XpenseTracker -> hledger makefile

Makefile

# XpenseTracker iOS app -> hledger import & reports
#
# Fetch csv files from XpenseTracker iphone app (requires curl, funzip).
# Procedure:
#  connect phone and computer to same network
#  make csv
#  start XpenseTracker app
#  enter each XpenseTracker log, as prompted on computer
#   (first time: options, set appropriate CSV fields)
#   options, export, export
#   (if phone shows IP different from above, ctrl-c, adjust IP below, redo)
#   computer prints ok and prompts for next log
#

PHONEIP=192.168.0.3

# one per XT log
CSVFILES=\
	2017-business.csv \
	2017-personal.csv \
#	2018-business.csv \
#	2018-personal.csv \

JOURNALS=\
	2017-business.journal \
	2017-personal.journal \
#	2018-business.journal \
#	2018-personal.journal \

REPORTS= \
	out/2017-business-is.html \
	out/2017-business-is-summary.html \
	out/2017-personal-is.html \
	out/2017-personal-is-summary.html \
	out/2017-business-is.csv \
	out/2017-business-is-summary.csv \
	out/2017-personal-is.csv \
	out/2017-personal-is-summary.csv \

REPORTOPTS=-MAT --flat --drop 1
REPORTOPTSSUMMARY=$(REPORTOPTS) depth:2

.PHONY: $(CSVFILES) $(JOURNALS) $(REPORTS)

default: journals reports open


# FETCH CSV

csv: $(CSVFILES) 

CURLOPTS=-s #--connect-timeout 1

# Fetch, unzip, clean space-terminated account names, report success/failure
$(CSVFILES):
	@echo "waiting for $(@:.csv= log) from $(PHONEIP) ..."
	@(curl $(CURLOPTS) 'http://$(PHONEIP):8080/XT Export.zip' | funzip >$@.tmp 2>/dev/null && echo ok || (echo failed; false)) \
		&& (mv $@.tmp $@ && perl -pi -e 's/ :/:/' $@) \
		|| rm -f $@.tmp


# CONVERT ALL TXNS TO JOURNAL

journals: $(JOURNALS)

# recreate journal from csv, and include corresponding accounts file
$(JOURNALS):
	[[ -x $@ ]] || printf "# converted from $(@:.journal=.csv)\n\ninclude $(@:.journal=.accounts)\n\n" > $@
	hledger -f $(@:.journal=.csv) print >> $@


# IMPORT NEW TXNS TO JOURNAL

# # make 2018-business-import: import new csv txns to this journal
# %-import:
# 	hledger import -f $*.journal $*.csv

# # YEAR-import-dry: convert YEAR csv and show new transactions
# %-import-dry:
# 	hledger import -f $*.journal $**.csv --dry-run

# # YEAR-import-dry-watch: convert YEAR csv and show first N new transactions, repeatedly
# %-import-dry-watch:
# 	ls $**.csv* | entr bash -c "echo; date; echo; hledger import -f $*.journal $**.csv --dry-run | head -20"

# # YEAR-import-unknowns-watch: convert YEAR csv and show uncategorised new transactions, repeatedly
# %-import-unknowns-watch:
# 	ls $**.csv* | entr bash -c "echo; date; echo; hledger import -f $*.journal $**.csv --dry-run | hledger -I -f- print unknown"


# GENERATE CSV REPORTS FROM CSV

# # Generate per-log csv income statement reports from csv and open in Numbers.
# reports-from-csv: \
# 	2017-business-monthly-is.csv \
# 	2017-personal-monthly-is.csv \

# 2017-%-monthly-is.csv: #2017-$*.csv
# 	hledger -f 2017-$*.csv bal -M ^income ^expenses depth:2 -o $@ && open -a Numbers $@


# GENERATE CSV REPORTS FROM JOURNAL

# # Generate csv income statement reports from journals and open in Numbers.
# %-is.csv: #$*.journal
# 	hledger -f $*.journal is -M depth:2 -o $@ && open -a Numbers $@
# #	hledger -f $*.journal bal ^income ^expenses -M depth:2 -o $@ && open -a Numbers $@


# GENERATE REPORTS FROM JOURNAL

reports: $(REPORTS)

# Generate per-year html income statement reports from journals
#$(REPORTS):
out/2017-business-is.html \
out/2017-personal-is.html:
	hledger -f $(subst out/,,$(@:-is.html=.journal)) is -o $@ $(REPORTOPTS)

#%-summary.html:
out/2017-business-is-summary.html \
out/2017-personal-is-summary.html:
	hledger -f $(subst out/,,$(subst -is-summary.html,.journal,$@)) is -o $@ $(REPORTOPTSSUMMARY)

out/2017-business-is.csv \
out/2017-personal-is.csv:
	hledger -f $(subst out/,,$(@:-is.csv=.journal)) is -o $@ $(REPORTOPTS)

out/2017-business-is-summary.csv \
out/2017-personal-is-summary.csv:
	hledger -f $(subst out/,,$(subst -is-summary.csv,.journal,$@)) is -o $@ $(REPORTOPTSSUMMARY)


# BROWSE HTML REPORTS

open:
	open $(REPORTS)


# MISC

clean:
	rm -f $(JOURNALS)
	rm -f $(REPORTS)

My workflow is to shoot a photo on the spot, 'forward' it to a CASH YYYY-MM-DD Note in iOS Notes (you can select an existing note when forwarding, making it possible to have one note with multiple photos), and later for each day/note, create a PDF which I can put in my cash statements directory. Since I use hledger-flow, there are some extension points for additional steps (OCR, extracting PDF images into individual images again, etc..). I still have to look at each image and type the entries manually in a CSV file, but I'm hoping to use Google Cloud Vision OCR to help with that later as well, just haven't found an easy command-line way to interface with it.

Nice process! Consider adding it to the Cookbook.