Makefile

#
# CSrankings
# http://csrankings.org
# Copyright (C) 2017-2020 by Emery Berger <http://emeryberger.org>
# See COPYING for license information.
#

TARGETS = csrankings.js csrankings.min.js submit/submit.js generated-author-info.csv

.PHONY: home-pages scholar-links fix-affiliations update-dblp clean-dblp download-dblp shrink-dblp clean-csrankings update-author-names update-dblp-full apply-author-names backup-dblp update-dblp-date download-prev-dblp

PYTHON = python3.12 # 3.7
PYPY   = python3.12 # pypy

# DBLP   = dblp.org
DBLP   = dblp.uni-trier.de

# Note: BaseX is no longer used for DBLP filtering. We now use a streaming Python SAX parser
# which uses constant memory (~50MB) instead of loading the entire document (~11GB).

all: generated-author-info.csv csrankings.js csrankings.min.js submit/submit.js csrankings.csv  # fix-affiliations home-pages scholar-links
	$(MAKE) clean-csrankings

clean:
	rm -f $(TARGETS)

csrankings.js: src/types.ts src/config.ts src/utils.ts src/data-loader.ts src/region.ts src/computation.ts src/verification.ts src/rendering.ts src/continents.ts src/app.ts
	@echo "Rebuilding JavaScript code."
	tsc --project tsconfig.json
	$(PYTHON) util/concat-js.py

submit/submit.js: src/submit.ts
	@echo "Rebuilding submit form JavaScript."
	tsc --project tsconfig.submit.json

csrankings.min.js: csrankings.js
	google-closure-compiler --js csrankings.js > csrankings.min.js

update-dblp:
	$(MAKE) download-dblp
	$(MAKE) shrink-dblp
	$(PYTHON) util/generate-aliases.py > dblp-aliases.csv
	@echo "Done."

clean-dblp:
	@echo "Fixing character encodings."
	sh ./util/fix-dblp.sh
	gzip dblp-fixed.xml
	$(PYTHON) util/find-missing-names-dblp.py

download-dblp:
	@echo "Downloading from DBLP."
	rm -f dblp.xml.gz
	curl -o dblp-original.xml.gz https://$(DBLP)/xml/dblp.xml.gz

shrink-dblp:
	@echo "Shrinking the DBLP file (streaming, low memory)."
	gunzip -dc dblp-original.xml.gz | $(PYTHON) util/filter-dblp.py | xz -9 > dblp.xml.xz
	@echo "Filtered DBLP saved to dblp.xml.xz"

faculty-affiliations.csv homepages.csv scholar.csv csrankings.csv: csrankings-*.csv
	@echo "Splitting main datafile."
	@$(PYTHON) util/split-csv.py
	@echo "Done."

clean-csrankings:
	@echo "Cleaning."
	@$(PYTHON) util/clean-csrankings.py
	@$(PYTHON) util/sort-csv-files.py
	@echo "Done."

home-pages: faculty-affiliations.csv homepages.csv
	@echo "Rebuilding home pages (homepages.csv)."
	@$(PYTHON) util/make-web-pages.py
	@echo "Cleaning home pages."
	@$(PYTHON) util/clean-web-pages.py
	@mv homepages-sorted.csv homepages.csv
	@echo "Done."

scholar-links: faculty-affiliations.csv homepages.csv
	@echo "Rebuilding Google Scholar links (scholar.csv)."
	@$(PYTHON) util/make-scholar-links.py
	@echo "Cleaning Scholar links."
	@$(PYTHON) util/clean-scholar-links.py
	@echo "Done."

fix-affiliations: faculty-affiliations.csv
	@echo "Updating affiliations."
	@$(PYTHON) util/fix-affiliations.py | sort -k2 -t"," | uniq > /tmp/f1.csv
	@echo "name,affiliation" | cat - /tmp/f1.csv >  /tmp/f2.csv
	@rm /tmp/f1.csv
	@mv /tmp/f2.csv faculty-affiliations.csv

faculty-coauthors.csv: dblp.xml.xz util/generate-faculty-coauthors.py util/csrankings.py
	@echo "Rebuilding the co-author database (faculty-coauthors.csv)."
	$(PYTHON) util/generate-faculty-coauthors.py
	@echo "Done."

generated-author-info.csv: faculty-affiliations.csv dblp.xml.xz util/regenerate_data.py util/csrankings.py dblp-aliases.csv
	@echo "Rebuilding the publication database (generated-author-info.csv)."
	@$(PYPY) util/split-csv.py
	@$(PYPY) util/regenerate_data.py
	@echo "Done."
#	@$(MAKE) clean-csrankings
#       @$(PYPY) util/split-csrankings.py
#	@$(MAKE) clean-csrankings

collab-graph: generated-author-info.csv faculty-coauthors.csv
	@echo "Generating the list of all publications (all-author-info.csv)."
	$(PYTHON) util/generate-all-pubs.py
	@echo "Building collaboration graph data."
	$(PYTHON) util/make-collaboration-graph.py

# Update the DBLP date in index.html to current month/year
update-dblp-date:
	@echo "Updating DBLP date in index.html..."
	@MONTH_YEAR=$$(date "+%B %Y"); \
	sed -i.bak -E "s/(DBLP<\/a> \()[A-Za-z]+ [0-9]{4}(\))/\1$$MONTH_YEAR\2/" index.html && \
	rm -f index.html.bak && \
	echo "Updated to: $$MONTH_YEAR"

# Backup current DBLP before downloading new one
backup-dblp:
	@if [ -f dblp-original.xml.gz ]; then \
		echo "Backing up dblp-original.xml.gz to prev-dblp.xml.gz"; \
		cp dblp-original.xml.gz prev-dblp.xml.gz; \
	else \
		echo "No dblp-original.xml.gz to backup (first run?)"; \
	fi

# Download the DBLP snapshot corresponding to the last update date (from index.html)
# This is used in CI where dblp-original.xml.gz is not checked in
# DBLP releases are at: https://drops.dagstuhl.de/storage/artifacts/dblp/xml/YYYY/dblp-YYYY-MM-01.xml.gz
download-prev-dblp:
	@echo "Downloading previous DBLP snapshot based on last update date..."
	@MONTH_YEAR=$$(grep -oE 'DBLP</a> \([A-Za-z]+ [0-9]{4}\)' index.html | grep -oE '[A-Za-z]+ [0-9]+'); \
	if [ -z "$$MONTH_YEAR" ]; then \
		echo "Error: Could not parse DBLP date from index.html"; \
		exit 1; \
	fi; \
	YEAR=$$(echo "$$MONTH_YEAR" | grep -oE '[0-9]+'); \
	MONTH_NAME=$$(echo "$$MONTH_YEAR" | grep -oE '[A-Za-z]+'); \
	MONTH_NUM=$$(echo "January February March April May June July August September October November December" | \
		tr ' ' '\n' | grep -n "^$$MONTH_NAME$$" | cut -d: -f1 | xargs printf "%02d"); \
	if [ -z "$$MONTH_NUM" ]; then \
		echo "Error: Could not parse month '$$MONTH_NAME'"; \
		exit 1; \
	fi; \
	RELEASE_DATE="$$YEAR-$$MONTH_NUM-01"; \
	echo "Last update was: $$MONTH_YEAR -> downloading dblp-$$RELEASE_DATE.xml.gz"; \
	curl -f -o dblp-original.xml.gz "https://drops.dagstuhl.de/storage/artifacts/dblp/xml/$$YEAR/dblp-$$RELEASE_DATE.xml.gz" || \
		(echo "Error: Failed to download DBLP release for $$RELEASE_DATE" && exit 1)

# Detect DBLP author name changes (dry-run preview)
# Requires: prev-dblp.xml.gz (previous DBLP dump) and dblp-original.xml.gz (current)
update-author-names:
	@echo "Detecting DBLP author name changes..."
	@if [ ! -f prev-dblp.xml.gz ]; then \
		echo "Error: prev-dblp.xml.gz not found. Save old DBLP before downloading new one:"; \
		echo "  cp dblp-original.xml.gz prev-dblp.xml.gz"; \
		exit 1; \
	fi
	$(PYTHON) util/new-name-detector.py --old prev-dblp.xml.gz --new dblp-original.xml.gz > name-changes.csv
	@echo "Name changes written to name-changes.csv"
	@echo "Previewing changes (dry-run):"
	$(PYTHON) util/update-new-names.py --diff name-changes.csv
	@echo ""
	@echo "To apply changes, run:"
	@echo "  make apply-author-names"

# Apply detected author name changes (modifies csrankings-*.csv files)
apply-author-names:
	@if [ ! -f name-changes.csv ]; then \
		echo "Error: name-changes.csv not found. Run 'make update-author-names' first."; \
		exit 1; \
	fi
	@CHANGES=$$(wc -l < name-changes.csv | tr -d ' '); \
	if [ "$$CHANGES" -le 1 ]; then \
		echo "No name changes to apply."; \
	else \
		echo "Applying $$(($$CHANGES - 1)) name change(s)..."; \
		$(PYTHON) util/update-new-names.py --diff name-changes.csv --in-place; \
	fi

# Fully automated DBLP update: backup, download, shrink, detect & apply name changes, regenerate data
# This is the one-command solution for updating DBLP
update-dblp-full:
	@echo "=== FULLY AUTOMATED DBLP UPDATE ==="
	@echo ""
	@echo "Step 1/9: Downloading previous DBLP snapshot (for name change detection)..."
	$(MAKE) download-prev-dblp
	@echo ""
	@echo "Step 2/9: Backing up to prev-dblp.xml.gz..."
	$(MAKE) backup-dblp
	@echo ""
	@echo "Step 3/9: Downloading new DBLP from $(DBLP)..."
	$(MAKE) download-dblp
	@echo ""
	@echo "Step 4/9: Filtering DBLP to CSRankings venues..."
	$(MAKE) shrink-dblp
	@echo ""
	@echo "Step 5/9: Generating DBLP aliases..."
	$(MAKE) faculty-affiliations.csv
	$(PYTHON) util/generate-aliases.py > dblp-aliases.csv
	@echo ""
	@echo "Step 6/9: Detecting and applying author name changes..."
	@if [ -f prev-dblp.xml.gz ]; then \
		$(PYTHON) util/new-name-detector.py --old prev-dblp.xml.gz --new dblp-original.xml.gz > name-changes.csv; \
		CHANGES=$$(wc -l < name-changes.csv | tr -d ' '); \
		if [ "$$CHANGES" -le 1 ]; then \
			echo "No name changes detected."; \
		else \
			echo "Detected $$(($$CHANGES - 1)) name change(s). Applying..."; \
			$(PYTHON) util/update-new-names.py --diff name-changes.csv --in-place; \
		fi; \
	else \
		echo "No previous DBLP backup found. Skipping name change detection."; \
	fi
	@echo ""
	@echo "Step 7/9: Regenerating publication data..."
	$(MAKE) generated-author-info.csv
	@echo ""
	@echo "Step 8/9: Rebuilding csrankings.csv with DBLP alias expansion..."
	@# clean-csrankings.py is the only step that expands dblp-aliases.csv into
	@# csrankings.csv. Without it this target writes an alias-stripped file (~2,100
	@# rows short), which the next post-merge 'make all' silently puts back.
	$(MAKE) clean-csrankings
	@echo ""
	@echo "Step 9/9: Updating DBLP date in index.html..."
	$(MAKE) update-dblp-date
	@echo ""
	@echo "=== DBLP UPDATE COMPLETE ==="
	@echo "You may want to run 'make all' to rebuild everything."