|
| 1 | +#!/usr/bin/env python3 |
| 2 | +"""Build a StringTrie from TSV input. |
| 3 | +
|
| 4 | +Input format: |
| 5 | + one UTF-8 line per record, with "key<TAB>value" |
| 6 | +""" |
| 7 | + |
| 8 | +from __future__ import annotations |
| 9 | + |
| 10 | +import argparse |
| 11 | +import sys |
| 12 | +from contextlib import nullcontext |
| 13 | + |
| 14 | +import marisa_trie |
| 15 | + |
| 16 | + |
| 17 | +def parse_args() -> argparse.Namespace: |
| 18 | + parser = argparse.ArgumentParser( |
| 19 | + description="Build a marisa_trie.StringTrie from TSV pairs." |
| 20 | + ) |
| 21 | + parser.add_argument( |
| 22 | + "input", |
| 23 | + help="Input TSV path, or '-' for stdin.", |
| 24 | + ) |
| 25 | + parser.add_argument( |
| 26 | + "output", |
| 27 | + help="Output binary path for StringTrie.save().", |
| 28 | + ) |
| 29 | + parser.add_argument( |
| 30 | + "--progress-every", |
| 31 | + type=int, |
| 32 | + default=0, |
| 33 | + help="Print progress every N input lines (0 disables progress logs).", |
| 34 | + ) |
| 35 | + return parser.parse_args() |
| 36 | + |
| 37 | + |
| 38 | +def iter_tsv_pairs(path: str, progress_every: int): |
| 39 | + if path == "-": |
| 40 | + ctx = nullcontext(sys.stdin) |
| 41 | + else: |
| 42 | + ctx = open(path, "r", encoding="utf-8", newline="") |
| 43 | + |
| 44 | + with ctx as f: |
| 45 | + for lineno, raw in enumerate(f, 1): |
| 46 | + line = raw.rstrip("\n") |
| 47 | + if not line: |
| 48 | + continue |
| 49 | + if progress_every > 0 and lineno % progress_every == 0: |
| 50 | + print(f"read {lineno} lines", file=sys.stderr) |
| 51 | + if "\t" not in line: |
| 52 | + raise ValueError( |
| 53 | + f"line {lineno}: expected a TAB separator between key and value" |
| 54 | + ) |
| 55 | + key, value = line.split("\t", 1) |
| 56 | + yield key, value |
| 57 | + |
| 58 | + |
| 59 | +def main() -> int: |
| 60 | + args = parse_args() |
| 61 | + try: |
| 62 | + trie = marisa_trie.StringTrie( |
| 63 | + iter_tsv_pairs(args.input, args.progress_every) |
| 64 | + ) |
| 65 | + trie.save(args.output) |
| 66 | + except Exception as exc: |
| 67 | + print(f"error: {exc}", file=sys.stderr) |
| 68 | + return 1 |
| 69 | + return 0 |
| 70 | + |
| 71 | + |
| 72 | +if __name__ == "__main__": |
| 73 | + raise SystemExit(main()) |
0 commit comments