[ dbx ][ mobile ][ todo ][ common ][ Info ]

A place to start software projects. Mostly shell based, but heck, some python likely.

This idea originated with the manufacture table I started yesterday (6/9/15).

dataFlow

From my moleskin notebook today

I'm giving myself the hour to formulate the data structure for a shell function to support the data flow diagram:

name () { command args files
         [ | command ... [  ... ] 
         > output ] ; }

 input file ... | -->  ( process ) --> | output file ...
                           ^
                           |
            ( args ) . . . +


 data structure := { table : name + { field } 1,m } 1,N

Since a function may invoke a list of piped commands, each with it's own arguments, some (many) of which are files.

A function is a list of commands

tables

command

name files arguments

flow

This is the principaal data-flow table:

• inputs – an RE of then input file name(s), which may not match any
• nargs – the minimum # of arguments, failing produces a report
• mode – always, append or newest, for the output dependency on the inputs
• output – a file name, which needn't exist
• command – to run on the inputs to produce the output, it may have local args
• dispose – if not empty, a command to cleanup the output, followed by the inputs

function

name commands

process

name function arguments inputs outputs

thinking, design

For starters, the /rdb list format looks like a good start

file    function

field name
field commands

file  command

field name
field files
field arguments

file fragment
field name
field function
field arguments
field inputs      # defaults to stdin
field outputs     # defaults to stdout

Let's start with a simple function

function invest_rdb
{
    set -- $(y_m_d $1)
    trace_call $*
    rdb_hdr $(invest_csv);
    lastFid tocancsv $1 | grep -iv 'name/number' | cawk 'NF > 12' | tr , '\t'
}

#+BEGIN_SRC

type fragment
name investhdr
inpt NULL
func rdb_hdr
args $(invest_csv)

type fragment
name lastFidA
inpt $(lastFid)
func tocansv
args $(y_m_d $1)

type fragment
name grepname
args -iv 'name/number'
func grep 

type fragment 
name fGT12
func cawk
args 'NF > 12'

type fragment
name csv2tab
func tr
args , '\t'

type function
name position
func column
args account symbol quantity
inpt invest_collect.rdb
outp position.rdb

type function
name tocancsv
func sed
args -f tocancsv.sed
inpt EITHER

type fragment
name 
func 
args 

type fragment
name 
func 
args 

type fragment
name 
func 
args