new vastly simplified plugin spec without pydantic

archivebox/plugins_search/ripgrep/__init__.py

@@ -0,0 +1,48 @@
+__package__ = 'plugins_search.ripgrep'
+__label__ = 'ripgrep'
+__version__ = '2024.10.14'
+__author__ = 'Nick Sweeting'
+__homepage__ = 'https://github.com/BurntSushi/ripgrep'
+__dependencies__ = []
+
+import abx
+
+
+@abx.hookimpl
+def get_PLUGIN():
+    return {
+        'ripgrep': {
+            'PACKAGE': __package__,
+            'LABEL': __label__,
+            'VERSION': __version__,
+            'AUTHOR': __author__,
+            'HOMEPAGE': __homepage__,
+            'DEPENDENCIES': __dependencies__,
+        }
+    }
+
+@abx.hookimpl
+def get_CONFIG():
+    from .config import RIPGREP_CONFIG
+    
+    return {
+        'ripgrep': RIPGREP_CONFIG
+    }
+
+
+@abx.hookimpl
+def get_BINARIES():
+    from .binaries import RIPGREP_BINARY
+    
+    return {
+        'ripgrep': RIPGREP_BINARY
+    }
+
+
+@abx.hookimpl
+def get_SEARCHBACKENDS():
+    from .searchbackend import RIPGREP_SEARCH_BACKEND
+    
+    return {
+        'ripgrep': RIPGREP_SEARCH_BACKEND,
+    }

archivebox/plugins_search/ripgrep/apps.py

@@ -1,114 +0,0 @@
-__package__ = 'archivebox.plugins_search.ripgrep'
-
-import re
-from pathlib import Path
-from subprocess import run
-from typing import List, Iterable
-# from typing_extensions import Self
-
-# Depends on other PyPI/vendor packages:
-from pydantic import InstanceOf, Field
-from pydantic_pkgr import BinProvider, BinaryOverrides, BinName
-
-# Depends on other Django apps:
-from abx.archivebox.base_plugin import BasePlugin
-from abx.archivebox.base_configset import BaseConfigSet
-from abx.archivebox.base_binary import BaseBinary, env, apt, brew
-from abx.archivebox.base_hook import BaseHook
-from abx.archivebox.base_searchbackend import BaseSearchBackend
-
-# Depends on Other Plugins:
-from archivebox.config import CONSTANTS
-from archivebox.config.common import SEARCH_BACKEND_CONFIG
-
-###################### Config ##########################
-
-class RipgrepConfig(BaseConfigSet):
-    RIPGREP_BINARY: str = Field(default='rg')
-    
-    RIPGREP_IGNORE_EXTENSIONS: str = Field(default='css,js,orig,svg')
-    RIPGREP_ARGS_DEFAULT: List[str] = Field(default=lambda c: [
-        # https://github.com/BurntSushi/ripgrep/blob/master/GUIDE.md
-        f'--type-add=ignore:*.{{{c.RIPGREP_IGNORE_EXTENSIONS}}}',
-        '--type-not=ignore',
-        '--ignore-case',
-        '--files-with-matches',
-        '--regexp',
-    ])
-    RIPGREP_SEARCH_DIR: Path = CONSTANTS.ARCHIVE_DIR
-
-RIPGREP_CONFIG = RipgrepConfig()
-
-
-
-class RipgrepBinary(BaseBinary):
-    name: BinName = RIPGREP_CONFIG.RIPGREP_BINARY
-    binproviders_supported: List[InstanceOf[BinProvider]] = [apt, brew, env]
-
-    overrides: BinaryOverrides = {
-        apt.name: {'packages': ['ripgrep']},
-        brew.name: {'packages': ['ripgrep']},
-    }
-
-RIPGREP_BINARY = RipgrepBinary()
-
-# regex to match archive/<ts>/... snapshot dir names
-TIMESTAMP_REGEX =  re.compile(r'\/([\d]+\.[\d]+)\/')
-
-class RipgrepSearchBackend(BaseSearchBackend):
-    name: str = 'ripgrep'
-    docs_url: str = 'https://github.com/BurntSushi/ripgrep'
-    
-    @staticmethod
-    def index(snapshot_id: str, texts: List[str]):
-        return
-
-    @staticmethod
-    def flush(snapshot_ids: Iterable[str]):
-        return
-
-    @staticmethod
-    def search(text: str) -> List[str]:
-        from core.models import Snapshot
-        
-        ripgrep_binary = RIPGREP_BINARY.load()
-        if not ripgrep_binary.version:
-            raise Exception("ripgrep binary not found, install ripgrep to use this search backend")
-    
-        cmd = [
-            ripgrep_binary.abspath, 
-            *RIPGREP_CONFIG.RIPGREP_ARGS_DEFAULT,
-            text,
-            str(RIPGREP_CONFIG.RIPGREP_SEARCH_DIR),
-        ]
-        proc = run(cmd, timeout=SEARCH_BACKEND_CONFIG.SEARCH_BACKEND_TIMEOUT, capture_output=True, text=True)
-        timestamps = set()
-        for path in proc.stdout.splitlines():
-            ts = TIMESTAMP_REGEX.findall(path)
-            if ts:
-                timestamps.add(ts[0])
-        
-        snap_ids = [str(id) for id in Snapshot.objects.filter(timestamp__in=timestamps).values_list('pk', flat=True)]
-    
-        return snap_ids
-
-RIPGREP_SEARCH_BACKEND = RipgrepSearchBackend()
-
-
-
-
-class RipgrepSearchPlugin(BasePlugin):
-    app_label: str ='ripgrep'
-    verbose_name: str = 'Ripgrep'
-
-    hooks: List[InstanceOf[BaseHook]] = [
-        RIPGREP_CONFIG,
-        RIPGREP_BINARY,
-        RIPGREP_SEARCH_BACKEND,
-    ]
-
-
-
-PLUGIN = RipgrepSearchPlugin()
-# PLUGIN.register(settings)
-DJANGO_APP = PLUGIN.AppConfig

archivebox/plugins_search/ripgrep/binaries.py

@@ -0,0 +1,23 @@
+__package__ = 'plugins_search.ripgrep'
+
+from typing import List
+
+from pydantic import InstanceOf
+from pydantic_pkgr import BinProvider, BinaryOverrides, BinName
+
+from abx.archivebox.base_binary import BaseBinary, env, apt, brew
+
+
+from .config import RIPGREP_CONFIG
+
+
+class RipgrepBinary(BaseBinary):
+    name: BinName = RIPGREP_CONFIG.RIPGREP_BINARY
+    binproviders_supported: List[InstanceOf[BinProvider]] = [apt, brew, env]
+
+    overrides: BinaryOverrides = {
+        apt.name: {'packages': ['ripgrep']},
+        brew.name: {'packages': ['ripgrep']},
+    }
+
+RIPGREP_BINARY = RipgrepBinary()

archivebox/plugins_search/ripgrep/config.py

@@ -0,0 +1,29 @@
+__package__ = 'plugins_search.ripgrep'
+
+from pathlib import Path
+from typing import List
+
+from pydantic import Field
+
+from abx.archivebox.base_configset import BaseConfigSet
+
+from archivebox.config import CONSTANTS
+from archivebox.config.common import SEARCH_BACKEND_CONFIG
+
+
+class RipgrepConfig(BaseConfigSet):
+    RIPGREP_BINARY: str = Field(default='rg')
+    
+    RIPGREP_IGNORE_EXTENSIONS: str = Field(default='css,js,orig,svg')
+    RIPGREP_ARGS_DEFAULT: List[str] = Field(default=lambda c: [
+        # https://github.com/BurntSushi/ripgrep/blob/master/GUIDE.md
+        f'--type-add=ignore:*.{{{c.RIPGREP_IGNORE_EXTENSIONS}}}',
+        '--type-not=ignore',
+        '--ignore-case',
+        '--files-with-matches',
+        '--regexp',
+    ])
+    RIPGREP_SEARCH_DIR: Path = CONSTANTS.ARCHIVE_DIR
+    RIPGREP_TIMEOUT: int = Field(default=lambda: SEARCH_BACKEND_CONFIG.SEARCH_BACKEND_TIMEOUT)
+
+RIPGREP_CONFIG = RipgrepConfig()

archivebox/plugins_search/ripgrep/searchbackend.py

@@ -0,0 +1,55 @@
+__package__ = 'plugins_search.ripgrep'
+
+import re
+import subprocess
+
+from typing import List, Iterable
+
+from abx.archivebox.base_searchbackend import BaseSearchBackend
+
+from .binaries import RIPGREP_BINARY
+from .config import RIPGREP_CONFIG
+
+
+
+# regex to match archive/<ts>/... snapshot dir names
+TIMESTAMP_REGEX =  re.compile(r'\/([\d]+\.[\d]+)\/')
+
+class RipgrepSearchBackend(BaseSearchBackend):
+    name: str = 'ripgrep'
+    docs_url: str = 'https://github.com/BurntSushi/ripgrep'
+    
+    @staticmethod
+    def index(snapshot_id: str, texts: List[str]):
+        return
+
+    @staticmethod
+    def flush(snapshot_ids: Iterable[str]):
+        return
+
+    @staticmethod
+    def search(text: str) -> List[str]:
+        from core.models import Snapshot
+        
+        ripgrep_binary = RIPGREP_BINARY.load()
+        if not ripgrep_binary.version:
+            raise Exception("ripgrep binary not found, install ripgrep to use this search backend")
+    
+        cmd = [
+            ripgrep_binary.abspath, 
+            *RIPGREP_CONFIG.RIPGREP_ARGS_DEFAULT,
+            text,
+            str(RIPGREP_CONFIG.RIPGREP_SEARCH_DIR),
+        ]
+        proc = subprocess.run(cmd, timeout=RIPGREP_CONFIG.RIPGREP_TIMEOUT, capture_output=True, text=True)
+        timestamps = set()
+        for path in proc.stdout.splitlines():
+            ts = TIMESTAMP_REGEX.findall(path)
+            if ts:
+                timestamps.add(ts[0])
+        
+        snap_ids = [str(id) for id in Snapshot.objects.filter(timestamp__in=timestamps).values_list('pk', flat=True)]
+    
+        return snap_ids
+
+RIPGREP_SEARCH_BACKEND = RipgrepSearchBackend()