From 142a2e4437fd155264d24687c5362daa33db9b59 Mon Sep 17 00:00:00 2001 From: Cyber MacGeddon Date: Fri, 6 Dec 2024 15:14:45 +0000 Subject: [PATCH] Dump some metadata --- trustgraph-cli/scripts/tg-dump-msgpack | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/trustgraph-cli/scripts/tg-dump-msgpack b/trustgraph-cli/scripts/tg-dump-msgpack index 537d35e3..18819649 100755 --- a/trustgraph-cli/scripts/tg-dump-msgpack +++ b/trustgraph-cli/scripts/tg-dump-msgpack @@ -25,7 +25,7 @@ def summary(input_file, action): triples = set() - max_records = 10 + max_records = 1000000 with open(input_file, 'rb') as f: @@ -41,22 +41,20 @@ def summary(input_file, action): if msg[0] == "t": for elt in msg[1]["m"]["m"]: - triples.add( + triples.add(( elt["s"]["v"], elt["p"]["v"], elt["o"]["v"], - ) + )) if rec_count > max_records: break rec_count += 1 print("Vector dimension:", vector_dim) - print(triples) -# print(f"Graph-embeddings: {ge:10d}") -# for k, v in sorted(ge_dim.items()): -# print(f" Dimension {k}: {v:10d}") -# print(f"Triples: {t:10d}") + for t in triples: + if t[1] == "http://www.w3.org/2000/01/rdf-schema#label": + print("-", t[2]) def main():