Alto Huambuno

• Part 1: Fixed bug that prohibitted processing files in subfolders.
• Part 1: Dynamic range can now be extracted from Axivity .cwa files when not equal to 8g.
• Part 2: Qwindows logically ordered in reports.
• Part 2: Fixed bug #516 regarding passing on of dateformat.
• Part 4: Fixed bug introduced in 2.6-0 relating to sleeplog guider being assined even when no sleeplog used.
• Part 5: Improved handling of missing sleep estimate for first night.
• General: Fix bug introduced in 2.6-0 relating to storage of configuration file on comma-separated machines
• General: GGIR reports are now saved with a maximum of 3 decimal places for numeric variables.

Ibarreta

• Part 1: Fixed warning when working with a vector of filenames as argument to datadir.
• Part 2: Added argument max_calendar_days to allow user to control the maximum number of calendar days to include in analysis, irrespective of argument strategy value.
• Part 2: Argument maxdur is no longer specific to strategy 1, but applied irrespective of value of argument strategy.
• Part 4: Fixed issue that sleeplog is indicated as the guider even when when corresponding timestamp cannot resolve to timestamp.
• Part 5: Time series output filenames now auto-stripped from [.]|rdata|csv|cwa|gt3x|bin character string.
• Fixed issues introduced in 2.5-6 release: maxNcores was not defined in part 1, params_output and not passed on to part 3, and params objects were unneccesarily repeatedly checked.

What's Changed

• Issue497 minorissues by @vincentvanhees in #498
• Issue499 move spo to df by @vincentvanhees in #500
• Add optional max_calendar_days argument by @vincentvanhees in #503
• Issue504 handling invalid sleeplog entries by @vincentvanhees in #505

Full Changelog: 2.5-6...2.6-0

Nueva Loja

Changes:

• Internal re-structuring: Arguments are now internally passed via parameter objects to make code and code documentation easier to navigate. This new structure preserve backward compatibility with the user's older R scripts.

Note: I have put this in a separate release from other recent changes in the previous release (2.5-5) earlier today, to help isolate the cause of issues if and when they arise.

Nueva Holanda

• Fixed bug #484 affecting part 5 report not being able to generated because non-matching columns in milestone data
• Part 2: Fix bug, BrondCounts and ZeroCrosssing Counts now no long auto-scaled by 1000 like all the other metrics.
• Part 3: Fixed warning when SRI is calculated on DST day with 23 hours.
• Part 5: sibreport and timeseries now stored per sib definition, previously only for last definition.

Puesto Esperanza

• Part 1: Now able to process modern .gt3x data with thanks to R package read.gt3x.
• Part 1: Timegaps and zeros across all three axes in ActiGraph data (.csv and .gt3x) now automatically imputed.
• Adding BrondCounts as optional acceleration metric and dependency on activityCounts, and enabling Sadeh and Galland algorithms to use it.

Frutillar Alto

• Part 1: Depricating function g.metric as its functionality has been taken over by g.applymetrics.
• Part 2: Adding warning when ID cannot be extracted from file based on argument idloc.
• Part 4: Empty or incomplete sleeplog rows now better ignored
• Part 4: No longer an R warning is given when ID is missing in sleeplog, because this is common and user can already see it in csv reports.

La Guamas

• Part 3: Now makes sure that HASPT is skipped when user configures def.noc.sleep as a set time window. This
  feature that probaly few people use nowadays broke with the 2.5-0 release.
• g.shell.GGIR: Now gives warning when user supplies double arguments.
• Part 4: Now warns when none of the IDs in the sleeplog could be matched with accelerometer data.
• Part 1: read.myacc.csv fix bug with argument rmc.check4timegaps
• Part 3: Fix #472 SRI calculation not possible when complete absence of sleep in recording
• Part 5: Experimental nap detection added to report and time series, currently only model for 3.5 year olds available.

Cachuela Esperanza

Minor fixes:

• g.shell.GGIR: Removing forced assignment of sleepwindowType argument to "SPT"
• g.shell.GGIR: Reviving setwindow option def.noc.sleep, which broke

Ranqueles

• Vignette expanded with documentation on non-default variables in part 4 csv report
• Vignette added with a tutorial on how to do segmented day analysis
• Part 1: The sensor fussion functionality introduced in version 2.2-2, as a bit
  of an experimental development, has now been removed from GGIR as added value turned out
  to be limited.
• Part 2: ID extraction code tidied up, and new idloc argument options added.
• Part 2: Activity log can now also have empty days.
• Part 2: Now also exported in long format if qwindow has length longer than 2.
• Part 3 and 4 expanded with Sleep Regularity Index.
• Part 4 number of awakenings is now also in the output per user request.
• Part 5 LUX per segment calculation bug fixes.
• visualreport code modified such that it visualises any day with at least 30 minutes of data
  unless sleep could not be estimated from the corresponding night.

LasGaviotas

• Part 1 now able to derive zero-crossing counts needed for Sadeh algorithm
• Part 2 function g.conv.actlog to use activity log to guide qwindow now able to tailor dateformat
• Part 3 now option to tailor SPT detection for hip data by using horizontal longitudinal angle as indicator of lying
• Part 3 code restructured to estimate SPT and SIB in separated function (HASIB and HASPT) such that g.sib.det is more readable.
• Part 3 now option to select SIB algorithm, including: vanHees2015, Sadeh1994, and Galland2012
• Part 4 now able to handle more advance sleeplog format that also contain nap and nonwear entries per date
• Part 4 report now includes guider estimates also in 'cleaned' report
• Part 4 WASO and NOA (Number of awakenings) added to output
• Part 4 Able to handle time in bed diaries/algorithms, and generates sleep latency and sleep efficiency estimate if available.
• Part 5 now able to export sustained inactivity bouts and self-reported naps and nonwear to csv report to aid nap analaysis.
• Part 5 numeric timing of MX-metrics changed to be hour of the day
• Part 5 Lux per segment variables now turn LUX during SPT to zero
• Part 4 legend added to visualisation
• Part 5 bug fixed that caused part5 single row output to be stored as a single column