zxcvbn-pl 0.1.0__tar.gz

zxcvbn_pl-0.1.0/.gitignore

@@ -0,0 +1,9 @@
+__pycache__/
+*.py[cod]
+.venv/
+dist/
+build/
+*.egg-info/
+.pytest_cache/
+.ruff_cache/
+.DS_Store

zxcvbn_pl-0.1.0/CHANGELOG.md

@@ -0,0 +1,7 @@
+# Changelog
+
+## 0.1.0 — 2026-10-24
+
+- First release: `pl_common_words`, `pl_first_names` and `pl_last_names`
+  dictionaries built from `@zxcvbn-ts/language-pl` 4.1.1, with
+  diacritics-free variants; `zxcvbn_pl.register()`.

zxcvbn_pl-0.1.0/LICENSE

@@ -0,0 +1,24 @@
+MIT License
+
+Copyright (c) 2026 Michał Pasternak, IPLweb
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+The bundled dictionaries in src/zxcvbn_pl/data are derived from third-party
+data under their own licenses (ODC-BY, CC0 1.0) — see NOTICE.md.

zxcvbn_pl-0.1.0/NOTICE.md

@@ -0,0 +1,37 @@
+# Data sources and licenses
+
+The code of zxcvbn-pl is licensed under the MIT License (see `LICENSE`).
+The dictionaries in `src/zxcvbn_pl/data/*.marisa` are derived from the
+[`@zxcvbn-ts/language-pl`](https://github.com/zxcvbn-ts/zxcvbn/tree/master/packages/languages/pl)
+npm package, version 4.1.1 (MIT), and carry the licenses of their sources:
+
+## `pl_common_words.marisa`
+
+Derived from OpenSubtitles via OPUS (`commonWords.json` in
+`@zxcvbn-ts/language-pl`).
+
+- Source dataset: OpenSubtitles 2024, https://opus.nlpl.eu/OpenSubtitles.php
+- Provider: Helsinki-NLP / OPUS
+- License: **ODC-BY** (Open Data Commons Attribution License),
+  https://opendatacommons.org/licenses/by/
+- Original data is attributed to subtitle contributors and rights holders
+  (http://www.opensubtitles.org/).
+
+This package contains a processed derivative: tokenized text, frequency
+aggregation, normalization and filtering (by zxcvbn-ts), then lowercasing,
+dropping entries shorter than 3 characters or containing non-letters, adding
+diacritics-free variants and converting to a marisa-trie index (by
+zxcvbn-pl). Any redistribution must retain this notice.
+
+## `pl_first_names.marisa`, `pl_last_names.marisa`
+
+Derived from the lists of first names and last names in the Polish PESEL
+registry, published by the Ministry of Digital Affairs on dane.gov.pl:
+
+- Lista imion występujących w rejestrze PESEL,
+  https://dane.gov.pl/pl/dataset/1667 — **CC0 1.0**
+- Nazwiska występujące w rejestrze PESEL,
+  https://dane.gov.pl/pl/dataset/1681 — **CC0 1.0**
+
+zxcvbn-ts keeps names with at least 200 occurrences; zxcvbn-pl merges the
+male and female lists, lowercases them and adds diacritics-free variants.

zxcvbn_pl-0.1.0/PKG-INFO

@@ -0,0 +1,152 @@
+Metadata-Version: 2.5
+Name: zxcvbn-pl
+Version: 0.1.0
+Summary: Polish dictionaries (common words, first and last names) for zxcvbn password strength estimation.
+Project-URL: Homepage, https://github.com/iplweb/zxcvbn-pl
+Project-URL: Repository, https://github.com/iplweb/zxcvbn-pl
+Project-URL: Changelog, https://github.com/iplweb/zxcvbn-pl/blob/main/CHANGELOG.md
+Author-email: Michał Pasternak <m@iplweb.pl>
+License: MIT
+License-File: LICENSE
+License-File: NOTICE.md
+Keywords: dictionary,hasło,password,password-strength,polish,security,zxcvbn
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Natural Language :: Polish
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Security
+Requires-Python: >=3.9
+Requires-Dist: marisa-trie>=1.0
+Requires-Dist: zxcvbn>=4.4
+Description-Content-Type: text/markdown
+
+# zxcvbn-pl
+
+[![PyPI](https://img.shields.io/pypi/v/zxcvbn-pl.svg)](https://pypi.org/project/zxcvbn-pl/)
+[![CI](https://github.com/iplweb/zxcvbn-pl/actions/workflows/ci.yml/badge.svg)](https://github.com/iplweb/zxcvbn-pl/actions/workflows/ci.yml)
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+
+Polskie słowniki dla [zxcvbn](https://github.com/dwolfhub/zxcvbn-python)
+(port w Pythonie) — pospolite słowa, imiona i nazwiska.
+
+*Polish dictionaries for the Python port of zxcvbn — common words, first
+names and last names. [English below.](#english)*
+
+---
+
+## 🇵🇱 Po polsku
+
+### Po co
+
+zxcvbn szacuje, ile prób potrzeba do odgadnięcia hasła, rozpoznając w nim
+słowa ze słowników, powtórzenia, sekwencje, układy klawiatury i daty. Ma
+jednak **tylko angielskie słowniki** — polskie słowo to dla niego losowe
+litery. Przez to hasła łatwe do zgadnięcia dla każdego, kto ma polską listę
+słów, dostają najwyższą ocenę:
+
+| Hasło | zxcvbn | zxcvbn + zxcvbn-pl |
+|---|---|---|
+| `słoneczko123456` | 4 | 2 |
+| `pomidorowa123456` | 3 | 2 |
+| `Katarzyna1985!` | 3 | 2 |
+| `agnieszka.nowak1` | 4 | 3 |
+| `Wojciechowski77` | 2 | 1 |
+| `ptaszek-na-drzewie` | 4 | 4 |
+| `zielonaherbatazcytryna` | 4 | 4 |
+
+Frazy z kilku słów nadal są oceniane wysoko — tak jak zaleca NIST
+SP 800-63B. Spada ocena haseł z jednego pospolitego słowa, imienia lub
+nazwiska z dopiskiem. Efekt jest umiarkowany: słowa spoza ok. 50 tys.
+najczęstszych (np. `kaczuszka2000!!`) nadal są dla zxcvbn losowymi literami.
+
+Port zxcvbn w JavaScripcie (zxcvbn-ts) ma pakiety językowe; port w Pythonie
+— nie. Ten pakiet wypełnia tę lukę danymi z `@zxcvbn-ts/language-pl`.
+
+### Instalacja i użycie
+
+```bash
+pip install zxcvbn-pl
+```
+
+```python
+import zxcvbn_pl
+from zxcvbn import zxcvbn
+
+zxcvbn_pl.register()  # raz, np. przy starcie aplikacji; wielokrotne wywołanie nic nie psuje
+
+wynik = zxcvbn("słoneczko123456")
+wynik["score"]  # 2
+[m.get("dictionary_name") for m in wynik["sequence"]]
+# ['pl_common_words', 'passwords']
+```
+
+`register()` dokłada trzy słowniki do globalnego rejestru zxcvbn
+(`zxcvbn.matching.RANKED_DICTIONARIES`), więc działa dla każdego kolejnego
+wywołania `zxcvbn()` w procesie:
+
+| Nazwa (`dictionary_name`) | Zawartość |
+|---|---|
+| `pl_common_words` | ok. 52 tys. najczęstszych słów (napisy filmowe) |
+| `pl_first_names` | imiona z rejestru PESEL (≥ 200 osób) |
+| `pl_last_names` | nazwiska z rejestru PESEL (≥ 200 osób) |
+
+Każde słowo z polskimi znakami jest też w wariancie bez nich (`żółw` →
+`zolw`) z tą samą pozycją. Pomijane są wpisy krótsze niż 3 znaki.
+
+Słowniki to indeksy [marisa-trie](https://github.com/pytries/marisa-trie)
+czytane przez mmap (~1 MB na dysku, znikomy narzut RAM).
+
+### Uwaga o wątkach
+
+zxcvbn trzyma słowniki — także `user_inputs` przekazane do `zxcvbn()` —
+w globalnym słowniku modułu. `register()` jest bezpieczne, ale jeśli
+wywołujesz `zxcvbn(..., user_inputs=...)` z wielu wątków, owiń wywołanie
+blokadą (to ograniczenie samego zxcvbn, nie tego pakietu).
+
+### Dane i licencje
+
+Kod: MIT. Dane pochodzą z `@zxcvbn-ts/language-pl` i mają własne licencje —
+szczegóły w [NOTICE.md](NOTICE.md):
+
+- pospolite słowa: OPUS OpenSubtitles 2024 — **ODC-BY** (wymaga podania źródła),
+- imiona i nazwiska: rejestr PESEL, dane.gov.pl — **CC0 1.0**.
+
+Odbudowa słowników (deterministyczna, z weryfikacją sha512 paczki npm):
+
+```bash
+uv run python scripts/build_dictionaries.py
+```
+
+---
+
+## English
+
+zxcvbn ships with English dictionaries only, so Polish words look like random
+letters to it and passwords such as `słoneczko123456` or `Katarzyna1985!`
+get the top score. `zxcvbn-pl` adds Polish common words, first names and last
+names (plus diacritics-free variants) to the Python port of zxcvbn:
+
+```python
+import zxcvbn_pl
+from zxcvbn import zxcvbn
+
+zxcvbn_pl.register()
+zxcvbn("słoneczko123456")["score"]  # 2 instead of 4
+```
+
+`register()` is idempotent and adds three dictionaries (`pl_common_words`,
+`pl_first_names`, `pl_last_names`) to zxcvbn's global
+`zxcvbn.matching.RANKED_DICTIONARIES`. Multi-word passphrases still score
+high, in line with NIST SP 800-63B.
+
+The data comes from the `@zxcvbn-ts/language-pl` npm package: common words
+from OPUS OpenSubtitles 2024 (ODC-BY), first and last names from the Polish
+PESEL registry published on dane.gov.pl (CC0 1.0). See [NOTICE.md](NOTICE.md).
+The code is MIT-licensed.

zxcvbn_pl-0.1.0/README.md

@@ -0,0 +1,123 @@
+# zxcvbn-pl
+
+[![PyPI](https://img.shields.io/pypi/v/zxcvbn-pl.svg)](https://pypi.org/project/zxcvbn-pl/)
+[![CI](https://github.com/iplweb/zxcvbn-pl/actions/workflows/ci.yml/badge.svg)](https://github.com/iplweb/zxcvbn-pl/actions/workflows/ci.yml)
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+
+Polskie słowniki dla [zxcvbn](https://github.com/dwolfhub/zxcvbn-python)
+(port w Pythonie) — pospolite słowa, imiona i nazwiska.
+
+*Polish dictionaries for the Python port of zxcvbn — common words, first
+names and last names. [English below.](#english)*
+
+---
+
+## 🇵🇱 Po polsku
+
+### Po co
+
+zxcvbn szacuje, ile prób potrzeba do odgadnięcia hasła, rozpoznając w nim
+słowa ze słowników, powtórzenia, sekwencje, układy klawiatury i daty. Ma
+jednak **tylko angielskie słowniki** — polskie słowo to dla niego losowe
+litery. Przez to hasła łatwe do zgadnięcia dla każdego, kto ma polską listę
+słów, dostają najwyższą ocenę:
+
+| Hasło | zxcvbn | zxcvbn + zxcvbn-pl |
+|---|---|---|
+| `słoneczko123456` | 4 | 2 |
+| `pomidorowa123456` | 3 | 2 |
+| `Katarzyna1985!` | 3 | 2 |
+| `agnieszka.nowak1` | 4 | 3 |
+| `Wojciechowski77` | 2 | 1 |
+| `ptaszek-na-drzewie` | 4 | 4 |
+| `zielonaherbatazcytryna` | 4 | 4 |
+
+Frazy z kilku słów nadal są oceniane wysoko — tak jak zaleca NIST
+SP 800-63B. Spada ocena haseł z jednego pospolitego słowa, imienia lub
+nazwiska z dopiskiem. Efekt jest umiarkowany: słowa spoza ok. 50 tys.
+najczęstszych (np. `kaczuszka2000!!`) nadal są dla zxcvbn losowymi literami.
+
+Port zxcvbn w JavaScripcie (zxcvbn-ts) ma pakiety językowe; port w Pythonie
+— nie. Ten pakiet wypełnia tę lukę danymi z `@zxcvbn-ts/language-pl`.
+
+### Instalacja i użycie
+
+```bash
+pip install zxcvbn-pl
+```
+
+```python
+import zxcvbn_pl
+from zxcvbn import zxcvbn
+
+zxcvbn_pl.register()  # raz, np. przy starcie aplikacji; wielokrotne wywołanie nic nie psuje
+
+wynik = zxcvbn("słoneczko123456")
+wynik["score"]  # 2
+[m.get("dictionary_name") for m in wynik["sequence"]]
+# ['pl_common_words', 'passwords']
+```
+
+`register()` dokłada trzy słowniki do globalnego rejestru zxcvbn
+(`zxcvbn.matching.RANKED_DICTIONARIES`), więc działa dla każdego kolejnego
+wywołania `zxcvbn()` w procesie:
+
+| Nazwa (`dictionary_name`) | Zawartość |
+|---|---|
+| `pl_common_words` | ok. 52 tys. najczęstszych słów (napisy filmowe) |
+| `pl_first_names` | imiona z rejestru PESEL (≥ 200 osób) |
+| `pl_last_names` | nazwiska z rejestru PESEL (≥ 200 osób) |
+
+Każde słowo z polskimi znakami jest też w wariancie bez nich (`żółw` →
+`zolw`) z tą samą pozycją. Pomijane są wpisy krótsze niż 3 znaki.
+
+Słowniki to indeksy [marisa-trie](https://github.com/pytries/marisa-trie)
+czytane przez mmap (~1 MB na dysku, znikomy narzut RAM).
+
+### Uwaga o wątkach
+
+zxcvbn trzyma słowniki — także `user_inputs` przekazane do `zxcvbn()` —
+w globalnym słowniku modułu. `register()` jest bezpieczne, ale jeśli
+wywołujesz `zxcvbn(..., user_inputs=...)` z wielu wątków, owiń wywołanie
+blokadą (to ograniczenie samego zxcvbn, nie tego pakietu).
+
+### Dane i licencje
+
+Kod: MIT. Dane pochodzą z `@zxcvbn-ts/language-pl` i mają własne licencje —
+szczegóły w [NOTICE.md](NOTICE.md):
+
+- pospolite słowa: OPUS OpenSubtitles 2024 — **ODC-BY** (wymaga podania źródła),
+- imiona i nazwiska: rejestr PESEL, dane.gov.pl — **CC0 1.0**.
+
+Odbudowa słowników (deterministyczna, z weryfikacją sha512 paczki npm):
+
+```bash
+uv run python scripts/build_dictionaries.py
+```
+
+---
+
+## English
+
+zxcvbn ships with English dictionaries only, so Polish words look like random
+letters to it and passwords such as `słoneczko123456` or `Katarzyna1985!`
+get the top score. `zxcvbn-pl` adds Polish common words, first names and last
+names (plus diacritics-free variants) to the Python port of zxcvbn:
+
+```python
+import zxcvbn_pl
+from zxcvbn import zxcvbn
+
+zxcvbn_pl.register()
+zxcvbn("słoneczko123456")["score"]  # 2 instead of 4
+```
+
+`register()` is idempotent and adds three dictionaries (`pl_common_words`,
+`pl_first_names`, `pl_last_names`) to zxcvbn's global
+`zxcvbn.matching.RANKED_DICTIONARIES`. Multi-word passphrases still score
+high, in line with NIST SP 800-63B.
+
+The data comes from the `@zxcvbn-ts/language-pl` npm package: common words
+from OPUS OpenSubtitles 2024 (ODC-BY), first and last names from the Polish
+PESEL registry published on dane.gov.pl (CC0 1.0). See [NOTICE.md](NOTICE.md).
+The code is MIT-licensed.

zxcvbn_pl-0.1.0/pyproject.toml

@@ -0,0 +1,59 @@
+[project]
+name = "zxcvbn-pl"
+version = "0.1.0"
+description = "Polish dictionaries (common words, first and last names) for zxcvbn password strength estimation."
+readme = "README.md"
+requires-python = ">=3.9"
+license = { text = "MIT" }
+authors = [{ name = "Michał Pasternak", email = "m@iplweb.pl" }]
+keywords = ["zxcvbn", "password", "password-strength", "polish", "dictionary", "hasło", "security"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Natural Language :: Polish",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+    "Topic :: Security",
+]
+dependencies = [
+    "marisa-trie>=1.0",
+    "zxcvbn>=4.4",
+]
+
+[project.urls]
+Homepage = "https://github.com/iplweb/zxcvbn-pl"
+Repository = "https://github.com/iplweb/zxcvbn-pl"
+Changelog = "https://github.com/iplweb/zxcvbn-pl/blob/main/CHANGELOG.md"
+
+[dependency-groups]
+dev = [
+    "pytest>=8.0",
+    "ruff>=0.6",
+    "pre-commit>=3.5",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/zxcvbn_pl"]
+
+[tool.hatch.build.targets.sdist]
+only-include = ["src", "tests", "scripts", "README.md", "LICENSE", "NOTICE.md", "CHANGELOG.md"]
+
+[tool.ruff]
+line-length = 88
+target-version = "py39"
+
+[tool.ruff.lint]
+select = ["E", "F", "W", "I", "B", "UP"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]

zxcvbn_pl-0.1.0/scripts/build_dictionaries.py

@@ -0,0 +1,104 @@
+"""Build the Polish zxcvbn dictionaries shipped in ``src/zxcvbn_pl/data``.
+
+Source data: the ``@zxcvbn-ts/language-pl`` npm package (pinned version,
+verified against the registry's sha512 integrity). Each upstream list is
+ordered by frequency; the position becomes the zxcvbn rank. Male and female
+lists are merged (lower rank wins). For every word with Polish diacritics a
+variant without them is added with the same rank (``żółw`` -> ``zolw``) —
+that is how such words are usually typed in passwords.
+
+Run from the repository root (needs only the runtime dependency)::
+
+    uv run python scripts/build_dictionaries.py
+
+The output is deterministic: the same upstream version gives byte-identical
+``.marisa`` files.
+"""
+
+import base64
+import hashlib
+import io
+import json
+import tarfile
+import unicodedata
+import urllib.request
+from pathlib import Path
+
+import marisa_trie
+
+UPSTREAM_VERSION = "4.1.1"
+UPSTREAM_URL = (
+    "https://registry.npmjs.org/@zxcvbn-ts/language-pl/-/"
+    f"language-pl-{UPSTREAM_VERSION}.tgz"
+)
+UPSTREAM_SHA512 = (
+    "hF6Qu9cyHx7sSEzNOQrJQntq8geoincsGvlOC0wkD43LnUVSTR49MKxdsifsemlQgtxR7aUK"
+    "DgK/e/RFHOljoQ=="
+)
+
+# Output dictionary name -> upstream lists merged into it.
+DICTIONARIES = {
+    "pl_common_words": ["commonWords"],
+    "pl_first_names": ["femaleFirstnames", "maleFirstnames"],
+    "pl_last_names": ["femaleLastnames", "maleLastnames"],
+}
+
+# Single letters and two-letter tokens match everywhere and only add noise.
+MIN_LENGTH = 3
+
+OUTPUT_DIR = Path(__file__).resolve().parent.parent / "src" / "zxcvbn_pl" / "data"
+
+
+def strip_diacritics(word):
+    word = word.replace("ł", "l")
+    word = unicodedata.normalize("NFKD", word)
+    return "".join(c for c in word if not unicodedata.combining(c))
+
+
+def download():
+    with urllib.request.urlopen(UPSTREAM_URL, timeout=60) as response:
+        payload = response.read()
+    digest = base64.b64encode(hashlib.sha512(payload).digest()).decode()
+    if digest != UPSTREAM_SHA512:
+        raise RuntimeError(
+            f"sha512 mismatch for {UPSTREAM_URL}: got {digest}, "
+            f"expected {UPSTREAM_SHA512}"
+        )
+    return payload
+
+
+def read_lists(payload):
+    lists = {}
+    with tarfile.open(fileobj=io.BytesIO(payload), mode="r:gz") as tar:
+        for name in {n for names in DICTIONARIES.values() for n in names}:
+            member = tar.extractfile(f"package/src/{name}.json")
+            lists[name] = json.load(member)
+    return lists
+
+
+def build_ranks(upstream_lists):
+    ranks = {}
+    for words in upstream_lists:
+        for position, word in enumerate(words, start=1):
+            word = word.strip().lower()
+            if len(word) < MIN_LENGTH or not word.isalpha():
+                continue
+            for variant in (word, strip_diacritics(word)):
+                if position < ranks.get(variant, position + 1):
+                    ranks[variant] = position
+    return ranks
+
+
+def main():
+    lists = read_lists(download())
+    OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
+    for dictionary, sources in DICTIONARIES.items():
+        ranks = build_ranks(lists[source] for source in sources)
+        trie = marisa_trie.RecordTrie("<I", ((w, (r,)) for w, r in ranks.items()))
+        path = OUTPUT_DIR / f"{dictionary}.marisa"
+        trie.save(str(path))
+        print(f"{path.name}: {len(trie)} entries, {path.stat().st_size // 1024} KB")
+
+
+if __name__ == "__main__":
+    main()

zxcvbn_pl-0.1.0/src/zxcvbn_pl/__init__.py

@@ -0,0 +1,78 @@
+"""Polish dictionaries for zxcvbn (the Python port).
+
+zxcvbn ships English dictionaries only, so to zxcvbn a Polish word looks like
+random letters and passwords such as ``słoneczko123456`` or
+``Katarzyna1985!`` get the top score. This package adds Polish common words,
+first names and last names (with diacritics-free variants) to zxcvbn::
+
+    import zxcvbn_pl
+    from zxcvbn import zxcvbn
+
+    zxcvbn_pl.register()
+    zxcvbn("słoneczko123456")["score"]  # 2 instead of 4
+
+The dictionaries are marisa-trie indices read via mmap (no big RAM cost).
+"""
+
+import threading
+from pathlib import Path
+
+import marisa_trie
+
+__all__ = ["DICTIONARY_NAMES", "MarisaRankedDictionary", "dictionaries", "register"]
+__version__ = "0.1.0"
+
+DATA_DIR = Path(__file__).parent / "data"
+
+#: Names under which the dictionaries are registered in zxcvbn — they show up
+#: as ``dictionary_name`` in zxcvbn's match sequence.
+DICTIONARY_NAMES = ("pl_common_words", "pl_first_names", "pl_last_names")
+
+_lock = threading.Lock()
+_dictionaries = None
+
+
+class MarisaRankedDictionary:
+    """Read-only ``word -> rank`` mapping over a marisa-trie index.
+
+    zxcvbn only needs ``word in d`` and ``d[word]`` from a ranked dictionary,
+    so this is a drop-in replacement for the plain dicts it builds itself.
+    """
+
+    def __init__(self, path):
+        self._trie = marisa_trie.RecordTrie("<I").mmap(str(path))
+
+    def __contains__(self, word):
+        return word in self._trie
+
+    def __getitem__(self, word):
+        return self._trie[word][0][0]
+
+    def __len__(self):
+        return len(self._trie)
+
+    def get(self, word, default=None):
+        return self[word] if word in self else default
+
+
+def dictionaries():
+    """Return ``{name: MarisaRankedDictionary}`` for all Polish dictionaries."""
+    global _dictionaries
+    with _lock:
+        if _dictionaries is None:
+            _dictionaries = {
+                name: MarisaRankedDictionary(DATA_DIR / f"{name}.marisa")
+                for name in DICTIONARY_NAMES
+            }
+        return _dictionaries
+
+
+def register():
+    """Add the Polish dictionaries to zxcvbn. Safe to call many times.
+
+    zxcvbn keeps its ranked dictionaries in a module-level dict, so this
+    affects every later ``zxcvbn()`` call in the process.
+    """
+    from zxcvbn import matching
+
+    matching.RANKED_DICTIONARIES.update(dictionaries())

zxcvbn_pl-0.1.0/tests/test_zxcvbn_pl.py

@@ -0,0 +1,78 @@
+import pytest
+from zxcvbn import matching, zxcvbn
+
+import zxcvbn_pl
+
+
+@pytest.fixture
+def zarejestrowane():
+    przed = dict(matching.RANKED_DICTIONARIES)
+    zxcvbn_pl.register()
+    yield
+    matching.RANKED_DICTIONARIES.clear()
+    matching.RANKED_DICTIONARIES.update(przed)
+
+
+def test_slowniki_maja_dane():
+    slowniki = zxcvbn_pl.dictionaries()
+    assert set(slowniki) == set(zxcvbn_pl.DICTIONARY_NAMES)
+    assert all(len(slownik) > 1000 for slownik in slowniki.values())
+
+
+@pytest.mark.parametrize(
+    "nazwa,slowo",
+    [
+        ("pl_common_words", "kot"),
+        ("pl_first_names", "katarzyna"),
+        ("pl_last_names", "kowalski"),
+        ("pl_last_names", "kowalska"),
+    ],
+)
+def test_slowo_w_slowniku(nazwa, slowo):
+    assert slowo in zxcvbn_pl.dictionaries()[nazwa]
+
+
+def test_wariant_bez_polskich_znakow_ma_te_sama_pozycje():
+    slownik = zxcvbn_pl.dictionaries()["pl_common_words"]
+    assert slownik["zolw"] == slownik["żółw"]
+    assert slownik["slonce"] == slownik["słońce"]
+
+
+def test_czestsze_slowo_ma_nizsza_pozycje():
+    slownik = zxcvbn_pl.dictionaries()["pl_common_words"]
+    assert slownik["nie"] < slownik["kot"] < slownik["żółw"]
+
+
+def test_brak_krotkich_i_nieliterowych_wpisow():
+    slownik = zxcvbn_pl.dictionaries()["pl_common_words"]
+    assert "na" not in slownik
+    assert "xqzvw" not in slownik
+    assert slownik.get("xqzvw") is None
+
+
+def test_register_jest_idempotentne(zarejestrowane):
+    zxcvbn_pl.register()
+    for nazwa in zxcvbn_pl.DICTIONARY_NAMES:
+        assert matching.RANKED_DICTIONARIES[nazwa] is zxcvbn_pl.dictionaries()[nazwa]
+
+
+@pytest.mark.parametrize(
+    "haslo,slownik",
+    [
+        ("słoneczko123456", "pl_common_words"),
+        ("Katarzyna1985!", "pl_first_names"),
+        ("Wojciechowski77", "pl_last_names"),
+    ],
+)
+def test_polskie_slowa_obnizaja_ocene(zarejestrowane, haslo, slownik):
+    wynik = zxcvbn(haslo)
+    assert wynik["score"] <= 2
+    assert slownik in [m.get("dictionary_name") for m in wynik["sequence"]]
+
+
+def test_bez_rejestracji_zxcvbn_nie_zna_polskich_slow():
+    assert zxcvbn("słoneczko123456")["score"] == 4
+
+
+def test_fraza_z_kilku_slow_nadal_mocna(zarejestrowane):
+    assert zxcvbn("zielonaherbatazcytryna")["score"] == 4