bquistorff

#!/bin/bash
#
# File to normalize Stata log files. Remove computer-specific,
# time-specific, and random components from log files.
# This enables identification of differences in Stata runs by
# using binary diffs (easy with a versioned project like one on SVN).

#To do: 
# deal with the unversioned env variable (quieted for now in ipums_migr...)

bquistorff

#!/usr/bin/sed -f

#### Standard Stata output ####

### Time/speed-specific specific
s/[ 0-9][0-9] [A-Z][a-z][a-z] [0-9]\{4\},\? ..:..\(:..\)\?/-normalized-/g

### Machine-specific
##OS-dependent: file path

bquistorff

#!/usr/bin/env python
#
# Rewrites a dataset without the timestamp nor junk in padding
# Helps making things bit-reproducible.
# Doesn't convert byte-order for numerical types.

import struct, sys, datetime, glob, tempfile, shutil, os, getopt

def error(message):
    sys.stderr.write(message + '\n')

bquistorff

#! /usr/bin/env python

'''
Normalizes a Stata .gph graph file to remove build-specific information.
Filepaths, dates, times, and random IDs are converted so that the file
will be the same no matter where it is built.

Pass in a file pattern (that is quoted so the shell doesn't expand it).

To do:

bquistorff

#! /bin/bash
# Removed non-versioned files recursively (otherwise pipe through grep -v '\\')
svn status | grep ^\? | cut -c 9- | sed -e 's/\\/\//g' | xargs -Ifile rm -rf file

bquistorff

#!/bin/bash
# renames files (and svn mv's them if in svn)
# not sure if this works if there is whitespace in these vars
# use from base of project
findstr=$1
replacestr=$2

# It currently doesn't search in hidden folders
# for no recursion, use:
# for i in *

bquistorff

# If you don't list all dependencies for code (because maybe it varies programmatically) then you can force a remake by doing 'make -B <target>. Note that it will remake all dependencies!

#nice-ness. Uncomment a line to enable. Default is +10
#nice_prefix := nice
#nice_prefix := nice -n10

############################# Project-specific Entries ####################

# which entries use parallel and need a windows gateway	
# from $grep -r eval_synth . --include="*.do"

bquistorff

* all mata objects used across files (or even just all)
* Should be made into an mlib so you don't have to worry
* about including multiple times.
* This should also help a bit with speed.

*Possibly get some project-specific settings
*qui include ....do

* Load all the mata files in ado/
local mata_files : dir "ado" files "*.mata"

bquistorff

*! converts a *.gph file to another format
* Required globals: dir_base
program gph2fmt
	args file fmt
	
	graph use "${dir_base}/fig/gph/`file'", name(toexport)
	
	local temp=regexm("`file'","(.+)\.gph")
	local file_name_base=regexs(1) 
	

bquistorff

*Callable from the command line for converting a gph to eps's in the right places
global dir_base ".."
adopath ++ "${dir_base}/code/ado/"
gph2fmt `1'.gph eps
if "`2'"!=""{
    gph2fmt `1'.gph `2'
}