dost

@@ -1,25 +1,15 @@

+#!/usr/bin/python3

+import argparse

 from collections import defaultdict

 import re

 import sys

 import time

 from typing import Generator

-def minima(options, key):

-    min_val = None

-    result = set()

-    for o in options:

-        k = key(o)

-        if min_val is None or k < min_val:

-            min_val = k

-            result = {o}

-        elif k == min_val:

-            result.add(o)

-    return result

-

 def cost(token_child: tuple[str, "Trie"]) -> tuple:

     token, child = token_child

     weight = child.deep_weight

-    return child.deep_usage / weight, -weight, child.min_depth, len(token), token

+    return child.deep_usage / weight, child.min_depth, len(token), token

 class Trie:

     def __init__(self, weight: int = 0):

@@ -73,13 +63,29 @@ class Trie:

             yield tokens

             self.use(tokens)

+if __name__ == "__main__":

+    parser = argparse.ArgumentParser(prog="dost", description="Diverse-order sampling by trie")

+    parser.add_argument("files", default="-", nargs="*")

+    parser.add_argument(

+        "-s", "--separators", default=" \t",

+        help="characters that separate tokens in an input item (default space and tab)"

+    )

+    parser.add_argument(

+        "-t", "--token",

+        help="structural regexp for a token of an input item, overrides -s"

+    )

+    args = parser.parse_args()

     re_flags = re.MULTILINE | re.DOTALL

-token_pattern = "[-_ \t]*[^-_\\s]+"

+    if args.token:

+        token_pattern = args.token

+    else:

+        token_pattern = f"[{args.separators}]*[^{args.separators}\r\n]+"

     entry_pattern = re.compile(f"({token_pattern})+", re_flags)

     token_pattern = re.compile(token_pattern, re_flags)

-words = [re.findall(token_pattern, m.group(0)) for m in re.finditer(entry_pattern, sys.stdin.read())]

     t = Trie()

-for w in words:

-    t.add(w)

-for w in t.traverse():

-    print("".join(w))

+    for filename in args.files:

+        with sys.stdin if filename == "-" else open(filename, "r") as handle:

+            for word in re.finditer(entry_pattern, handle.read()):

+                t.add(re.findall(token_pattern, word.group(0)))

+    for word in t.traverse():

+        print("".join(word))