To have in input a vector of string of NMRExperiments

[philibe]

Hello.

Thank you very much for your powerful Nmr package, well documented with examples, integrated, and not too complex to use.

I saw that you have implemented Ensure unique exp names pull #48, it's useful, but when we have >100 experiments for a PCA, for example, it's not easy to analyze "10...54" experiment name and not the real name of experiment.

Here is my workaround :) (written before the pull #48 fix, but still useful for renaming duplicate names)

# my analyses (my_dir_of_NMR_analysis)
#   analysis L1 bla bla
#     10 : my_experiments[[1]]
#   analysis L2 bla bla
#     10 : my_experiments[[2]]
#     20 : my_experiments[[3]]

my_experiments<- fs::dir_ls( 
  my_dir_of_NMR_analysis   , 
  type="directory", 
  recurse=TRUE, 
  glob="*0"
)

dataset<- AlpsNMR::nmr_read_samples(my_experiments) 

dataset$NMRExperiment<-paste0(
  basename(dirname(dataset$metadata$info$info_sample_path)),
  "_exp_", 
  basename(dataset$metadata$info$info_sample_path) 
)

for (metadata_name_niv1 in names(dataset$metadata)) {
  for (metadata_name_niv2 in names(dataset$metadata[[metadata_name_niv1]])) {
    if (DescTools::Coalesce(dplyr::contains ("NMRExperiment", ignore.case = TRUE,metadata_name_niv2),0)==1) { 
      dataset$metadata[[metadata_name_niv1]][[metadata_name_niv2]]<- dataset$NMRExperiment
    }
  } 
}    

[zeehio]

Thanks for your feedback! If you have your samples from different groups in different folders then that's a way to name them, but not everyone uses that schema and some experiment designs may be more complex and not take care of naming so easily I guess...

I am currently heavily working on improving some parts of the package. Naming samples is important and we need to provide a convenient solution for everyone.

I could easily provide a function to set the sample names where you just give the sample names as a character vector and then AlpsNMR replaces the names where needed.

I could provide an additional helper to allow the renaming of the sample based on the sample path. I will explore some options, provide you some solutions and if they work for you then I will merge them for everyone. I hope to find the time to implement this soon.

[zeehio]

I've implemented two functions that will be helpful:

• names(<nmr_dataset>) allows you to get/set sample names
• nmr_read_samples(....) has a smarter approach to set the default sample names.

For instance, if the sample names are repeated, in different subfolders, such as:

- your_dataset/
  + control/
     * 10/
     * 20/
     * 30/
  + mutated/
     * 10/
     * 20/
     * 30/
``

Using:

```r
dataset <- nmr_read_samples_dir(c("your_dataset/control", "your_dataset/mutated"))
dataset

Will give you as sample names control/10, control/20, control/30, mutated/10, mutated/20, mutated/30.

I hope this behaves as you expect, you can always choose your own names, with names()<-

I will close this issue in a few days. Feel free to comment or close it yourself if you like.

Have a nice day!

[philibe]

Thanks. I will look your new feature next week. :)

[philibe]

I cannot upgrade AlpsNMR now because I have not yet R >=4.2 (R 4.1.2).

I will look in the future.

• names(<nmr_dataset>) could be useful, if I can use use with the same feature as my for example.
• nmr_read_samples_dir(c("your_dataset/control", "your_dataset/mutated")) could be useful but not for me now :) See below.

In fact I have many folders of Bruker datasets, each folder is different. I have never the same folder like control or dataset. My wish would have been like my for example below:

my analyses (my_dir_of_NMR_analysis)
analysis L1 bla bla
10 : my_experiments[[1]] , new NMRExperiment name in AlpsNMR : analysis L1 bla bla_10
analysis L2 bla bla
10 : my_experiments[[2]], new NMRExperiment name in AlpsNMR : analysis L2 bla bla_10
20 : my_experiments[[3]], new NMRExperiment name in AlpsNMR : analysis L2 bla bla_20

I see that you prefer to close my issue, even if I cannot test, and you prefer to reopen or create a new one to not have too many opened issues during months and months :)

Thank you anyway for your work.

[zeehio]

I'll try to relax the 4.2 dependency to 4.1.

Besides, would something like:

dataset <- nmr_read_samples_dir(list.dirs("your_dataset_dir/", recursive=FALSE))

list.dirs() would list all your analysis bla folders.

Work that work for you?

[philibe]

nmr_read_samples_dir(list.dirs("your_dataset_dir/", recursive=FALSE)) : yes : It's my use case :)

[zeehio]

The package seems to work on R 4.1 as well. Feel free to try it

[philibe]

Ok thanks. I installed it.

But I think we don't understand each other :D

See below for a new explanation of wish :)

Issue of names(<nmr_dataset>)

my_list<- fs::dir_ls( ROOT_DIRECTORY_ANALYSIS, type="directory", recurse=0)
dataset<-NULL
dataset<-AlpsNMR::nmr_read_samples_dir(
    my_list
)
names(dataset[1])<- my_list[1]
Warning message in names(dataset[1]) <- my_list[1]:
“number of items to replace is not a multiple of replacement length”

It doesn't recurse into each NmrExperiment or infoNmrExp to rename it. And it's normal, it's a vector.

Issue of nmr_read_samples_dir(list.dirs("your_dataset_dir/", recursive=FALSE))

my_list<- fs::dir_ls( ROOT_DIRECTORY_ANALYSIS, type="directory", recurse=0)
nmr_read_samples_dir(my_list)
New names:
* `1` -> `1...1`
* `10` -> `10...2`
* `1` -> `1...3`
* `10` -> `10...4`
* `1` -> `1...5`

New explanation :)

Here is my actual tree of directories. The RMN scientific where I work says it's not usual to work like we work.
We work to have one experiment by analysis to have only one spectrum by analysis.

directories

/mnt/rmn  <- ROOT_DIRECTORY_ANALYSIS
- /mnt/rmn/Flower1 <- bruker directory
   +  10 : Automatic name of NmrExperiment
       * pdata
       * acqus, fid, etc..

- /mnt/rmn/Flower2 <- bruker directory
   +  10 : Automatic name of NmrExperiment
       * pdata
       * acqus, fid, etc..       
- /mnt/rmn/Flower3 <- bruker directory
   +  10 : Automatic name of NmrExperiment
       * pdata
       * acqus, fid, etc..       

AlpsNMR Structure

to have in your structure, for example in the first and second one

dataset[1]

str(dataset[1])
List of 4
 $ metadata   :List of 7
  ..$ info    : tibble [1 × 8] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment      : chr "10...1"  => to be renamed "Flower1_10"
  .. ..$ info_NMRExperiment : chr "10" => to be renamed "Flower1_10"
  .. ..$ info_file_format   : chr "Bruker NMR directory"
  .. ..$ info_sample_path   : chr "/mnt/rmn/Flower1/10"
  .. ..$ info_dimension     : int 1

  ..$ orig    : tibble [1 × 1] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...1" => to be renamed "Flower1_10"
  ..$ title   : tibble [1 × 28] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment                 : chr "10...1"
  .. ..$ title_V1                      : chr "m=758.9mg eq 750ul"

  ..$ acqus   : tibble [1 × 239] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment     : chr "10...1" => to be renamed "Flower1_10"
  .. ..$ acqus_TITLE       : chr "Parameter file, TopSpin 3.5 pl 6"
  ..$ procs   : tibble [1 × 137] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment     : chr "10...1" => to be renamed "Flower1_10"

  ..$ levels  : tibble [1 × 2] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...1" => to be renamed "Flower1_10"
  .. ..$ levels_levels: logi NA
  ..$ external: tibble [1 × 1] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...1" => to be renamed "Flower1_10"
 $ data_1r    :List of 1
  ..$ : num [1:131072] -175 257 686 875 947 ...
 $ axis       :List of 1
  ..$ :List of 1
  .. ..$ : num [1:131072] 15.1 15.1 15.1 15.1 15.1 ...
 $ num_samples: int 1
 - attr(*, "class")= chr [1:2] "nmr_dataset" "nmr_dataset_family"

dataset[2]

str(dataset[2])
List of 4
 $ metadata   :List of 7
  ..$ info    : tibble [1 × 8] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment      : chr "10...2"  => to be renamed "Flower2_10"
  .. ..$ info_NMRExperiment : chr "10" => to be renamed "Flower2_10"
  .. ..$ info_file_format   : chr "Bruker NMR directory"
  .. ..$ info_sample_path   : chr "/mnt/rmn/Flower2/10"
  .. ..$ info_dimension     : int 1

  ..$ orig    : tibble [1 × 1] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...2" => to be renamed "Flower2_10"
  ..$ title   : tibble [1 × 28] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment                 : chr "10...2"
  .. ..$ title_V1                      : chr "m=758.9mg eq 750ul"

  ..$ acqus   : tibble [1 × 239] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment     : chr "10...2" => to be renamed "Flower2_10"
  .. ..$ acqus_TITLE       : chr "Parameter file, TopSpin 3.5 pl 6"
  ..$ procs   : tibble [1 × 137] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment     : chr "10...2" => to be renamed "Flower2_10"

  ..$ levels  : tibble [1 × 2] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...2" => to be renamed "Flower2_10"
  .. ..$ levels_levels: logi NA
  ..$ external: tibble [1 × 1] (S3: tbl_df/tbl/data.frame)
  .. ..$ NMRExperiment: chr "10...2" => to be renamed "Flower2_10"
 $ data_1r    :List of 1
  ..$ : num [1:131072] -175 257 686 875 947 ...
 $ axis       :List of 1
  ..$ :List of 1
  .. ..$ : num [1:131072] 15.1 15.1 15.1 15.1 15.1 ...
 $ num_samples: int 1
 - attr(*, "class")= chr [1:2] "nmr_dataset" "nmr_dataset_family"

Except that, AlpsNMR works

But except the issue (only for us) about the names of NmrExperiments, we are used to using AlpsNMR in particular nmr_read_samples_dir(), nmr_read_samples(), nmr_meta_add(), nmr_interpolate_1D(),nmr_pca_outliers_robust(), nmr_pca_outliers_plot(), plot(dataset) etc etc :) But with curves of spectrums or points with many many of their names like "10..1" "10..2" "10..50". So my loop of renaming at the beginning of this issue.

[zeehio]

I will prepare some unit tests to ensure your directory structure gives sensible names.

Meanwhile, I would expect something like this to work:

my_list<- fs::dir_ls( ROOT_DIRECTORY_ANALYSIS, type="directory", recurse=0)
ds <- nmr_read_samples_dir(my_list)
names(ds) <- c("Flower1", "Flower2", "Flower3")
# Or:
# names(ds) <- basename(my_list)

But I will do further testing. I'm writing this from my phone right now :-)