regex-based POC

Uses ua-parser/uap-rust#3

Fixes ua-parser#166

Author: masklinn

pyproject.toml

@@ -9,7 +9,6 @@ version = "1.0.0a1"
 readme = "README.rst"
 requires-python = ">=3.8"
 dependencies = []
-optional-dependencies = { yaml = ["PyYaml"], re2 = ["google-re2"] }
 
 license = {text = "Apache 2.0"}
 urls = {repository = "https://github.com/ua-parser/uap-python"}
@@ -39,10 +38,16 @@ classifiers = [
         "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
         "Programming Language :: Python :: 3.11",
+        "Programming Language :: Python :: 3.12",
         "Programming Language :: Python :: Implementation :: CPython",
         "Programming Language :: Python :: Implementation :: PyPy"
 ]
 
+[project.optional-dependencies]
+yaml = ["PyYaml"]
+re2 = ["google-re2"]
+regex = ["ua-parser-rs"]
+
 [tool.ruff.lint]
 select = ["F", "E", "W", "I", "RET", "RUF", "PT"]
 ignore = [

src/ua_parser/__main__.py

@@ -39,11 +39,13 @@
 from .caching import Cache, Local
 from .loaders import load_builtins, load_yaml
 from .re2 import Resolver as Re2Resolver
+from .regex import Resolver as RegexResolver
 from .user_agent_parser import Parse
 
 CACHEABLE = {
     "basic": True,
     "re2": True,
+    "regex": True,
     "legacy": False,
 }
 
@@ -178,6 +180,8 @@ def get_parser(
         r = BasicResolver(rules)
     elif parser == "re2":
         r = Re2Resolver(rules)
+    elif parser == "regex":
+        r = RegexResolver(rules)
     else:
         sys.exit(f"unknown parser {parser!r}")
 
@@ -327,6 +331,7 @@ def run_threaded(args: argparse.Namespace) -> None:
         ("locking-lru", CachingResolver(basic, caching.Lru(CACHESIZE))),
         ("local-lru", CachingResolver(basic, Local(lambda: caching.Lru(CACHESIZE)))),
         ("re2", Re2Resolver(load_builtins())),
+        ("regex", RegexResolver(load_builtins())),
     ]
     for name, resolver in resolvers:
         print(f"{name:11}: ", end="", flush=True)
@@ -436,14 +441,14 @@ def __call__(
 bench.add_argument(
     "--bases",
     nargs="+",
-    choices=["basic", "re2", "legacy"],
-    default=["basic", "re2", "legacy"],
+    choices=["basic", "re2", "regex", "legacy"],
+    default=["basic", "re2", "regex", "legacy"],
     help="""Base resolvers to benchmark. `basic` is a linear search
     through the regexes file, `re2` is a prefiltered regex set
-    implemented in C++, `legacy` is the legacy API (essentially a
-    basic resolver with a clearing cache of fixed 200 entries, but
-    less layered so usually slightly faster than an equivalent
-    basic-based resolver).""",
+    implemented in C++, `regex` is a prefiltered regex set implemented
+    in Rust, `legacy` is the legacy API (essentially a basic resolver
+    with a clearing cache of fixed 200 entries, but less layered so
+    usually slightly faster than an equivalent basic-based resolver).""",
 )
 bench.add_argument(
     "--caches",

src/ua_parser/regex.py

@@ -0,0 +1,76 @@
+__all__ = ["Resolver"]
+
+from operator import attrgetter
+
+import ua_parser_rs  # type: ignore
+
+from .core import (
+    Device,
+    Domain,
+    Matchers,
+    OS,
+    PartialResult,
+    UserAgent,
+)
+
+
+class Resolver:
+    ua: ua_parser_rs.UserAgentExtractor
+    os: ua_parser_rs.OSExtractor
+    de: ua_parser_rs.DeviceExtractor
+
+    def __init__(self, matchers: Matchers) -> None:
+        ua, os, de = matchers
+        self.ua = ua_parser_rs.UserAgentExtractor(
+            map(
+                attrgetter("regex", "family", "major", "minor", "patch", "patch_minor"),
+                ua,
+            )
+        )
+        self.os = ua_parser_rs.OSExtractor(
+            map(
+                attrgetter("regex", "family", "major", "minor", "patch", "patch_minor"),
+                os,
+            )
+        )
+        self.de = ua_parser_rs.DeviceExtractor(
+            map(
+                attrgetter("regex", "regex_flag", "family", "brand", "model"),
+                de,
+            )
+        )
+
+    def __call__(self, ua: str, domains: Domain, /) -> PartialResult:
+        user_agent = os = device = None
+        if Domain.USER_AGENT in domains:
+            if m := self.ua.extract(ua):
+                user_agent = UserAgent(
+                    m.family,
+                    m.major,
+                    m.minor,
+                    m.patch,
+                    m.patch_minor,
+                )
+        if Domain.OS in domains:
+            if m := self.os.extract(ua):
+                os = OS(
+                    m.os,
+                    m.major,
+                    m.minor,
+                    m.patch,
+                    m.patch_minor,
+                )
+        if Domain.DEVICE in domains:
+            if m := self.de.extract(ua):
+                device = Device(
+                    m.family,
+                    m.brand,
+                    m.model,
+                )
+        return PartialResult(
+            domains=domains,
+            string=ua,
+            user_agent=user_agent,
+            os=os,
+            device=device,
+        )

tests/test_core.py

@@ -53,6 +53,19 @@
 else:
     PARSERS.append(pytest.param(Parser(re2.Resolver(load_builtins())), id="re2"))
 
+try:
+    from ua_parser import regex
+except ImportError:
+    PARSERS.append(
+        pytest.param(
+            None,
+            id="regex",
+            marks=pytest.mark.skip(reason="regex parser not available"),
+        )
+    )
+else:
+    PARSERS.append(pytest.param(Parser(regex.Resolver(load_builtins())), id="regex"))
+
 UA_FIELDS = {f.name for f in dataclasses.fields(UserAgent)}
 
 

tox.ini

@@ -20,6 +20,7 @@ deps =
      pytest
      pyyaml
      google-re2
+     ua-parser-rs
 commands =
     pytest -Werror --doctest-glob="*.rst" {posargs}