Split Parser and reorganise package

Parser turns out to not really make sense as a superclass / ABC: it
really only has one useful method, and because parsers use delegation
there's no real way to override the utility methods / shortcuts, so
they're only useful on the caller / client side but they constrain the
implementor (who has to extend the ABC and then possibly deal with
multiple-inheritance shenanigans).

Making the core object just a callable protocol instead makes the
implementation somewhat simpler and more flexible (e.g. just a
function or HoF can be a "parser"), however the convenient utility
methods *are* important for end users and should not be discounted.

For that, keep a wrapper `Parser` object which can be wrapped around a
"parser" in order to provide the additional convenience (similar to
the free functions at the root). Importantly, `Parser` methods can
also be used as free functions by passing a "parser" as `self`, they
are intended to be compatible. It doesn't work super well from the
typechecking perspective, but it works fine enough.

Consideration was given to making the free functions at the package
root parametric on the parser e.g.

    def parse(ua: str, resolver: Optional[Resolver] = None, /) -> ParseResult:
        if resolver is None:
            from . import parser as resolver

        return resolver(ua, Domain.ALL).complete()

but that feels like it would be pretty error prone, in the sense that
it would be too easy to forget to pass in the resolver, compared to
consistently resolving via a bespoke parser, or just installing a
parser globally.

Also move things around a bit:

- move matcher utility functions out of the core, un-prefix them since
  we're using `__all__` for visibility anyway
- move eager matchers out of the core, similar to the lazy matchers

Fixes #189

Author: masklinn

setup.py

@@ -182,7 +182,7 @@ class EagerWriter(Writer):
 __all__ = ["MATCHERS"]
 
 from typing import Tuple, List
-from .core import UserAgentMatcher, OSMatcher, DeviceMatcher
+from .matchers import UserAgentMatcher, OSMatcher, DeviceMatcher
 
 MATCHERS: Tuple[List[UserAgentMatcher], List[OSMatcher], List[DeviceMatcher]] = ([
 """

src/ua_parser/__init__.py

@@ -16,29 +16,25 @@
 This way importing anything but the top-level package should not be
 necessary unless you want to *implement* a parser.
 """
+from __future__ import annotations
 
 __all__ = [
-    "BasicParser",
-    "CachingParser",
+    "BasicResolver",
+    "CachingResolver",
     "Clearing",
     "DefaultedParseResult",
     "Device",
-    "DeviceMatcher",
     "Domain",
     "LRU",
     "Locking",
     "Matchers",
     "OS",
-    "OSMatcher",
     "ParseResult",
-    "Parser",
+    "Resolver",
     "PartialParseResult",
     "UserAgent",
-    "UserAgentMatcher",
     "load_builtins",
     "load_lazy_builtins",
-    "load_data",
-    "load_yaml",
     "parse",
     "parse_device",
     "parse_os",
@@ -48,43 +44,89 @@
 import contextlib
 from typing import Callable, Optional
 
-from .basic import Parser as BasicParser
-from .caching import CachingParser, Clearing, Locking, LRU
+from .basic import Resolver as BasicResolver
+from .caching import CachingResolver, Clearing, Locking, LRU
 from .core import (
     DefaultedParseResult,
     Device,
-    DeviceMatcher,
     Domain,
     Matchers,
     OS,
-    OSMatcher,
-    Parser,
     ParseResult,
     PartialParseResult,
+    Resolver,
     UserAgent,
-    UserAgentMatcher,
 )
-from .loaders import load_builtins, load_data, load_lazy_builtins, load_yaml
+from .loaders import load_builtins, load_lazy_builtins
 
-Re2Parser: Optional[Callable[[Matchers], Parser]] = None
+Re2Resolver: Optional[Callable[[Matchers], Resolver]] = None
 with contextlib.suppress(ImportError):
-    from .re2 import Parser as Re2Parser
+    from .re2 import Resolver as Re2Resolver
 
 
 VERSION = (1, 0, 0)
+
+
+class Parser:
+    @classmethod
+    def from_matchers(cls, m: Matchers, /) -> Parser:
+        if Re2Resolver is not None:
+            return cls(Re2Resolver(m))
+        else:
+            return cls(
+                CachingResolver(
+                    BasicResolver(m),
+                    Locking(LRU(200)),
+                )
+            )
+
+    def __init__(self, resolver: Resolver) -> None:
+        self.resolver = resolver
+
+    def __call__(self, ua: str, domains: Domain, /) -> PartialParseResult:
+        """Parses the ``ua`` string, returning a parse result with *at least*
+        the requested :class:`domains <Domain>` resolved (whether to success or
+        failure).
+
+        A parser may resolve more :class:`domains <Domain>` than
+        requested, but it *must not* resolve less.
+        """
+        return self.resolver(ua, domains)
+
+    def parse(self, ua: str) -> ParseResult:
+        """Convenience method for parsing all domains, and falling back to
+        default values for all failures.
+        """
+        return self(ua, Domain.ALL).complete()
+
+    def parse_user_agent(self, ua: str) -> Optional[UserAgent]:
+        """Convenience method for parsing the :class:`UserAgent` domain,
+        falling back to the default value in case of failure.
+        """
+        return self(ua, Domain.USER_AGENT).user_agent
+
+    def parse_os(self, ua: str) -> Optional[OS]:
+        """Convenience method for parsing the :class:`OS` domain, falling back
+        to the default value in case of failure.
+        """
+        return self(ua, Domain.OS).os
+
+    def parse_device(self, ua: str) -> Optional[Device]:
+        """Convenience method for parsing the :class:`Device` domain, falling
+        back to the default value in case of failure.
+        """
+        return self(ua, Domain.DEVICE).device
+
+
 parser: Parser
 
 
 def __getattr__(name: str) -> Parser:
     global parser
     if name == "parser":
-        if Re2Parser is not None:
-            parser = Re2Parser(load_lazy_builtins())
-        else:
-            parser = CachingParser(
-                BasicParser(load_builtins()),
-                Locking(LRU(200)),
-            )
+        parser = Parser.from_matchers(
+            load_builtins() if Re2Resolver is None else load_lazy_builtins()
+        )
         return parser
     raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
 
@@ -105,7 +147,7 @@ def parse(ua: str) -> ParseResult:
     # parser, a `global` access fails to and we get a NameError
     from . import parser
 
-    return parser.parse(ua)
+    return parser(ua, Domain.ALL).complete()
 
 
 def parse_user_agent(ua: str) -> Optional[UserAgent]:
@@ -114,7 +156,7 @@ def parse_user_agent(ua: str) -> Optional[UserAgent]:
     """
     from . import parser
 
-    return parser.parse_user_agent(ua)
+    return parser(ua, Domain.USER_AGENT).user_agent
 
 
 def parse_os(ua: str) -> Optional[OS]:
@@ -123,7 +165,7 @@ def parse_os(ua: str) -> Optional[OS]:
     """
     from . import parser
 
-    return parser.parse_os(ua)
+    return parser(ua, Domain.OS).os
 
 
 def parse_device(ua: str) -> Optional[Device]:
@@ -132,4 +174,4 @@ def parse_device(ua: str) -> Optional[Device]:
     """
     from . import parser
 
-    return parser.parse_device(ua)
+    return parser(ua, Domain.DEVICE).device

src/ua_parser/_matchers.pyi

@@ -2,7 +2,7 @@ __all__ = ["MATCHERS"]
 
 from typing import List, Tuple
 
-from .core import DeviceMatcher, OSMatcher, UserAgentMatcher
+from .matchers import DeviceMatcher, OSMatcher, UserAgentMatcher
 
 MATCHERS: Tuple[
     List[UserAgentMatcher],

src/ua_parser/basic.py

@@ -1,3 +1,5 @@
+__all__ = ["Resolver"]
+
 from operator import methodcaller
 from typing import List
 
@@ -7,13 +9,12 @@
     Matcher,
     Matchers,
     OS,
-    Parser as AbstractParser,
     PartialParseResult,
     UserAgent,
 )
 
 
-class Parser(AbstractParser):
+class Resolver:
     """A simple pure-python parser based around trying a numer of regular
     expressions in sequence for each domain, and returning a result
     when one matches.
@@ -27,9 +28,7 @@ def __init__(
         self,
         matchers: Matchers,
     ) -> None:
-        self.user_agent_matchers = matchers[0]
-        self.os_matchers = matchers[1]
-        self.device_matchers = matchers[2]
+        self.user_agent_matchers, self.os_matchers, self.device_matchers = matchers
 
     def __call__(self, ua: str, domains: Domain, /) -> PartialParseResult:
         parse = methodcaller("__call__", ua)

src/ua_parser/bench.py

@@ -7,18 +7,18 @@
 from typing import Any, Callable, Iterable, List, Optional
 
 from . import (
-    BasicParser,
-    CachingParser,
+    BasicResolver,
+    CachingResolver,
     Clearing,
     Locking,
     LRU,
     Matchers,
     Parser,
-    load_builtins,
-    load_yaml,
+    Resolver,
 )
 from .caching import Cache
-from .re2 import Parser as Re2Parser
+from .loaders import load_builtins, load_yaml
+from .re2 import Resolver as Re2Resolver
 from .user_agent_parser import Parse
 
 CACHEABLE = {
@@ -222,19 +222,19 @@ def run_csv(args: argparse.Namespace) -> None:
 def get_parser(
     parser: str, cache: str, cachesize: int, rules: Matchers
 ) -> Callable[[str], Any]:
-    p: Parser
+    r: Resolver
     if parser == "legacy":
         return Parse
     elif parser == "basic":
-        p = BasicParser(rules)
+        r = BasicResolver(rules)
     elif parser == "re2":
-        p = Re2Parser(rules)
+        r = Re2Resolver(rules)
     else:
         sys.exit(f"unknown parser {parser!r}")
 
     c: Callable[[int], Cache]
     if cache == "none":
-        return p.parse
+        return Parser(r).parse
     elif cache == "clearing":
         c = Clearing
     elif cache == "lru":
@@ -244,7 +244,7 @@ def get_parser(
     else:
         sys.exit(f"unknown cache algorithm {cache!r}")
 
-    return CachingParser(p, c(cachesize)).parse
+    return Parser(CachingResolver(r, c(cachesize))).parse
 
 
 def run(

src/ua_parser/caching.py

@@ -1,20 +1,20 @@
 import abc
 import threading
 from collections import OrderedDict
-from typing import Dict, Optional
+from typing import Dict, Optional, Protocol
 
-from .core import Domain, Parser, PartialParseResult
+from .core import Domain, PartialParseResult, Resolver
 
 __all__ = [
-    "CachingParser",
+    "CachingResolver",
     "Cache",
     "Clearing",
     "Locking",
     "LRU",
 ]
 
 
-class Cache(abc.ABC):
+class Cache(Protocol):
     """Cache abstract protocol. The :class:`CachingParser` will look
     values up, merge what was returned (possibly nothing) with what it
     got from its actual parser, and *re-set the result*.
@@ -33,7 +33,7 @@ def __getitem__(self, key: str) -> Optional[PartialParseResult]:
         ...
 
 
-class Clearing(Cache):
+class Clearing:
     """A clearing cache, if the cache is full, just remove all the entries
     and re-fill from scratch.
 
@@ -62,7 +62,7 @@ def __setitem__(self, key: str, value: PartialParseResult) -> None:
         self.cache[key] = value
 
 
-class LRU(Cache):
+class LRU:
     """Cache following a least-recently used replacement policy: when
     there is no more room in the cache, whichever entry was last seen
     the least recently is removed.
@@ -103,7 +103,7 @@ def __setitem__(self, key: str, value: PartialParseResult) -> None:
             self.cache.popitem(last=False)
 
 
-class Locking(Cache):
+class Locking:
     """Locking cache decorator. Takes a non-thread-safe cache and
     ensures retrieving and setting entries is protected by a mutex.
 
@@ -122,7 +122,7 @@ def __setitem__(self, key: str, value: PartialParseResult) -> None:
             self.cache[key] = value
 
 
-class CachingParser(Parser):
+class CachingResolver:
     """A wrapping parser which takes an underlying concrete :class:`Cache`
     for the actual caching and cache strategy.
 
@@ -134,8 +134,8 @@ class CachingParser(Parser):
     really, they're immutable).
     """
 
-    def __init__(self, parser: Parser, cache: Cache):
-        self.parser: Parser = parser
+    def __init__(self, parser: Resolver, cache: Cache):
+        self.parser: Resolver = parser
         self.cache: Cache = cache
 
     def __call__(self, ua: str, domains: Domain, /) -> PartialParseResult: