|
| 1 | +""" |
| 2 | +Analogy queries for embeddings. |
| 3 | +""" |
| 4 | +import argparse |
| 5 | +import sys |
| 6 | +from typing import List, Set |
| 7 | + |
| 8 | +from finalfusion.scripts.util import Format |
| 9 | + |
| 10 | + |
| 11 | +def main() -> None: # pylint: disable=missing-function-docstring |
| 12 | + formats = ["word2vec", "finalfusion", "fasttext", "text", "textdims"] |
| 13 | + parser = argparse.ArgumentParser(prog="ffp-analogy", |
| 14 | + description="Analogy queries.") |
| 15 | + parser.add_argument( |
| 16 | + "-f", |
| 17 | + "--format", |
| 18 | + choices=formats, |
| 19 | + type=str, |
| 20 | + default="finalfusion", |
| 21 | + help=f"Valid choices: {formats} Default: 'finalfusion'", |
| 22 | + metavar="INPUT_FORMAT") |
| 23 | + parser.add_argument("embeddings", |
| 24 | + help="Input embeddings", |
| 25 | + type=str, |
| 26 | + metavar="EMBEDDINGS") |
| 27 | + parser.add_argument( |
| 28 | + "-i", |
| 29 | + "--include", |
| 30 | + choices=["a", "b", "c"], |
| 31 | + nargs="+", |
| 32 | + default=[], |
| 33 | + help= |
| 34 | + "Specify query parts that should be allowed as answers. Valid choices: ['a', 'b', 'c']" |
| 35 | + ) |
| 36 | + parser.add_argument("-k", |
| 37 | + type=int, |
| 38 | + default=10, |
| 39 | + help=f"Number of neighbours. Default: 1", |
| 40 | + metavar="K") |
| 41 | + parser.add_argument("input", nargs='?', default=0) |
| 42 | + args = parser.parse_args() |
| 43 | + if args.include != [] and len(args.include) > 3: |
| 44 | + print("-i/--include can take up to 3 unique values: a, b and c.", |
| 45 | + file=sys.stderr) |
| 46 | + sys.exit(1) |
| 47 | + embeds = Format(args.format).load(args.embeddings) |
| 48 | + with open(args.input) as queries: |
| 49 | + for query in queries: |
| 50 | + query_a, query_b, query_c = query.strip().split() |
| 51 | + skips = get_skips(query_a, query_b, query_c, args.include) |
| 52 | + res = embeds.analogy(query_a, |
| 53 | + query_b, |
| 54 | + query_c, |
| 55 | + k=args.k, |
| 56 | + skip=skips) |
| 57 | + if res is None: |
| 58 | + print( |
| 59 | + f"Could not compute for: {query_a} : {query_b}, {query_c} : ? ", |
| 60 | + file=sys.stderr) |
| 61 | + else: |
| 62 | + print("\n".join(f"{ws.word} {ws.similarity}" for ws in res)) |
| 63 | + |
| 64 | + |
| 65 | +def get_skips( # pylint: disable=missing-function-docstring |
| 66 | + query_a: str, query_b: str, query_c: str, |
| 67 | + includes: List[str]) -> Set[str]: |
| 68 | + if includes == []: |
| 69 | + return {query_c, query_b, query_a} |
| 70 | + skips = set() |
| 71 | + if 'a' not in includes: |
| 72 | + skips.add(query_a) |
| 73 | + if 'b' not in includes: |
| 74 | + skips.add(query_b) |
| 75 | + if 'c' not in includes: |
| 76 | + skips.add(query_b) |
| 77 | + return skips |
| 78 | + |
| 79 | + |
| 80 | +if __name__ == '__main__': |
| 81 | + main() |
0 commit comments