IO (Input Output)*

work-in-progress

Working with file paths

Reading a file

$ cat > test.txt
can we all be happy!
also the sad people! :P

$ rye

read %test.txt
; returns string with two lines

read\lines %test.txt
; returns block with two strings
; one for each line

Writing to a file

"Save me to a disc"
|write %save.txt

"Me too"
|write\append %save.txt   ; TODO

Copying a file

Rye can copy a file directly from reader to writer without loading it full into memory.

$ cat > source.txt
hello copy!

$ rye

open %source.txt |copy create %destination.txt

Appending to file

Append next number to a file each second, forever.

open\append %seconds.txt :wr

forever { .to-string + "\n" |write\string* wr , sleep 1000 }

[Ctrl-z]

$ tail -f seconds.txt

TODO: #bug Ctrl-z or Ctrl-c don’t work when in forever loop.

Tailing a file

The tail -f opens a file and waits for appends to a file and prints them. We can also to this in Rye, retrieving additions to a file as they come.

open %seconds.txt :f
.seek\end
.reader :r
forever { read\string r "\n" |print }

HTTP Download

Downloading a bigger file without loading it into memmory. Open with htts URI opens a reader, and copy copies (streams) it over to a writer, which we can create with create and a file path.

open https://fu.gov.si/fileadmin/prenosi/DURS_zavezanci_PO_csv.zip
|copy open %download.zip

Downloading from FTP

FTP is more complex (stateful?) protocol, but we can use it similarly as HTTP above to download big files with directly streaming them to disc for example.

; load the ftp username and password from a file
info: context { do load %.ftpinfo }

open ftp://prenos.example.com:21
|login info/user info/pwd
|retrieve "rtr/MES/BIG_FILE.zip"
|copy create %local3.zip