def main(anthology):
variants = defaultdict(list)
slugs = {}
for person in anthology.people.personids():
name = anthology.people.get_canonical_name(person)
name_slug = slugify(repr(name))
if name_slug in slugs:
variants[slugs[name_slug]].append(repr(name))
else:
slugs[name_slug] = repr(name)
canonical_variants = []
for var1, varlist in variants.items():
varlist.append(var1)
# Determine best canonical variant
scores = {var: score_variant(var) for var in varlist}
canonical = max(scores.items(), key=lambda k: k[1])[0]
canonical_variants.append({
"canonical":
to_dict(PersonName.from_repr(canonical)),
"variants": [
to_dict(PersonName.from_repr(var)) for var in varlist
if var != canonical
],
})
canonical_variants.sort(
key=lambda x: (x["canonical"]["last"], x["canonical"]["first"]))
# flow style to match format of file name_variants.yaml
print(
yaml.dump(canonical_variants,
allow_unicode=True,
default_flow_style=None))
def main(anthology):
variants = defaultdict(list)
slugs = {}
for name in anthology.people.names():
name_slug = slugify(repr(name))
if name_slug in slugs:
variants[slugs[name_slug]].append(repr(name))
else:
slugs[name_slug] = repr(name)
canonical_variants = []
for var1, varlist in variants.items():
varlist.append(var1)
# Determine best canonical variant
scores = {var: score_variant(var) for var in varlist}
canonical = max(scores.items(), key=lambda k: k[1])[0]
canonical_variants.append(
{
"canonical": to_dict(PersonName.from_repr(canonical)),
"variants": [
to_dict(PersonName.from_repr(var))
for var in varlist
if var != canonical
],
}
)
print(yaml.dump(canonical_variants, allow_unicode=True))
def main(anthology):
variants = defaultdict(list)
slugs = {}
for name in anthology.people.names():
name_slug = slugify(repr(name))
if name_slug in slugs:
variants[slugs[name_slug]].append(repr(name))
else:
slugs[name_slug] = repr(name)
canonical_variants = []
for var1, varlist in variants.items():
varlist.append(var1)
# Determine best canonical variant
scores = {var: score_variant(var) for var in varlist}
canonical = max(scores.items(), key=lambda k: k[1])[0]
canonical_variants.append({
"canonical":
to_dict(PersonName.from_repr(canonical)),
"variants": [
to_dict(PersonName.from_repr(var)) for var in varlist
if var != canonical
],
})
print(yaml.dump(canonical_variants, allow_unicode=True))
def disambiguate_name(node):
name = PersonName.from_element(node)
ids = people.get_ids(name)
if node.tag == "editor":
# meta block, get volume
anth_id = build_anthology_id(
collection_id,
node.getparent().getparent().attrib["id"])
elif node.tag == "author":
# paper, get full ID
anth_id = build_anthology_id(
collection_id,
node.getparent().getparent().attrib["id"],
node.getparent().attrib["id"],
)
if len(ids) > 1:
choice = -1
while choice < 0 or choice >= len(ids):
print(
f"({anth_id}): ambiguous author {name}; Please choose from the following:"
)
for i, id_ in enumerate(ids):
print(f"[{i}] {id_} ({people.get_comment(id_)})")
choice = int(input("--> "))
node.attrib["id"] = ids[choice]
def disambiguate_name(node, anth_id):
name = PersonName.from_element(node)
ids = people.get_ids(name)
choice = -1
if len(ids) > 1:
while choice < 0 or choice >= len(ids):
print(
f"({anth_id}): ambiguous author {name}; Please choose from the following:"
)
for i, id_ in enumerate(ids):
print(f"[{i}] {id_} ({people.get_comment(id_)})")
choice = int(input("--> "))
return ids[choice], choice
tree_being_added = etree.parse(args.infile)
# Ensure nested format
root_being_added = make_nested(tree_being_added.getroot(),
pdf_path=os.path.dirname(args.infile))
collection_id = root_being_added.attrib["id"]
# Ensure names are properly identified
ambiguous = {}
for paper in root_being_added.findall(".//paper"):
anth_id = build_anthology_id(collection_id,
paper.getparent().attrib["id"],
paper.attrib["id"])
for node in chain(paper.findall("author"), paper.findall("editor")):
name = PersonName.from_element(node)
ids = people.get_ids(name)
if len(ids) > 1:
print(
f"WARNING ({anth_id}): ambiguous author {name}, defaulting to first of {ids}"
)
ambiguous[anth_id] = (name, ids)
node.attrib["id"] = ids[0]
# Ensure PDF exists. PDF should be in the same directory as the XML file being ingested.
pdf_path = os.path.join(pdf_directory, f"{anth_id}.pdf")
if not os.path.exists(pdf_path):
print(f"FATAL: cannot file PDF {pdf_path}!")
sys.exit(1)
def main(args):
collections = defaultdict(OrderedDict)
volumes = {}
anthology_datadir = os.path.join(os.path.dirname(sys.argv[0]), "..", "data")
venue_index = VenueIndex(srcdir=anthology_datadir)
venue_keys = [venue["slug"].lower() for _, venue in venue_index.items()]
# Build list of volumes, confirm uniqueness
unseen_venues = []
for proceedings in args.proceedings:
meta = read_meta(os.path.join(proceedings, "meta"))
venue_abbrev = meta["abbrev"]
venue_slug = venue_index.get_slug(venue_abbrev)
if venue_slug not in venue_keys:
unseen_venues.append((venue_slug, venue_abbrev, meta["title"]))
meta["path"] = proceedings
meta["collection_id"] = collection_id = meta["year"] + "." + venue_slug
volume_name = meta["volume"].lower()
volume_full_id = f"{collection_id}-{volume_name}"
if volume_full_id in volumes:
print("Error: ")
collections[collection_id][volume_name] = {}
volumes[volume_full_id] = meta
# Make sure all venues exist
if len(unseen_venues) > 0:
for venue in unseen_venues:
slug, abbrev, title = venue
print(f"Creating venue '{abbrev}' ({title})")
venue_index.add_venue(abbrev, title)
venue_index.dump(directory=anthology_datadir)
# Copy over the PDFs and attachments
for volume, meta in volumes.items():
root_path = os.path.join(meta["path"], "cdrom")
collection_id = meta["collection_id"]
venue_name = meta["abbrev"].lower()
volume_name = meta["volume"].lower()
year = meta["year"]
pdfs_dest_dir = os.path.join(args.pdfs_dir, venue_name)
if not os.path.exists(pdfs_dest_dir):
os.makedirs(pdfs_dest_dir)
# copy the book
book_src_filename = meta["abbrev"] + "-" + year
book_src_path = os.path.join(root_path, book_src_filename) + ".pdf"
book_dest_path = None
if os.path.exists(book_src_path):
book_dest_path = (
os.path.join(pdfs_dest_dir, f"{collection_id}-{volume_name}") + ".pdf"
)
if not args.dry_run:
maybe_copy(book_src_path, book_dest_path)
# copy the paper PDFs
pdf_src_dir = os.path.join(root_path, "pdf")
for pdf_file in os.listdir(pdf_src_dir):
# Skip . files
if os.path.basename(pdf_file).startswith("."):
continue
# names are {abbrev}{number}.pdf
match = re.match(rf".*\.(\d+)\.pdf", pdf_file)
if match is not None:
paper_num = int(match[1])
paper_id_full = f"{collection_id}-{volume_name}.{paper_num}"
bib_path = os.path.join(
root_path,
"bib",
pdf_file.replace("/pdf", "/bib/").replace(".pdf", ".bib"),
)
pdf_src_path = os.path.join(pdf_src_dir, pdf_file)
pdf_dest_path = os.path.join(
pdfs_dest_dir, f"{collection_id}-{volume_name}.{paper_num}.pdf"
)
if not args.dry_run:
maybe_copy(pdf_src_path, pdf_dest_path)
collections[collection_id][volume_name][paper_num] = {
"anthology_id": paper_id_full,
"bib": bib_path,
"pdf": pdf_dest_path,
"attachments": [],
}
# copy the attachments
if os.path.exists(os.path.join(root_path, "additional")):
attachments_dest_dir = os.path.join(args.attachments_dir, venue_name)
if not os.path.exists(attachments_dest_dir):
os.makedirs(attachments_dest_dir)
for attachment_file in os.listdir(os.path.join(root_path, "additional")):
if os.path.basename(attachment_file).startswith("."):
continue
attachment_file_path = os.path.join(
root_path, "additional", attachment_file
)
match = re.match(
rf"{year}\.{venue_name}-\w+\.(\d+)_?(\w+)\.(\w+)$", attachment_file
)
if match is None:
print(
f"* Warning: no attachment match for {attachment_file}",
file=sys.stderr,
)
sys.exit(2)
paper_num, type_, ext = match.groups()
paper_num = int(paper_num)
file_name = f"{collection_id}-{volume_name}.{paper_num}.{type_}.{ext}"
dest_path = os.path.join(attachments_dest_dir, file_name)
if not args.dry_run and not os.path.exists(dest_path):
log(f"Copying {attachment_file} -> {dest_path}", args.dry_run)
shutil.copyfile(attachment_file_path, dest_path)
collections[collection_id][volume_name][paper_num]["attachments"].append(
(dest_path, type_)
)
people = AnthologyIndex(None, srcdir=anthology_datadir)
def correct_caps(person, name_node, anth_id):
"""
Many people submit their names in "ALL CAPS" or "all lowercase".
Correct this with heuristics.
"""
name = name_node.text
if name.islower() or name.isupper():
# capitalize all parts
corrected = " ".join(list(map(lambda x: x.capitalize(), name.split())))
print(
f"-> Correcting capitalization of '{name}' to '{corrected}'",
file=sys.stderr,
)
name_node.text = corrected
def disambiguate_name(node, anth_id):
name = PersonName.from_element(node)
ids = people.get_ids(name)
if len(ids) > 1:
choice = -1
while choice < 0 or choice >= len(ids):
print(
f"({anth_id}): ambiguous author {name}; Please choose from the following:"
)
for i, id_ in enumerate(ids):
print(f"[{i}] {id_} ({people.get_comment(id_)})")
choice = int(input("--> "))
node.attrib["id"] = ids[choice]
for collection_id, collection in collections.items():
# Newly added volumes, so we can normalize and name-disambig later
newly_added_volumes = []
collection_file = os.path.join(
args.anthology_dir, "data", "xml", f"{collection_id}.xml"
)
if os.path.exists(collection_file):
root_node = etree.parse(collection_file).getroot()
else:
root_node = make_simple_element("collection", attrib={"id": collection_id})
for volume_id, volume in collection.items():
volume_node = make_simple_element(
"volume",
attrib={"id": volume_id, "ingest-date": args.ingest_date},
)
# Replace the existing one if present
existing_volume_node = root_node.find(f"./volume[@id='{volume_id}']")
for i, child in enumerate(root_node):
if child.attrib["id"] == volume_id:
root_node[i] = volume_node
break
else:
root_node.append(volume_node)
meta_node = None
for paper_num, paper in sorted(volume.items()):
paper_id_full = paper["anthology_id"]
bibfile = paper["bib"]
paper_node = bib2xml(bibfile, paper_id_full)
if paper_node.attrib["id"] == "0":
# create metadata subtree
meta_node = make_simple_element("meta", parent=volume_node)
title_node = paper_node.find("title")
title_node.tag = "booktitle"
meta_node.append(title_node)
for author_or_editor in chain(
paper_node.findall("./author"), paper_node.findall("./editor")
):
meta_node.append(author_or_editor)
author_or_editor.tag = "editor"
meta_node.append(paper_node.find("publisher"))
meta_node.append(paper_node.find("address"))
meta_node.append(paper_node.find("month"))
meta_node.append(paper_node.find("year"))
if book_dest_path is not None:
make_simple_element(
"url",
text=f"{collection_id}-{volume_name}",
attrib={"hash": compute_hash_from_file(book_dest_path)},
parent=meta_node,
)
# modify frontmatter tag
paper_node.tag = "frontmatter"
del paper_node.attrib["id"]
else:
# remove unneeded fields
for child in paper_node:
if child.tag in [
"editor",
"address",
"booktitle",
"publisher",
"year",
"month",
]:
paper_node.remove(child)
url = paper_node.find("./url")
if url is not None:
url.attrib["hash"] = compute_hash_from_file(paper["pdf"])
for path, type_ in paper["attachments"]:
make_simple_element(
"attachment",
text=os.path.basename(path),
attrib={
"type": type_,
"hash": compute_hash_from_file(path),
},
parent=paper_node,
)
if len(paper_node) > 0:
volume_node.append(paper_node)
# Normalize
for oldnode in paper_node:
normalize(oldnode, informat="latex")
for name_node in chain(
paper_node.findall("./author"), paper_node.findall("./editor")
):
disambiguate_name(name_node, paper_id_full)
person = PersonName.from_element(name_node)
for name_part in name_node:
correct_caps(person, name_part, paper_id_full)
# Other data from the meta file
if "isbn" in meta:
make_simple_element("isbn", meta["isbn"], parent=meta_node)
indent(root_node)
tree = etree.ElementTree(root_node)
tree.write(
collection_file, encoding="UTF-8", xml_declaration=True, with_tail=True
)
def main(args):
volumes = {}
anthology_datadir = os.path.join(os.path.dirname(sys.argv[0]), "..", "data")
venue_index = VenueIndex(srcdir=anthology_datadir)
venue_keys = [venue["slug"].lower() for _, venue in venue_index.items()]
sig_index = SIGIndex(srcdir=anthology_datadir)
people = AnthologyIndex(srcdir=anthology_datadir)
people.bibkeys = load_bibkeys(anthology_datadir)
def correct_caps(name):
"""
Many people submit their names in "ALL CAPS" or "all lowercase".
Correct this with heuristics.
"""
if name.islower() or name.isupper():
# capitalize all parts
corrected = " ".join(list(map(lambda x: x.capitalize(), name.split())))
print(
f"-> Correcting capitalization of '{name}' to '{corrected}'",
file=sys.stderr,
)
name = corrected
return name
def disambiguate_name(node, anth_id):
name = PersonName.from_element(node)
ids = people.get_ids(name)
choice = -1
if len(ids) > 1:
while choice < 0 or choice >= len(ids):
print(
f"({anth_id}): ambiguous author {name}; Please choose from the following:"
)
for i, id_ in enumerate(ids):
print(f"[{i}] {id_} ({people.get_comment(id_)})")
choice = int(input("--> "))
return ids[choice], choice
# Build list of volumes, confirm uniqueness
unseen_venues = []
for proceedings in args.proceedings:
meta = read_meta(os.path.join(proceedings, "meta"))
venue_abbrev = meta["abbrev"]
venue_slug = venue_index.get_slug(venue_abbrev)
if str(datetime.now().year) in venue_abbrev:
print(f"Fatal: Venue assembler put year in acronym: '{venue_abbrev}'")
sys.exit(1)
if re.match(r".*\d$", venue_abbrev) is not None:
print(
f"WARNING: Venue {venue_abbrev} ends in a number, this is probably a mistake"
)
if venue_slug not in venue_keys:
unseen_venues.append((venue_slug, venue_abbrev, meta["title"]))
meta["path"] = proceedings
meta["collection_id"] = collection_id = meta["year"] + "." + venue_slug
volume_name = meta["volume"].lower()
volume_full_id = f"{collection_id}-{volume_name}"
if volume_full_id in volumes:
print("Error: ")
volumes[volume_full_id] = meta
if "sig" in meta:
print(
f"Add this line to {anthology_datadir}/sigs/{meta['sig'].lower()}.yaml:"
)
print(f" - {meta['year']}:")
print(f" - {volume_full_id} # {meta['booktitle']}")
# Make sure all venues exist
if len(unseen_venues) > 0:
for venue in unseen_venues:
slug, abbrev, title = venue
print(f"Creating venue '{abbrev}' ({title})")
venue_index.add_venue(abbrev, title)
venue_index.dump(directory=anthology_datadir)
# Copy over the PDFs and attachments
for volume_full_id, meta in volumes.items():
root_path = os.path.join(meta["path"], "cdrom")
collection_id = meta["collection_id"]
venue_name = meta["abbrev"].lower()
volume_name = meta["volume"].lower()
year = meta["year"]
pdfs_dest_dir = os.path.join(args.pdfs_dir, venue_name)
if not os.path.exists(pdfs_dest_dir):
os.makedirs(pdfs_dest_dir)
# copy the book from the top-level proceedings/ dir, named "VENUE-year.pdf"
book_path = f'cdrom/{venue_name.upper()}-{year}.pdf'
book_src_path = os.path.join(meta["path"], book_path)
book_dest_path = None
if os.path.exists(book_src_path) and not args.dry_run:
book_dest_path = (
os.path.join(pdfs_dest_dir, f"{collection_id}-{volume_name}") + ".pdf"
)
maybe_copy(book_src_path, book_dest_path)
# temp holder for papers in each volume
volume = dict()
# copy the paper PDFs
pdf_src_dir = os.path.join(root_path, "pdf")
for pdf_file in os.listdir(pdf_src_dir):
# Skip . files
if os.path.basename(pdf_file).startswith("."):
continue
# names are {abbrev}{number}.pdf
match = re.match(rf".*\.(\d+)\.pdf", pdf_file)
if match is not None:
paper_num = int(match[1])
paper_id_full = f"{collection_id}-{volume_name}.{paper_num}"
bib_path = os.path.join(
root_path,
"bib",
pdf_file.replace("/pdf", "/bib/").replace(".pdf", ".bib"),
)
pdf_src_path = os.path.join(pdf_src_dir, pdf_file)
pdf_dest_path = os.path.join(
pdfs_dest_dir, f"{collection_id}-{volume_name}.{paper_num}.pdf"
)
if not args.dry_run:
maybe_copy(pdf_src_path, pdf_dest_path)
volume[paper_num] = {
"anthology_id": paper_id_full,
"bib": bib_path,
"pdf": pdf_dest_path,
"attachments": [],
}
# copy the attachments
if os.path.exists(os.path.join(root_path, "additional")):
attachments_dest_dir = os.path.join(args.attachments_dir, venue_name)
if not os.path.exists(attachments_dest_dir):
os.makedirs(attachments_dest_dir)
for attachment_file in os.listdir(os.path.join(root_path, "additional")):
if os.path.basename(attachment_file).startswith("."):
continue
attachment_file_path = os.path.join(
root_path, "additional", attachment_file
)
# Find the attachment file, using a bit of a fuzzy
# match. The fuzzy match is because sometimes people
# generate the proceedings with the wrong venue
# code. If we correct it, we still need to be able to
# find the file.
match = re.match(
rf"{year}\..*-\w+\.(\d+)_?(\w+)\.(\w+)$", attachment_file
)
if match is None:
print(
f"* Warning: no attachment match for {attachment_file}",
file=sys.stderr,
)
sys.exit(2)
paper_num, type_, ext = match.groups()
paper_num = int(paper_num)
file_name = f"{collection_id}-{volume_name}.{paper_num}.{type_}.{ext}"
dest_path = os.path.join(attachments_dest_dir, file_name)
if not args.dry_run and not os.path.exists(dest_path):
log(f"Copying {attachment_file} -> {dest_path}", args.dry_run)
shutil.copyfile(attachment_file_path, dest_path)
volume[paper_num]["attachments"].append((dest_path, type_))
# create xml
collection_file = os.path.join(
args.anthology_dir, "data", "xml", f"{collection_id}.xml"
)
if os.path.exists(collection_file):
root_node = etree.parse(collection_file).getroot()
else:
root_node = make_simple_element("collection", attrib={"id": collection_id})
volume_node = make_simple_element(
"volume",
attrib={"id": volume_name, "ingest-date": args.ingest_date},
)
# Replace the existing one if present
existing_volume_node = root_node.find(f"./volume[@id='{volume_name}']")
for i, child in enumerate(root_node):
if child.attrib["id"] == volume_name:
root_node[i] = volume_node
break
else:
root_node.append(volume_node)
meta_node = None
for paper_num, paper in sorted(volume.items()):
paper_id_full = paper["anthology_id"]
bibfile = paper["bib"]
paper_node = bib2xml(bibfile, paper_id_full)
if paper_node.attrib["id"] == "0":
# create metadata subtree
meta_node = make_simple_element("meta", parent=volume_node)
title_node = paper_node.find("title")
title_node.tag = "booktitle"
meta_node.append(title_node)
for author_or_editor in chain(
paper_node.findall("./author"), paper_node.findall("./editor")
):
disamb_name, name_choice = disambiguate_name(
author_or_editor, paper_id_full
)
if name_choice != -1:
author_or_editor.attrib["id"] = disamb_name
person = PersonName.from_element(author_or_editor)
for name_part in author_or_editor:
name_part.text = correct_caps(name_part.text)
meta_node.append(author_or_editor)
author_or_editor.tag = "editor"
# Here, we grab the publisher from the meta file, in case it's not in the
# frontmatter paper. We don't handle the situation where it's in neither!
publisher_node = paper_node.find("publisher")
if publisher_node is None:
publisher_node = make_simple_element("publisher", meta["publisher"])
meta_node.append(publisher_node)
# Look for the address in the bib file, then the meta file
address_node = paper_node.find("address")
if address_node is None:
address_node = make_simple_element("address", meta["location"])
meta_node.append(address_node)
meta_node.append(paper_node.find("month"))
meta_node.append(paper_node.find("year"))
if book_dest_path is not None:
make_simple_element(
"url",
text=f"{collection_id}-{volume_name}",
attrib={"hash": compute_hash_from_file(book_dest_path)},
parent=meta_node,
)
# modify frontmatter tag
paper_node.tag = "frontmatter"
del paper_node.attrib["id"]
else:
# remove unneeded fields
for child in paper_node:
if child.tag in [
"editor",
"address",
"booktitle",
"publisher",
"year",
"month",
]:
paper_node.remove(child)
url = paper_node.find("./url")
if url is not None:
url.attrib["hash"] = compute_hash_from_file(paper["pdf"])
for path, type_ in paper["attachments"]:
make_simple_element(
"attachment",
text=os.path.basename(path),
attrib={
"type": type_,
"hash": compute_hash_from_file(path),
},
parent=paper_node,
)
if len(paper_node) > 0:
volume_node.append(paper_node)
# Normalize
for oldnode in paper_node:
normalize(oldnode, informat="latex")
# Adjust the language tag
language_node = paper_node.find("./language")
if language_node is not None:
try:
lang = iso639.languages.get(name=language_node.text)
except KeyError:
raise Exception(f"Can't find language '{language_node.text}'")
language_node.text = lang.part3
# Fix author names
for name_node in chain(
paper_node.findall("./author"), paper_node.findall("./editor")
):
disamb_name, name_choice = disambiguate_name(name_node, paper_id_full)
if name_choice != -1:
name_node.attrib["id"] = disamb_name
person = PersonName.from_element(name_node)
for name_part in name_node:
name_part.text = correct_caps(name_part.text)
# Other data from the meta file
if "isbn" in meta:
make_simple_element("isbn", meta["isbn"], parent=meta_node)
indent(root_node)
tree = etree.ElementTree(root_node)
tree.write(
collection_file, encoding="UTF-8", xml_declaration=True, with_tail=True
)
def main(args):
collections = defaultdict(OrderedDict)
volumes = {}
# Build list of volumes, confirm uniqueness
for proceedings in args.proceedings:
meta = read_meta(os.path.join(proceedings, "meta"))
meta["path"] = proceedings
meta["collection_id"] = collection_id = (meta["year"] + "." +
meta["abbrev"].lower())
volume_name = meta["volume_name"]
volume_full_id = f"{collection_id}-{volume_name}"
if volume_full_id in volumes:
print("Error: ")
collections[collection_id][volume_name] = {}
volumes[volume_full_id] = meta
# Copy over the PDFs and attachments
for volume, meta in volumes.items():
root_path = os.path.join(meta["path"], "cdrom")
collection_id = meta["collection_id"]
venue_name = meta["abbrev"].lower()
volume_name = meta["volume_name"]
pdfs_dest_dir = os.path.join(args.pdfs_dir, venue_name)
if not os.path.exists(pdfs_dest_dir):
os.makedirs(pdfs_dest_dir)
print(f"VOLUME: {volume}")
# copy the book
book_src_filename = meta["abbrev"] + "-" + meta["year"]
book_src_path = os.path.join(root_path, book_src_filename) + ".pdf"
book_dest_path = None
if os.path.exists(book_src_path):
book_dest_path = (
os.path.join(pdfs_dest_dir, f"{collection_id}-{volume_name}") +
".pdf")
log(f"Copying {book_src_path} -> {book_dest_path}", args.dry_run)
if not args.dry_run:
shutil.copyfile(book_src_path, book_dest_path)
# copy the paper PDFs
pdf_src_dir = os.path.join(root_path, "pdf")
for pdf_file in os.listdir(pdf_src_dir):
# names are {abbrev}{number}.pdf
abbrev = meta["abbrev"]
match = re.match(rf"{abbrev}(\d+)\.pdf", pdf_file)
if match is not None:
paper_num = int(match[1])
paper_id_full = f"{collection_id}-{volume_name}.{paper_num}"
bib_path = os.path.join(
root_path,
"bib",
pdf_file.replace("/pdf", "/bib/").replace(".pdf", ".bib"),
)
pdf_src_path = os.path.join(pdf_src_dir, pdf_file)
pdf_dest_path = os.path.join(
pdfs_dest_dir,
f"{collection_id}-{volume_name}.{paper_num}.pdf")
log(
f"Copying [{paper_id_full}] {pdf_src_path} -> {pdf_dest_path}",
args.dry_run,
)
if not args.dry_run:
shutil.copyfile(pdf_src_path, pdf_dest_path)
collections[collection_id][volume_name][paper_num] = {
"anthology_id": paper_id_full,
"bib": bib_path,
"pdf": pdf_dest_path,
"attachments": [],
}
# copy the attachments
if os.path.exists(os.path.join(root_path, "additional")):
attachments_dest_dir = os.path.join(args.attachments_dir,
collection_id)
if not os.path.exists(attachments_dest_dir):
os.makedirs(attachments_dest_dir)
for attachment_file in os.listdir(
os.path.join(root_path, "additional")):
match = re.match(rf"{abbrev}(\d+)_(\w+)\.(\w+)")
if match is not None:
paper_num, type_, ext = match.groups()
paper_num = int(paper_num)
file_name = f"{collection_id}-{volume_name}.{paper_num}.{type_}.{ext}"
dest_path = os.path.join(attachments_dest_dir, file_name)
log(f"Copying {attachment_file} -> {dest_path}",
args.dry_run)
if not args.dry_run:
shutil.copyfile(attachment_file, dest_path)
collections[collection_id][volume_name][paper_num][
"attachments"].append(dest_path)
people = AnthologyIndex(None,
srcdir=os.path.join(os.path.dirname(sys.argv[0]),
"..", "data"))
for collection_id, collection in collections.items():
collection_file = os.path.join(args.anthology_dir, "data", "xml",
f"{collection_id}.xml")
root_node = make_simple_element("collection",
attrib={"id": collection_id})
for volume_id, volume in collection.items():
volume_node = make_simple_element("volume",
attrib={"id": volume_id},
parent=root_node)
meta = None
for paper_num, paper in sorted(volume.items()):
paper_id_full = paper["anthology_id"]
bibfile = paper["bib"]
paper_node = bib2xml(bibfile, paper_id_full)
# print(etree.tostring(paper_node, pretty_print=True))
if paper_node.attrib["id"] == "0":
# create metadata subtree
meta = make_simple_element("meta", parent=volume_node)
title_node = paper_node.find("title")
title_node.tag = "booktitle"
meta.append(title_node)
for editor in paper_node.findall("editor"):
meta.append(editor)
meta.append(paper_node.find("publisher"))
meta.append(paper_node.find("address"))
meta.append(paper_node.find("month"))
meta.append(paper_node.find("year"))
if book_dest_path is not None:
make_simple_element(
"url",
text=f"{collection_id}-{volume_name}",
parent=meta)
# modify frontmatter tag
paper_node.tag = "frontmatter"
del paper_node.attrib["id"]
else:
# remove unneeded fields
for child in paper_node:
if child.tag in [
"editor",
"address",
"booktitle",
"publisher",
"year",
"month",
]:
paper_node.remove(child)
for attachment in paper["attachments"]:
make_simple_element(
"attachment",
text=attachment.path,
attrib={
"type": attachment.type,
},
parent=paper_node,
)
if len(paper_node) > 0:
volume_node.append(paper_node)
# Normalize
for paper in root_node.findall(".//paper"):
for oldnode in paper:
normalize(oldnode, informat="latex")
# Ensure names are properly identified
ambiguous = {}
for paper in root_node.findall(".//paper"):
anth_id = build_anthology_id(collection_id,
paper.getparent().attrib["id"],
paper.attrib["id"])
for node in chain(paper.findall("author"), paper.findall("editor")):
name = PersonName.from_element(node)
ids = people.get_ids(name)
if len(ids) > 1:
print(
f"WARNING ({anth_id}): ambiguous author {name}, defaulting to first of {ids}"
)
ambiguous[anth_id] = (name, ids)
node.attrib["id"] = ids[0]
indent(root_node)
tree = etree.ElementTree(root_node)
tree.write(collection_file,
encoding="UTF-8",
xml_declaration=True,
with_tail=True)
