def main(ranking, all_lemmas, language_code, top_n, dump_enriched,
dump_top_lemmas, pid_batch, prop_batch):
"""
Extract FrameNet data given a ranking of corpus Lexical Units (lemmas).
Return frames only if FEs map to Wikidata properties via exact matching of labels and aliases.
"""
logger.info("Loading ranked corpus Lexical Units (LUs) from '%s' ..." %
ranking.name)
# Remember to preserve the order
lus = json.load(ranking, object_pairs_hook=OrderedDict)
logger.info("Loaded %d LUs" % len(lus))
logger.info("Will consider the top %d LUs" % top_n)
top = get_top_n_lus(lus, top_n)
logger.debug("Top LUs: %s" % top)
logger.info("Retrieving the full list of Wikidata properties ...")
all_pids = get_property_ids(pid_batch)
all_properties = get_entities(all_pids, prop_batch)
logger.info("Extracting label and aliases only ...")
clean_properties = get_labels_and_aliases(all_properties, language_code)
enriched = intersect_lemmas_with_framenet(top, clean_properties)
logger.info("Managed to enrich %d LUs with FrameNet data" % len(enriched))
logger.info("Dumping top enriched LUs to '%s' ..." % dump_enriched.name)
json.dump(enriched, dump_enriched, indent=2)
top = extract_top_corpus_tokens(enriched, json.load(all_lemmas))
logger.info("Dumping top lemmas with tokens to '%s' ..." %
dump_top_lemmas.name)
json.dump(top, dump_top_lemmas, indent=2)
return 0
def main(ranking, all_lemmas, language_code, top_n, dump_enriched, dump_top_lemmas, pid_batch, prop_batch):
"""
Extract FrameNet data given a ranking of corpus Lexical Units (lemmas).
Return frames only if FEs map to Wikidata properties via exact matching of labels and aliases.
"""
logger.info("Loading ranked corpus Lexical Units (LUs) from '%s' ..." % ranking.name)
# Remember to preserve the order
lus = json.load(ranking, object_pairs_hook=OrderedDict)
logger.info("Loaded %d LUs" % len(lus))
logger.info("Will consider the top %d LUs" % top_n)
top = get_top_n_lus(lus, top_n)
logger.debug("Top LUs: %s" % top)
logger.info("Retrieving the full list of Wikidata properties ...")
all_pids = get_property_ids(pid_batch)
all_properties = get_entities(all_pids, prop_batch)
logger.info("Extracting label and aliases only ...")
clean_properties = get_labels_and_aliases(all_properties, language_code)
enriched = intersect_lemmas_with_framenet(top, clean_properties)
logger.info("Managed to enrich %d LUs with FrameNet data" % len(enriched))
logger.info("Dumping top enriched LUs to '%s' ..." % dump_enriched.name)
json.dump(enriched, dump_enriched, indent=2)
top = extract_top_corpus_tokens(enriched, json.load(all_lemmas))
logger.info("Dumping top lemmas with tokens to '%s' ..." % dump_top_lemmas.name)
json.dump(top, dump_top_lemmas, indent=2)
return 0
def main(corpus_frames, language_code, pid_batch, prop_batch, outfile):
""" Map FEs to Wikidata properties via exact matches """
all_pids = get_property_ids(pid_batch)
all_properties = get_entities(all_pids, prop_batch)
clean_properties = get_labels_and_aliases(all_properties, language_code)
logger.debug(json.dumps(clean_properties, indent=2))
logger.info("Computing exact matches mapping ...")
exact_matches = compute_exact_matches(json.load(corpus_frames), clean_properties)
logger.info("Total matches: %d Will dump to '%s' ..." %(len(exact_matches), outfile.name))
json.dump(exact_matches, outfile, indent=2)
return 0