Compare:

library(plater)
one <- read_plate("~/../Desktop/one.csv")
two <- read_plate("~/../Desktop/two.csv")

files.zip

Seems to be due to whether columns are all numeric or contain a string.

Hello @seaaan !

I've been working on some other plate-adjacent packages. {plater} is fantastic for the reading of plate-based metadata and the like. I have having the need to have more general purpose plate-related functions for dealing with well IDs, column / row data and plate layouts and the like. I started compiling them into a more general-purpose utility package (currently {wellr}.

I thought before I go any further with it, I should ask if maybe incorporating into plater, or exposing some of the internal functions inside of plater might be the better way to go. My overarching idea would be to have a utility package that multiple different packages, such as plater and tidyqpcr could then utilise, to try and start ensuring consistent API etc across packages that are working on the same data types (plate-based data), and to have it come under the ropensci banner.

I'm open to any of your thoughts on the matter!

I like the clear intention and focus of the package but is there any reason to not generalize it to other platedimensions than 12, 24, 48, 96 or 384 wells? For example I frequently work with 1536 well format and I could imagine that sometimes the 6 well format is required.

Detect programmatically

E.g.
Error in validatePlate(plate, plateSize) : Correct row labels not found. Found 'A B C D E G H I' but expected 'a b c d e f g h' or 'A B C D E F G H'. or Error: Invalid plate dimensions. Found 7 rows and 12 columns. Must be (8, 12) for a 96-well plate. should both specify the name of the file causing the error.

http://www.r-bloggers.com/plotting-microtiter-plate-maps/

Currently there is a very stringent check for the plate format, and some existing plates are not supported.
For example a plate with 15 wells - 5 columns and 3 rows. Could this be relaxed?
From what I can see, the number of rows and columns is inferred from the plate_size argument of view_plate()

Right now, passing in a file with a semi-colon as delimiter gives an unhelpful error message:

Error: Could not guess plate size from number of columns. Invalid number of columns: -Inf In addition: 

Warning message:
In max(as.numeric(first_line_vector)) :
  no non-missing arguments to max; returning -Inf

Warning messages:
1: In readLines(file) :
  incomplete final line found on '../../DATA/16-10-12 sort C4B 3 donors/gels/161012_978_2_G1.csv'
2: In readLines(file) :
  incomplete final line found on '../../DATA/16-10-12 sort C4B 3 donors/gels/161012_978_2_G1.csv'

example files.zip

e.g. cell containing: 5' LTR

COPASutils
platetools

Develop a simple interface where users can submit plateR format files and receive the data back in the shape of their choice.

tibble(well = c("A01", "F12", "G09"), Droplets = 1:3) %>% group_by(Droplets) %>% view_plate("well", "Droplets", 96)
Error: Error in cbind_all(x) : cannot convert object to a data frame

tibble(well = c("A01", "F12", "G09"), Droplets = 1:3) %>% group_by(well) %>% view_plate("well", "Droplets", 96)
Error: Error: Well IDs are invalid.

Running ungroup() or as.data.frame() in before view_plate resolves the issue.

Wrapper around lapply(..., FUN = read_plate) that also combines all of the plates into a single data frame.

Things to think about:

• What if plates have different columns (how to combine that into a df)

A function or a guide to how to do it

Sometimes I have meta data in plater format with no corresponding experimental data because of a technical problem. It would be nice if I did not have to remove this well from my plater meta data in order to get the rest of the merge to work,. Maybe there could just be NA's if plater can't find a match for that well? all.x = TRUE or all.y = TRUE kind of thing... :)