Batch import with lr_get_spec() and lr_get_metadata()

Hugo Gruson

2020-06-30

lightr provides three main functions for patch import of spectral data and metadata:

Those three functions contain a loop and can directly be used to import/convert whole folders.

They also allow for recursive search in the folder tree with the argument subdir. In this example, the data that contains a subdirectory named procspec_files, which contains :

└──+ data
   ├── avantes_export.ttt
   ├── avantes_export2.trt
   ├── avantes_export_long.ttt
   └──+ procspec_files
      ├── OceanOptics_badencode.ProcSpec
      ├── OceanOptics_Linux.ProcSpec
      ├── OceanOptics_Windows.ProcSpec
      └── whiteref.ProcSpec

We first demonstrate these features on lr_get_spec() but they work in the same way for lr_get_metadata() and lr_spec2csv()

library(lightr)

Import spectral data: lr_get_spec()

lr_get_spec() is one the core functions of lightr. It finds spectral data files, extract the reflectance / transmittance / absorbance data and returns a data.frame where the first column (named wl) contains the wavelengths and the subsequent columns contain the spectral data, interpolated every nanometre:

res <- lr_get_spec(where = "data", ext = "ttt", lim = c(300, 700))
## 2 files found; importing spectra:
head(res)
##    wl avantes_export avantes_export_long
## 1 300         3.1487           13.624678
## 2 301         3.1589            5.276307
## 3 302         3.5700           11.023560
## 4 303         3.9215           10.307208
## 5 304         3.4034            8.980705
## 6 305         3.7878            8.278200

lr_get_spec() also supports setting multiple file extensions at once by passing a character vector to ext:

res <- lr_get_spec(where = "data", ext = c("ttt", "trt"), lim = c(300, 700))
## 3 files found; importing spectra:
## Warning: Could not import one or more files:
## data/avantes_export2.trt
head(res)
##    wl avantes_export avantes_export_long
## 1 300         3.1487           13.624678
## 2 301         3.1589            5.276307
## 3 302         3.5700           11.023560
## 4 303         3.9215           10.307208
## 5 304         3.4034            8.980705
## 6 305         3.7878            8.278200

Finally, lr_get_spec() can also recursively search in your folder tree with the subdir argument:

res <- lr_get_spec(where = "data", ext = "procspec", lim = c(300, 700), subdir = TRUE)
## 6 files found; importing spectra:
head(res)
##    wl OceanOptics_Linux OceanOptics_Windows OceanOptics_badencode  whiteref
## 1 300          126.5502            3.199635             -6.905214  98.30193
## 2 301          125.3005            3.420500             -7.034905  98.67972
## 3 302          127.0825            3.224495             -7.656868  98.10391
## 4 303          128.0483            3.320803             -8.577880 101.34410
## 5 304          128.9909            3.407551             -9.182934  99.86908
## 6 305          127.4218            3.492118             -9.367868 101.32638
##   BR_PF26_1 BR_PF27_3
## 1  19.35800  16.81450
## 2  19.40243  16.69849
## 3  19.42713  16.54195
## 4  19.47335  16.47455
## 5  19.48856  16.38495
## 6  19.41967  16.31837

As you may have noticed, lr_get_spec() does not care about the file extension case by default. This can be changed by using the ignore.case switch:

res <- lr_get_spec(where = "data", ext = "procspec", subdir = TRUE, ignore.case = FALSE)
## Warning: No files found. Try a different value for argument "ext".

If all your input files sample the wavelengths (this would be the case if you use the same spectrometer model and same recording software), you can also get uninterpolated data, by changing the value of the interpolate boolean argument:

res <- lr_get_spec(where = "data/heliomaster", ext = "jdx", interpolate = FALSE)
## Warning: No files found. Try a different value for argument "ext".
head(res)
## NULL

Import spectral metadata: lr_get_metadata()

lr_get_metadata() extracts metadata captured by the spectrophotometer during the recording. This metadata should be reported in your scientific articles to ensure reproducibility of your measurements and ultimately of your findings. The amount of information strongly depends on the brand and model of the spectrometer.

Similarly to lr_get_spec(), it can handle multiple extensions at once and perform recursive searches:

res <- lr_get_metadata(where = "data", ext = c("trt", "procspec"), subdir = TRUE)
## 7 files found; importing metadata:
head(res)
##                    name       user       date  spec_model     spec_ID
## 1       avantes_export2       <NA>       <NA>        <NA>   1305084U1
## 2     OceanOptics_Linux       hugo 2016-03-16     USB4000  USB4C00008
## 3   OceanOptics_Windows doutrelant 2015-12-04      JazUSB    JAZA2982
## 4 OceanOptics_badencode       user 2016-12-02 USB2000Plus USB2+H06330
## 5              whiteref      gomez 2018-08-02     USB4000  USB4C00008
## 6             BR_PF26_1 Adminlocal 2006-06-24     USB4000  USB4C00008
##   white_inttime dark_inttime sample_inttime white_avgs dark_avgs sample_avgs
## 1            95           95             95         20        20          20
## 2           200          200            200          5         5           5
## 3            60           60             60         15        15          15
## 4            20           20             20        100       100         100
## 5           500          500            500          5         5           5
## 6            10           10             10         40        40          40
##   white_boxcar dark_boxcar sample_boxcar
## 1            1           1             1
## 2            0           0             0
## 3            0           0             0
## 4            5           5             5
## 5            0           0             0
## 6           10          10            10

Convert spectral data to csv: lr_convert_tocsv()

lr_convert_tocsv() is designed for people that want to a machine readable version for each individual input file, possibly allowing them to carry on with their analysis using another programming language or software.

It works in a very similar way to lr_get_spec() and will create csv files with the same file names as the input files (but a different extension).

lr_convert_tocsv(where = "data", ext = "procspec", subdir = TRUE)