1.x will fail to run or return the same results. Using the information in this file an investigator should be able to install respR v1.1.1 and its dependencies and reproduce these analyses.
Lots of updates! First, we have a new function, calc_pcrit(), which we will use to include new methods in the future. We’ve also made some improvements to import_file() with increased support to more files from Vernier, PRESENS and Loligo systems.
Vignettes have been migrated into a new repository on GitHub, and are now updated separately. This gives us several advantages – the documentation can be updated without needing to compile a new package, and the size of the package is significantly reduced. Users can also contribute to the documentation more easily now that it is not tied to the package.
respR by running citation("respR").respR will print a startup message containing links to our published manuscript and vignettes.inspect() can now plot data with multiple columns when more than 2 columns are detected. For some people, this may provide a great overview of the data.calc_pcrit() is a new function that will be developed in parallel to the current pcrit() function. In the future we intend to use calc_pcrit() to incorporate new methods.import_file() supports even more files! Added more parsers for Vernier, PRESENS and Loligo systems.format_time() can now calculate time elapsed even if date information is not provided (e.g. “HMS”-only data).by input in auto_rate(), calc_rate(), calc_rate.bg() and subset_data() is now more forgiving with string input values (e.g. "oxygen", "Oxygen", "o2", "O2", etc. are recognised). To achieve this we created a string matching function, verify_by() which uses brute force matching to recognise different ways of writing the same text.convert_DO():
"P" (pressure) was ignored even when a value was already specified, resulting in the default used. The function will now respect the "P" value provided.inspect():
convert_rate():
"mass" input is provided but "output.unit" is not a mass-specific unit.calc_rate.bg.output.unit in the code.calc_rate.ft():
"time" input has been removed as it was not used.inspect.pcrit():
inspect and inspect_data. Please note that the inspect_data() function is deprecated and will be removed in the near future.This is a quick fix for images not showing in the online vignettes.
This version is aimed at improving the functionality and usability of our conversion functions. We have decided to remove the default values for temperature and salinity inputs in convert_DO() and convert_rate() (i.e. switched t and S numerics to NULL). This was no easy decision, but we noticed that some users were running the functions by default without considering (or even knowing the existence of) these two important input variables. With this change, existing workflows using calc_DO() and calc_rate()are likely to break. Do note that this is a very rare modification – we know that changing core functionality that breaks prior code is not something to be taken lightly. Please update these two functions respectively.
convert_DO() and convert_rate() now require the user to explicitly provide t (temperature) and S (salinity) values.convert_DO() and convert_rate() will warn the user when the default P value is used for O2 units that are strongly influenced by atmospheric pressure.unit_args() has been updated to indicate which O2 units need inputs of t, S and P.inspect() has a better-looking plot for rolling regressions.This version updates the new functions from 1.0.0 and tries to improve piping workflows.
import_data() was not importing firesting files properly.subset_data() - objects produced by this function can now be immediately passed on to other functions. subset_data() is a lesser known function but it’s very useful, e.g. for truncating intermittent data before analyses. It’s been updated work well with other functions, and can also be piped.import_data() - works with a lot more files. Now supports: Firesting Logger | Pyro Oxygen Logger (also Firesting) | PRESENS OXY10 | PRESENS (generic) | MiniDOT | Loligo Witrox Logger | Loligo AutoResp (software output). Get more files to us, people!auto_rate().The big update! Maybe it’s time to submit to CRAN?
dplyr and magrittr are now imported into respR.import_file(): automatically import and format raw files from common devices. This function is not complete but needs to be put out there for user feedback. It’s stable, just doesn’t support all available machines yet. We expect this function to be improved dramatically once users send in feedback and sample files.dplyr pipes now work after print() and summary() commands. Will add support for plot() depending on feedback.adjust_rate(): minor print issue (new line required) that was messing with the console output.adjust_rate() documentation has been updated.inspect() should now work with pipes.inspect() had issues with data.table objects, so we switched to data.frame.calc_rate.bg() did not subset data properly in certain usage scenarios (see #49)We are getting close to a 0.1 release, which will be ready for CRAN.
format_time(), which fixes some errors with zeroing in some numeric data.format_time(). This function is a wrapper for lubridate functions and is used specifically to convert date-time to numeric time.inspect() that contains much cleaner code. The old function inspect_data() was tough to fix. It will be deprecated in a future update, but will still be supported for at least 18 months after we announce the deprecation. For now, users need not worry about using either function.calc_rate(), calc_rate.bg(), calc_rate.ft(), and auto_rate() are updated to work with inspect(). (TODO: calc_rate.ft() still needs to be fixed.)auto_rate() is now available.auto_rate()’s linear detection method is now available.sim_data(): simulate data for benchmark analyses.test_lin(): perform benchmark analyses specific for auto_rate()’s linear detection (i.e. methods = "linear").auto_rate() fixed (it was messing with print() commands).convert_rate() fails when output contains “mmol”. Fixed.pcrit() example on website.subset_data() now contains examples.inspect_data() plot output was plotting only columns 1 and 2 by default, even when other columns are selected in inputs.convert_rate() was not accepting objects of class calc_rate.ft.